Railway Transportation : Policies, Technology and Perspectives [1 ed.] 9781617285745, 9781606928639

Citation preview

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

TRANSPORTATION ISSUES, POLICIES AND R&D SERIES

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

RAILWAY TRANSPORTATION: POLICIES, TECHNOLOGY AND PERSPECTIVES

No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information Railway Transportation : Policies, Technology edited byis Nicholas P. Scott, Publishers,that Incorporated, 2008. ProQuest contained herein. and ThisPerspectives, digital document sold with the Nova clearScience understanding the publisher is not engaged in

TRANSPORTATION ISSUES, POLICIES AND R&D SERIES Public Transit Issues and Developments Calvin B. Lang (Editor) 2009. ISBN 978-1-60692-689-5

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Railway Transportation: Policies, Technology and Perspectives Nicholas P. Scott (Editor) 2009. ISBN 978-1-60692-8639

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

TRANSPORTATION ISSUES, POLICIES AND R&D SERIES

RAILWAY TRANSPORTATION: POLICIES, TECHNOLOGY AND PERSPECTIVES

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

NICHOLAS P. SCOTT EDITOR

Nova Science Publishers, Inc. New York

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2009 by Nova Science Publishers, Inc.

All rights reserved. No part of this book may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical photocopying, recording or otherwise without the written permission of the Publisher. For permission to use material from this book please contact us: Telephone 631-231-7269; Fax 631-231-8175 Web Site: http://www.novapublishers.com NOTICE TO THE READER The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained in this book. The Publisher shall not be liable for any special, consequential, or exemplary damages resulting, in whole or in part, from the readers’ use of, or reliance upon, this material.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Independent verification should be sought for any data, advice or recommendations contained in this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage to persons or property arising from any methods, products, instructions, ideas or otherwise contained in this publication. This publication is designed to provide accurate and authoritative information with regard to the subject matter cover herein. It is sold with the clear understanding that the Publisher is not engaged in rendering legal or any other professional services. If legal, medical or any other expert assistance is required, the services of a competent person should be sought. FROM A DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS. Library of Congress Cataloging-in-Publication Data Railway transportation : policies, technology, and perspectives / editor, Nicholas P. Scott. p. cm. Includes index. ISBN H%RRN 1. Railroads. 2. Railroad engineering. I. Scott, Nicholas P. TF145.R25 2009 385--dc22 2009000147

Published by Nova Science Publishers, Inc.

New York

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

CONTENTS

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Preface

vii

Chapter 1

Railway Infrastructure Market in Europe Christina Nikolova

Chapter 2

Explaining Occupancy Rates in the European Railways: A Reduced-Form Approach Luis Orea, Ana Rodríguez-Álvarez and Subal C. Kumbhakar

Chapter 3

Railway-Generated Magnetic Field: Environmental Aspects N. G. Ptitsyna, G. Villoresi and Y. A. Kopytenko

Chapter 4

Marginal Cost Pricing of Noise in Railway Infrastructure Henrik Andersson and Mikael Ögren

Chapter 5

Planners, Cognition and the Route towards Improved Planning Support: An Empirical Study into the Use of APS’s in the Netherlands Railways René J. Jorna, Wout van Wezel and Joep Bos

Chapter 6

Remaining Fatigue Life Estimation of Existing Railway Bridges Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake and Kaita Tatsumasa

Chapter 7

Maintenance Management Strategies for Steel Railway Bridges: A Case Study from Vietnam Dinh Tuan Hai

Chapter 8

Chapter 9

Chapter 10

Noise and Vibrations of Railway Wheels: Generation Mechanisms and Attenuation Alfredo Cigada, Stefano Manzoni, Matteo Redaelli and Marcello Vanali Train Wheel Detection Systems: Present Technologies and Future Trends Patricio G. Donato and Jesús Ureña Track Stiffness Considerations for High Speed Railway Lines Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg and Eric Berggren

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

1

63 87 141

163 181

217

237

279 303

vi Chapter 11

Contents Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail D. Vangi and A. Virga

Expert Commentary Edmond Fortier and the Lagos Tram (1908) Liora Bigon and Frank Brown

355 379 379

Short Communications A Security Risk Management of Railway Transportation Systems Francesco Flammini and Nicola Mazzocca

385

Short Communications B Study on Railway Accessibility Mode Choice: A Case study on Urban Railway in Beijing Huang Shan, Guan Hongzhi and Yan Hai

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Index

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

399 413

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

PREFACE This book provides the latest scientific research regarding the importance of infrastructure charges in establishing competitive conditions in the railway market. The current charging regimes applied throughout the EU member states are analyzed as well as the planning and scheduling that determine how and when the company's resources will be used in the case of railway organizations. Railway noise emission and its reduction are considered among the most important topics in the future development of transportation systems. This book gives an overview on the noise emitted by wheels and rails from the basic emission mechanisms up to noise attenuation by means of passive/active control. The importance of the vertical track stiffness as a means to guide railway track bed design for high speed railway lines is discussed as well. A rational approach to substructure design is described, which it is hoped will further an understanding of the process of appropriate track design and enable the adaptation of existing design procedures to provide a realistic design for the conditions at hand. Chapter 1 – This chapter presents a brief review of the European Transport Policy in railway transport. It provides an outline of the development of the railway infrastructure market in Europe and examines the means for achieving the main goals concerning revitalization of the railways and especially for using the railway infrastructure in a more efficient way. The current state of restructuring in the different countries and the models for the development of the operators and infrastructure enterprises are described as well. The emphasis is laid on the importance of the measures for regulating the rail infrastructure as a natural monopoly. This chapter also takes into account the management and regulation of the railway infrastructure as key factors in the development of the rail infrastructure market. The latest scientific research regarding the importance of infrastructure charges in establishing competitive conditions in the railway market and in defining possible problems is presented. The current charging regimes applied throughout the EU member states are analyzed. An accent is placed on the caused external costs and possibilities for their internalization. A comparative analysis between railway and other modes of transport is made. From the point of view of the expected traffic growth in Europe, the chapter looks through the possibilities for better integration of the separate national and regional networks and the optimization of the use of existing capacity. It takes into account the opportunities for linking up railways and other modes of transport through offering combined transport services and building up intermodal terminals. Finally, some implications regarding the sustainability of the railway transport are drawn up, as well as issues linked to certain economic, social and environmental aspects.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

viii

Nicholas P. Scott

Chapter 2 – Occupancy rates of vehicles are often used to analyze the firm performance in the transport industry. The occupancy rates tend to be low in many of the European countries and there are important differences in occupancy rates among countries. To understand better the reserved capacity that transport companies are maintaining in the European railway sector, the authors propose using a reduced-form approach to identify the determinants of the observed occupancy rates. Our results suggest that demand uncertainty, adjustment costs, “fidelity” strategies and social obligations are important determinants of annual rates of passenger occupation. Chapter 3 – Recent epidemiological studies suggest a link between magnetic fields generated by electrified transportation systems and certain adverse health effects, even though these magnetic fields are generally lower than international limits. Since many people ride on railways daily, such magnetic field exposures should be examined thoroughly. Here the authors will present a short review of epidemiological research among electrical transportation workers and research results of our waveform measurements of complex “realworld” magnetic fields generated by DC- and AC-powered railways. The obtained measurement information allowed quantitative characterization of specific magnetic field features that could be biologically important: frequency and intensity “windows”, interaction with DC-field, geometry (e.g., polarization) and other aspects. Special attention will be drawn to explore the extent to which railway magnetic fields contribute to the total magnetic field environment, and to compare these fields with natural geomagnetic field background. The authors have conducted an extensive study of magnetic fields on different electrified rail systems. The studied transport technologies were: (i) Russian DC-powered locomotives; (ii) Swiss AC-powered (16.67 Hz) locomotives; and (iii) Russian DC-powered trains formed by a number of self-powered electric motor units and units without motors. Measurements have been performed in the ULF-ELF frequency range (0–50 Hz) by means of a novel sophisticated portable computer-based waveform capture system. This measurement system (sampling rate up to 200 Hz) allows continuous recording and characterization of a field profile over time and frequency along three axes. The authors demonstrated the practicality of the data acquisition and analysis protocols used in this study to estimate exposure characteristics beyond TWA (time-weighted average) that might be of interest to healthrelated magnetic field researchers. The authors developed a set of improved methods, algorithms and software to quantify characteristics of complex-spectra railway magnetic field, such as variability in different frequency ranges, amplitude-frequency dependence, polarization and intermittency. These exposure parameters have been determined in engineers’ workplaces and passengers’ coaches of both DC- and AC-powered railway systems. Magnetic fields encountered on DC- and AC-powered transport systems are different from power-line fields, which are predominantly sinusoidal with main frequency at 50 or 60 Hz. Railway magnetic fields present complex patterns resulting from the superposition of variations with different amplitudes, frequencies and geometries; the main energy is concentrated in the lowest frequencies. Possible recognition of the main peculiarities of the railway magnetic field could be useful in the identification of sources of these specific features. It will allow developing design-related preventive measures to diminish the health-hazardous potential of magnetic field. Magnetic field survey and data analysis have been conducted in the framework of the European INCO-Copernicus project “Improvement of Methods of Exposure Assessment for Magnetic Fields from Electric Traction with Regard to Coronary Heart Diseases”.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Preface

ix

Chapter 4 – In order to mitigate negative effects from traffic it has been decided that infrastructure charges in the European Union (EU) should be based on short run marginal costs. The Swedish Parliament has legislated that operators in the Swedish railway infrastructure must pay charges based on short run marginal social costs in order to mitigate externalities in railway infrastructure. Internalization of the social cost of noise is of particular interest, since it is the only environmental problem perceived as more troublesome today than in the early 1990s. Inclusion of a noise component in rail infrastructure charges raises two issues: (i) the monetary evaluation of noise abatement, since noise is a non-market good, and (ii) the estimation of the effect on the noise level that one extra train will create. Regarding the latter, the authors are interested in the marginal noise, since infrastructure charges based on the short-run marginal cost principle should be based on the effect from the marginal train, not the noise level itself. Using already existing knowledge, this study shows that it is possible to implement a noise component in the rail infrastructure charges. The values that are used today to estimate the social cost of noise exposure in cost benefit analysis can also be used to calculate the marginal cost. The authors recommend, however, that further research be carried out in order to get more robust estimates based on railway traffic. The authors also show that the existing noise estimation models can easily be modified to estimate the marginal noise. Noise infrastructure charges give the operators incentives to reduce their noise emissions. They believe that this kind of charge can be used to reduce overall emission levels to an optimal social level, but that it is important for the charge to be based on monetary estimates for rail-traffic and not road-traffic. Chapter 5 – In many organizations, planning and scheduling are extremely important because they determine how and when the company’s resources will be used. This is especially the case in railway organizations, where many resources must be coordinated nationwide: train coaches, train drivers, ticket collectors, railroad tracks, shunting yards, shunting staff, etc. For such complex planning and scheduling situations, computer support is indispensable. In the past decades, mathematical algorithms have been incorporated in scheduling systems to improve schedules, but in many cases, this has not resulted in the expected improvements in performance. The authors performed experiments to investigate this problem for the decision support system that is used by staff planners of The Netherlands Railways. In the experiments, planners solve simple and complex problems with and without support of an algorithm. The authors report on the effect of the use of the algorithm on mental load and task performance. Chapter 6 – The assessment of the remaining fatigue life of a railway bridge for continuing service has become more important. The present day accepted fatigue life assessment approach to railway bridges is generally based on a combination of measured stress histories, Miner’s rule and railway-code–provided fatigue curve. Even though the past strain measurements are available for major bridges, most of the old bridges do not have past strain measurements. Therefore, the application of available methods is limited to particular bridges where the information of previous loading histories is available. Furthermore, in the case of existing railway bridges where the detailed loading history is known, Miner’s rule might provide incorrect results because of its omission of load sequence effect. Therefore, it is inaccurate to use Miner’s rule for the remaining fatigue life estimation of railway bridges because most railway bridges are subjected to variable amplitude loadings. Meanwhile, a new damage-indicator–based sequential law was developed to capture the loading sequence effect of variable amplitude loads more precisely. The authors have applied this theory to estimate

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

x

Nicholas P. Scott

the remaining fatigue life of railway bridges in their previous publications, and it was proven that sequential law is more applicable to railway bridges than the previous theories. However, there are more issues to be addressed in generalizing sequential law as it applies to railway bridges. This chapter therefore proposes a reasonably accurate method to predict the past stress histories from present day measured strains and a method to apply the new damageindicator–based sequential law for remaining fatigue life estimation of railway bridges. Initially, this chapter describes the proposed method to predict the past stress histories. Then, the method of application of new fatigue law is described comprehensively. It describes i) the basic concept of fatigue law, ii) a technique that utilizes transfer of the partially-known codeprovided Wöhler curve to the fully-known curve, iii) the proposed extension of new fatigue law to evaluate secondary stress- (stress concentration- ) based fatigue life of bridge connections and iv) experimental verifications of proposed methods. Then the proposed method is applied to estimate the remaining fatigue life of an old existing railway bridge. Finally, comparisons of the results are made with Miner’s rule-based previous estimation. Hence validity and applicability of the proposed approach are discussed. Chapter 7 – Bridge defects have become a major social concern in recent times. Defects on railway bridges in Vietnam have been identified as structural failure, corrosion, fatigue, functional obsolescence, aging and human intrusion. In addition, problems of low maintenance budget, rigidity and long lines of communication between relevant stakeholders, inadequate management attention and lack of prioritization in maintenance selection adversely impacts bridge structures. This study proposes a series of maintenance strategies for overcoming the current problems. Essential strategies are suggested for the short-term to delay the onset of deterioration by extending service life. Preventive strategies aim to eliminate potential causes early, thus the required resources and time expanded on remedies of future defects can be substantially reduced. It is recommended that preventive strategies should be given priority to eliminate causes of potential problems. Moreover, essential maintenance must be regularly carried out to eliminate serious defects. These goals can be achieved only if training of involved parties and their in-house personnel becomes imperative to identify failure modes and their root causes, and to decide on appropriate preventive and corrective solutions. Further research is needed to verify the proposed maintenance strategies on site and to modify them to fit the specific condition of Vietnam. Chapter 8 – Railway noise emission and its reduction are considered among the most important topics in the future development of transportation systems, as they have a strong impact, which comes with a new line, on the people living nearby. Many elements contribute to the overall noise emission, but it has been evidenced that the most critical factor is the interaction between wheels and rails. This contribution is intended to give an overview of the noise emitted by wheels and rails from the basic emission mechanisms (rolling noise, squeal noise and impact noise) up to noise attenuation by means of passive/active control. To this purpose, the vibro-acoustic behaviour of different wheels is investigated both by means of experimental data and numerical simulations. Solid wheels, resilient wheels and wheels provided with constrained layers are deeply analysed to underline how different design strategies affect wheel emission under operating conditions. It is shown that a strict link exists between wheel mode shapes and the emitted noise. Particular attention is devoted to resilient wheels, as they are often adopted on city-tramcars, and thus their vibroacoustic behaviour has a strong impact as an annoyance to people. Many aspects are discussed, especially those that have not yet been deeply considered in the state of the art.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Preface

xi

An example is the effect of the tread thickness value of resilient wheels on the noise emission. This comparison among different wheels also shows that a particular design approach can reduce the noise emission due to a certain generation mechanism but, on the other hand, it has no effects on other kinds of noise. The role of rails is also considered, focusing on their influence on the wheel dynamic behaviour and so on their effect on the global noise emission. Data gathered during laboratory tests are integrated and compared to data measured in different situations on real railway lines throughout Italy. City-tram lines and high-speed trains have been considered. A deep comprehension of the vibro-acoustic behaviour of different wheels is the starting point towards any attempt to reduce wheel noise emission. The absence of a design approach that is effective for all of the main emission mechanisms makes the wheel-rail interaction a topic still open and where research is under continuous development. The final part of the contribution presents different possible approaches for noise mitigation and points to the new challenge: active noise reduction. The main difficulties associated with this strategy are discussed and some experimental results are presented too in order to show that, in some situations, technology is closing in on the possibility of finding new effective solutions for noise reduction. Chapter 9 – Safe automation and traffic control are two of the main R&D lines in the railways scope. New technologies in infrastructure and vehicles require more sophisticated railway signalling systems. Train wheel detectors or axle counters perform an essential safety function in the railway signalling system. These modules can detect the passing of a train over critical points, also known as detection points, along the railway network without using any vehicle-mounted device. The train wheel detectors are used to determine the train position and direction, to measure its speed, to confirm track circuits and treadles, among others. Almost all commercial electronic train wheel detectors use magnetic or inductive devices as physical sensors. Most of them are based on RLC circuits located next to the rail at the detection points, where the impedance of the sensor is directly related to the magnetic permeability of the physical medium. Other train wheel detectors use a pair of coils (located on either side of the rail) at each detection point, which work as emitter and receiver. The interruptions in the magnetic flux between the coils are used as a detection signal in these cases. Over the past few years, new solutions based on the use of coded signals and new sensor array implementations for magnetic or inductive train wheel detectors have been proposed. These novel detection systems allow to work with low signal-to-noise ratios (SNR) without the presence of signal processing electronic equipment near the track. Train wheel detectors based on non-magnetic principles, such as fiber optic or infrared (IR) sensors, have also been proposed and developed. These new technologies are aimed at ensuring reliability levels similar to those of magnetic or inductive detectors, while reducing maintenance tasks and costs. Future developments in the field should encourage new or improved solutions for critical aspects regarding railway safety applications, such as: multiple detection points, fault tolerance, robustness regarding external noise, interferences from other sources and attenuation of wires, and easy installation and maintenance tasks. Chapter 10 – This chapter discusses the importance of the vertical track stiffness as a means to guide high speed railway track bed design and inform decisions regarding track maintenance and renewal. To this end, a rational approach to substructure track-bed design is described which it is hoped will facilitate appropriate track design and enable the adaptation of existing design procedures to provide a realistic design for the conditions at hand. Despite adopting suitable design and construction standards, however, a number of factors may still

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

xii

Nicholas P. Scott

cause the track stiffness to vary spatially. Sometimes the stiffness variation may be very large within a short distance. These are likely to cause variations in the wheel/rail interaction force and will have a detrimental effect on track degradation increasing wear, fatigue of track components, and track settlement. This chapter discusses these issues and describes the use of numerical models to; assess the influence of track stiffness variations on the wheel/rail contact force, and; investigate the use of possible countermeasures. Techniques which have been developed to measure track stiffness, including the novel approach developed by one of the authors, are described, and the possible uses of such measurements to make appropriate and timely railway maintenance and renewal decisions are discussed. Devices from different countries are described in relation to the evolution from static to rolling devices capable of effecting measurements at train speeds. Such techniques have undergone a process of continual evolution in recent years and several are now ready to be developed from research tools to those capable of being used practically by commercial organisations. Chapter 11 – Ultrasonic systems are since long time used to measure stress on materials, utilizing the acoustoelastic effect. Research has also been carried out in railway engineering, basically to estimate stress on the rails. So far a widely used commercial system, based on ultrasonics or on other methods, does not exist. In this chapter a methodology for monitoring thermally induced loads on continuous welded rails is described, based on the use of ultrasonic waves. The technique allowed an estimation of rail’s neutral temperature and instantaneous longitudinal loads, by means of a new data elaboration method. A complete monitoring system was built and run for about two years on a 3-km track. The method proved to be expensive and time-consuming, if a large amount of railway track is to be monitored, because of the cost of the instrumentation and the necessity of rail adjustment. To find a solution to these problems, a new study was undertaken, aiming to design and test a portable ultrasonic device capable of measuring stress on rails. When using instruments of this kind, several problems must be faced, like those arising from the variation of contact characteristics between ultrasonic probes and rail surface. To solve these problems a double couple probe instrument was designed and built: it proved capable to allow for variation in couplant thickness and type, contact pressure and surface roughness. A calibration with respect to stress and temperature for some of the materials commonly used for rails was carried out. The influence of material’s internal structure and residual stress was also investigated. Expert Commentary – The French photographer Edmond Fortier, who lived and worked in Dakar (Senegal) from the 1900s, produced more than 3,000 pictures/postcards during his extensive trips in French West Africa. While his photographs are celebrated today in scholarly and commercial circles, his 1908 series of twenty-two Lagos pictures has not so far attracted the attention it deserves. Partly representing views from the Lagos steam tram and partly depicting the tramline as a background feature, this photographic series constitutes one of the rare pieces of contemporary evidence that relates to the tram. Operating between 1902 and 1933, the history of this line has hardly been researched as well. This commentary reflects on the historiography of these two relatively unknown episodes in colonial West Africa: that of Fortier’s visit in Lagos and the subsequent series of photographs; and that of the Lagos steam tramway which served as their background. Short Commentary A – Railway operators and suppliers can count on well established techniques for safety assurance. However, the recent terrorist strikes have shown that security is also an important issue to be addressed by system engineers. While safety refers to the possible hazardous effect of the system upon the external environment, security studies the

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Preface

xiii

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

effect of the external environment upon the system, including natural as well as malicious threats. Terrorists are not the only adversaries: vandals, thieves and other perpetrators use to attack mass transit rail based system on a daily rate. Procedures and protection mechanisms are therefore needed for the safeguard of the infrastructure against external threats. This paper describes a systemic approach to the security risk management of railway infrastructures. The analysis addresses both methodological and technological means, showing that while a correspondence can be found between safety and security taxonomies, novel approaches are needed to cope with security specific issues. Short Commentary B – Effective use of urban rapid railway systems requires that the railway systems be effectively connected with other transportation modes. Under such conditions, a higher level of accessibility can be achieved. This paper analyzes different attributes between arrival and departure access modes through a trip survey in Beijing. Using this data, this paper sets up the nested logit model access mode choice paradigm and analyzes the influence of three factors: access time, access cost, and access distance on the access mode choice. It is confirmed to be effective based on the combined estimated methodology. It is shown that these three factors are all negatively correlated with the access mode choice. It also appears more sensitive to the bus access fee. The conclusions provide an analysis tool for urban railway planning and construction.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

In: Railway Transportation Editor: Nicholas P. Scott

ISBN: 978-1-60692-863-9 © 2009 Nova Science Publishers, Inc.

Chapter 1

RAILWAY INFRASTRUCTURE MARKET IN EUROPE Christina Nikolova Department of Economics of Transport, University of National and World Economy Sofia, Bulgaria

Abstract The development of railway transport in Europe in the past 15 years is related to the process of restructuring. European Transport Policy and certain legislation impose vertical separation between the railway infrastructure and the transport services. This process has been accompanied by a number of problems linked to railway competitiveness in the transport market. The main development trends in the railway sector in the European Union are as follows:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

– – – –

ensuring the independence of the infrastructure enterprises and management of the transport operators; establishing administrative capacities for both activities; implementing the business-oriented structures for established companies; applying appropriate infrastructure charging regimes.

This chapter presents a brief review of the European Transport Policy in railway transport. It provides an outline of the development of the railway infrastructure market in Europe and examines the means for achieving the main goals concerning revitalization of the railways and especially for using the railway infrastructure in a more efficient way. The current state of restructuring in the different countries and the models for the development of the operators and infrastructure enterprises are described as well. The emphasis is laid on the importance of the measures for regulating the rail infrastructure as a natural monopoly. This chapter also takes into account the management and regulation of the railway infrastructure as key factors in the development of the rail infrastructure market. The latest scientific research regarding the importance of infrastructure charges in establishing competitive conditions in the railway market and in defining possible problems is presented. The current charging regimes applied throughout the EU member states are analyzed. An accent is placed on the caused external costs and possibilities for their internalization. A comparative analysis between railway and other modes of transport is made. From the point of view of the expected traffic growth in Europe, the chapter looks through the possibilities for better integration of the separate national and regional networks and the optimization of the use of existing capacity. It takes into account the opportunities for

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

2

Christina Nikolova linking up railways and other modes of transport through offering combined transport services and building up intermodal terminals. Finally, some implications regarding the sustainability of the railway transport are drawn up, as well as issues linked to certain economic, social and environmental aspects.

I. Introduction The transport market is usually described as a specific sphere of demand and supply of transport services and powers connected to them in order to make possible spatial movement of goods and passengers and satisfying the transport services demand (Mutafchiev, 1995); i.e., it is by nature an intermediate good. It is necessary to have equilibrium in the transport market as well as in the transport infrastructure market so that it is possible to reach common market equilibrium. The transport infrastructure market is a peculiar factor market, as it provides the required prerequisites and conditions for spatial movement of passengers and goods. In the context of the macroeconomic turnover (Brown & Jackson, 1998), the relations among the participants of those two kinds of market are shown in figure 1. Furthermore, the starting point for the analysis is the general equilibrium model (Quinet, 2005). The figure presents a simple three-sector model, which includes users of transport services, transport companies as well as the government (state) as main economic subjects that operate together, striving to achieve common market equilibrium, and among whom circulate the main finance flows in the transport sector. –

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

–

–

–

Users of transport services: potential passengers and shippers shape the transport demand—they incur or are willing to make expenditures to meet their needs of transport services; Transport companies offer, on one hand, transport services, while on the other hand seek access to the given infrastructure, for which they pay the infrastructure companies; Infrastructure companies, in connection with the aforementioned, receive income by way of infrastructure charges and at the same time are financed by the state on account of the ownership of the infrastructure, which is usually public; The government (state) accumulates the required budget funds through levying taxes on income, properties and usage of transport services and charges on the transport infrastructure. On the other hand, it finances the activities of managing infrastructure objects, which are explicitly state property, and respectively assigns state orders for transport services to some of the carriers (usually for passenger services and urban transport) and offers transfer payments for the citizens and subsidies to some of the transport operators. The state regulates transport market and infrastructure functioning through price regulation (transport prices or infrastructure charges), applying regulatory legislation to each mode of transport and creating regulatory bodies that pursue this policy.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Regulatory Body Regulation

Transfers Taxes

Infrastructure Managers Fundings

Transport Infrastructure Supply

Taxes

Transport services’ Supply

Transport Infrastructure Demand

Payments

Taxes

Subsidies

Revenues

Users

Government Regulation

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest Ebook

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Transport Services’ Demand

Transport Operators Figure 1. Interrelations among the participants in transport and infrastructure markets

4

Christina Nikolova

By virtue of that model, the place and the role of every economic subject in the transport sector in circulation flows of income and services could be defined. Furthermore, we could define accurately the functioning of the transport infrastructure market. Similarly, this model precisely presents the conception related to the necessity of transport infrastructure supply as a public good, which contributes to transport services’ performance. Considering this method, we enter inevitably into the continuation of the public choice theory in the field of transport economics. To understand the role and significance of the railways in the EU requires in-depth study of their current state, of the development and utilization of the rail infrastructure, the approaches for operators’ and infrastructure management and the structures and elements of the infrastructure charging system in the railway subsector. In that case—defining the infrastructure as a common good—requires answering the following questions: • • • • • •

What is the level of infrastructure capacity demand? In what way its supply is organized? To what extent is the allocation of infrastructure capacity effective? What is the mechanism in providing the required funds for building, maintenance, repair and exploitation of the railways infrastructure? Are these funds utilized effectively? In what way does the distribution of the accumulated funds and infrastructure capacities influence the common wealth as whole?

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The answer to each of these questions requires in-depth study of the status, development and utilization of the railway infrastructure, the approaches for its management and the structures and elements of the infrastructure charging system.

II. The Current State of Railway Transport in Europe During the last decade, the total length of the rail infrastructures in the European Union experienced a slight decrease. The indices for railway infrastructure providing on the territory of the EU countries (see Table 1 below), reflect its density based on territoriality. The insignificant reductions in the values of these indices during the reviewed period (1990–2005) are based on reduction of rail tracks in some sections in the certain countries, categorized as low-performance routes due to the low volume of demand and accumulation of losses, which makes it impossible for them to be properly maintained. The length of the rail tracks in the EU-27 is 231,582 km, of which 49.8% (107,373) are electrified. The data regarding the rail lines in use per 1000 citizens demonstrate a slight decline (about 5% average). The reason is, of course, the slight decrease in the length of rail lines as in the same time the number of population increases.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

5

Table 1. Infrastructure Providing Indices Length of lines in use Electrified lines Territory Population Network density

Years 1990 2005 2005 %

Measure km km

EU 25

EU 15

EU 12

231582

215930

162132

69450

215439

200337

153513

61926

107373

100526

80249

27124

49.8

50.18

52.3

43.8

4292.99

3238.99

2944.96

2005 thousand square km 4641.99 1990 2005 1990 2005

million citizens km per square km

Rail lines in use per 1000 citizens

1990 km per 1000 2005 citizens

Number of locomotives and railcars Number of passenger coaches, railcars and trailers Number of freight wagons

EU 27

470.39

436.47

363,49

106,90

489.21

457.03

385.69

103.53

49,89

50.30

50.06

23.58

46.41

46.67

47,40

21,03

0.49

0.49

0.45

0.65

0,44

0,44

0,40

0.60

1990 number 2005

65981 46808

59503 43953

43710 33062

22271 13746

1990

114355

105617

78687

76331

90621

85753

66997

64727

825398 403630

808022 395717

2005

number

1990 number 2005

1457046 1248501 686024 604338

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Legend: European Union – 27 member states (EU 27), European Union – 25 member states (EU 25), European Union – 15 member states (EU 15), European Union – 12 member states (EU 12). Source: EUROSTAT (2007). Directorate-General for Energy and Transport, EU Energy and Transport in Figures

Country codes: Belgium (BE), Bulgaria (BG), the Czech Republic (CZ), Denmark (DK), Germany (DE), Estonia (EE), Ireland (IE), Greece (EL), Spain (ES), France (FR), Italy (IT), Cyprus (CY), Latvia (LV), Lithuania (LT), Luxembourg (LU), Hungary (HU), Malta (MT), the Netherlands (NL), Austria (AT), Poland (PL), Portugal (PT), Romania (RO), Slovenia (SI), Slovakia (SK), Finland (FI), Sweden (SE) and the United Kingdom (UK). Source: EUROSTAT (2007). Directorate-General for Energy and Transport, EU Energy and Transport in Figures

Figure 2. Network density in individual EU member states

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

6

Christina Nikolova

In terms of network density things look the same. The decrease is almost 7% per whole union. The Czech Republic and Belgium have the highest rail network density. The lowest density within the EU-27 can be found in Finland and Greece (see figure 2). Regarding the structure of the rolling stock, as we could see from the figure 3 below, the railway companies decrease the number of their rolling stock average by 50% between 1990 and 2005 but the percentage defers from country to country.

Source: EUROSTAT (2007). Directorate-General for Energy and Transport, EU Energy and Transport in Figures

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 3. Number of locomotives and railcars

This is as a result of the decrease in the demand for railway haulage and the necessity of optimization and modernization of the vehicles. In view of the age structure of the inventory, it should be noted, that there is a positive trend towards better degree of renewal and maintenance. This is an important measure for improving the quality of tendered services along with enhancing infrastructure capacity. A greater mutual dependence of the world’s and European markets as well as the increase in the international trade, as one of the basic goals of the European transport policy, emphasizes the inner transport connections as well as the connections with other countries and requires a modern transport network. The traffic policy programme of the European Union, especially its part that relates to the development of the international relations puts the railway infrastructure into the first plan. Therefore, a special attention needs to be directed to the Pan-European corridors concept. The railway routes of the Pan-European transport corridors include over 25200 km railway network. Through the territory of the EU these are as follows: Corridor I with total railway length of 1710 km: Helsinki – Tallinn – Riga – Kaunas – Warsaw with the components: - Rail Corridor (Rail Baltica) from Tallinn – Riga – Warsaw - Branch (road/rail) from Riga – Kaliningrad – Gdansk Corridor II – 2300 km: Road and rail link connecting Berlin – Warsaw – Minsk – Moscow –Nizhny Novgorod

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

7

Corridor III – 1650 km: Road and rail connection between Dresden – Wroclaw – L’viv – Kiev Corridor IV – 4340 km: Road and rail connection between Dresden– Prague – Vienna – Bratislava – Budapest Branches to Nuremberg, Bucarest – Constanta and Sofia – Thessaloniki / Istanbul Corridor V – 3270 km : Road and rail connection between Venice – Trieste – Koper – Ljubljana – Budapest – Uzgorod – L’viv Branch a: Bratislava – Kosice – (Uzhgorod) – L’viv Branch b: (railway): Rijeka – Zagreb – Koprivnica – Dombovar Branch c: Ploce – Mostar – Sarajevo – Osijek – Budapest Corridor VI – 1800 km: Road and rail connection between Gdansk – Grudziadz/Warsaw – Katowice – Zilina Branch to Brno Corridor VII - 2411 km: The Danube waterway Corridor VIII – 1270 km: Road and rail connection between Bari and Brindisi – Durres and Vlore – Tirana – Skopje – Sofia – Varna and Burgas Branch 1: Cafasan – Kaphstice/Kristallopigi Branch 2: Sofia – Pleven – Byala (road)/Gorna Oriahovica (rail) Branch 3: Burgas – Svilengrad – Ormenion

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Corridor IX – 6500 km: Road and rail connection between Helsinki – St. Petersburg – Pskov/Moscow – Kiev – Ljubasevka – Chisinau – Bucarest – Dimitrovgrad – Alexandroupolis Corridor X-2360 km: Road and rail connection between Salzburg – Ljubljana – Zagreb – Beograd – Nis – Skopje – Veles – Thessaloniki Branches to Graz, Budapest, Sofia and Florina The ten Pan-European transport corridors were defined at the second Pan-European transport Conference in Crete, March 1994, as routes in Central and Eastern Europe that required major investment over the next ten to fifteen years. Additions were made at the third conference in Helsinki in 1997. A tenth corridor was proposed after the end of hostilities between the states of the former Yugoslavia. These development corridors are distinct from the Trans-European transport networks, which include all major established routes in the European Union, although there are proposals to combine the two systems. The Trans-European Transport Network Concept is developed in order to help the EU countries to set up the foundations for their future infrastructure that would stimulate trade between members, expand commodity flows, enable easier flow of traffic means and enhance social relationships (Jugovic, 2006). The Trans-European transportation network consists of the following components: the Pan-European transportation corridors, located on the territory of the new independent states, joined members of the EU or potential candidates for the EU membership; the TINA network (Transport Infrastructure Need Assessment)

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

8

Christina Nikolova

includes: the Pan-European transport corridors and additional infrastructure components in the states being the potential candidates for the EU membership; the four Pan-European transportation areas (Pan-European Transport Areas-PATRAS) that cover the maritime traffic; the Eurasian connection known as TRACECA (Transport Corridor Europe-CaucasusAsia). Along with this concept a special attention should be placed on the development of railway traffic. Since 1995, the railways’ share of the EU market for both freight and passenger transport has increased slightly. Considering only inland transport, it appears that the considerable growth in transport has been realized mainly by road transport. In 2006, it represented around 75% of the tonne kilometers (tkm) performed in the EU (see Table 2). Table 2. Inland transport Indices Freight transport by road Freight transport by rail Passenger car transport Bus transport of passengers Rail transport of passengers

Years 1995 2006 1995 2006 1995 2006 1995 2006 1995 2006

measure thousand mio tkm thousand mio tkm thousand mio pkm thousand mio pkm thousand mio pkm

EU 27 1288.7 1887.6 386 434.6 3854.7 4601,7 500.6 522.5 347.9 384

EU 25 1263.76 1816.22 359.52 413.42 3800.7 4510,7 472.6 499.7 324.3 373.5

EU 15 1138.2 1537.2 222.61 280.37 3520.9 4052,2 377.1 416.9 273,6 334.1

EU 12 150.4 350.4 163.4 154.2 333.8 549.5 123.7 105.6 74.4 49.9

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Source: EUROSTAT, (2007). Directorate-General for Energy and Transport, EU Energy and Transport in Figures

The increase between 1995 and 2006 was very high in Germany, Spain, France, Italy, Poland and Romania. No country had a decrease during the same reference period and the growth of the total tonne kilometers transported by road in all the EU countries was equal to 40% (see Figure 4).

Source: EUROSTAT (2007). Directorate-General for Energy and Transport, EU Energy and Transport in Figures

Figure 4. Freight transport by road

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

9

In the freight market the decline in share of the railways has been combined with an absolute decline in transport volumes. There is an increase only in Germany and UK, and in some of the new member states such as Estonia, Latvia and Lithuania. It shows that the railways’ share of the EU freight transport market increased during the last years, but not at all in the same pace as road transport (see figure 5 below).

Source: EUROSTAT, (2007).Directorate-General for Energy and Transport, EU Energy and Transport in Figures

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5. Freight transport by rail

The conclusion is that the number of tonne kilometers performed by road is much greater than those performed by rail. This increase is there in spite of a decreasing network and less rolling stock, which must indicate that the efficiency of the rail industry today is higher. Regarding the passenger carriages, the increased mobility demand has mainly been satisfied by passenger cars, performing roughly three-quarters of all trips (see figure 6 below). In the middle of revised period between 7,000 and 13,000 passenger kilometers (pkm) where performed pro person in the EU. The use of a car offers a high degree of independence and flexibility. On the other hand, we must not forget the negative impact on the environment and the number of killed each year in road accidents (around 40,000).

Source: EUROSTAT, (2007). Directorate-General for Energy and Transport, EU Energy and Transport in Figures

Figure 6. Passenger car transport

Between 1995 and 2006, the average transport by buses and coaches in EU-15 has been increasing by nearly 10%, arriving at a total of more than 416 billion passenger kilometers

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

10

Christina Nikolova

(see figure 6). With above 102 billion passenger kilometers, Italy has the highest figure in the EU in absolute terms this corresponds to roughly 5 km per person per day. It is however the Danish, Luxembourg’s and Greek population that travels the most on bus and coach per person in the EU-15 with between 5.5 and 6 km per person and day.

Source: EUROSTAT (2007). Directorate-General for Energy and Transport, EU Energy and Transport in Figures

Figure 7. Bus transport of passengers

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Compared to the other modes, transport performances of rail experienced only a modest increase for the EU as whole. Since the early 1990s, growth has been slow in most countries and a certain increase in some countries can be observed (i.e., in Germany, France, Italy, Sweden, UK). Still the EU average of kilometers travelled per person and day is above 2. This increase is there in spite of a decreasing network and less rolling stock, which must indicate that the efficiency is higher. In the EU passenger transport market the railways have, over the period 1995 to 2005, lost market share to other modes but increased total passenger kilometers by 10.4 per cent (see figure 8).

Source: EUROSTAT (2007). Directorate-General for Energy and Transport, EU Energy and Transport in Figures

Figure 8. Rail transport of passengers

The significant increase in overall traffic has been accompanied by a parallel increase in mobility and has also led to the deterioration of the environment through pollution, congestion, noise and land-take for transport infrastructure.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Railway Infrastructure Market in Europe

11

Over recent years, political attention in Europe has been focused on the more negative interpretation of the above analysis—specifically it has focused on the fact that whilst the overall transport market has grown strongly the railways have performed less well, and lost modal share with adverse financial consequences. Many factors lie behind this trend, some of which are outside the direct control of the railways such as government policies that have favored the development of road transport. Other factors however originate within the rail sector. European Conference of Ministers of Transport (ECMT) Round Table 43 suggested that the railways of the EU increasingly faced a lack of correspondence between the organization of their train services and the rapidly changing patterns of transport demand. In particular, freight services no longer corresponded to the demand for transport at short-notice in small loads arising from ‘Just-in-Time’ industrial production methods. Many see the situation for freight haulage as particularly problematic and want to see the railways increase their market share in order to reduce congestion and the negative impact of the rapidly growing number of trucks on the roads of Europe. Such proposals tend to reflect real concerns that rail is losing market share because its quality of service is in some cases poor and its labor productivity relatively low compared to other modes. Whilst many countries have pursued their own rail reform agendas driven by national priorities, the European Union has played a leading role in promoting railway reform. The EU has targeted international rail freight for priority in reform, with the prime reason put forward for poor service and low productivity being the lack of competition in train operations and the monopoly powers enjoyed by national railways. The approach adopted in Europe, by the ECMT and EU, to improving the efficiency and competitiveness of railway undertakings has centered on promoting commercial freedom. A series of EU Directives have been issued since 1991, based on the premise that the railways are a vital part of the transport system and crucial to achieving greater integration of the EU’s transport sector as an essential element in the creation of the internal market. Their provisions are intended to improve the efficiency of the railway system by promoting a competitive market, whilst taking account of the special features of the railways. The specific policies have included: providing a clear definition and separation of the roles of railway undertakings vis-à-vis the state; improvements in the financing of railways through greater transparency; and the progressive opening of railway infrastructure for specified services, through the provision of non-discriminatory rights of access. These developments have been facilitated through the separation of accounts for railway infrastructure and train services, and the isolation of non-commercial debts. Common rules have been applied to debt restructuring in the EU. The railways of European countries are experiencing a fundamental overhaul the result of which has not been seen yet. They are being transformed from monopolistic state-run organizations into commercially-led enterprises. Policymakers, however, recognizing the technological development achieved by railways as well as their operational efficiency over the years, decided that these traffic trends should not mask the inherent advantages possessed by the railways. Railway transport is an excellent substitute for road and air transport, in medium distance inter-city travel and helps to reduce congestion for the benefit of other traffic, such as short distance road transport and intra-European air traffic. To exploit these advantages and to ensure an important place for rail services in transport markets on a sustainable basis, the EU and non-EU states have promoted policies designed to render their railways more efficient and competitive by promoting commercial freedom.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

12

Christina Nikolova

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

III. Combined Transport In its White Paper (Commission of European Communities, 2001), the European Commission spoke of an increase in intra-European Transport by all modes around 38% in ten years time. The forecast is toward a growth in the rail freight’s market share of around 8 to 15% to the year 2020. To meet this challenge, a large number of railway companies has adopted an aggressive strategy where intermodal transport plays an important role. Contrary to the general trend of rail freight, rail-road intermodal transport more than doubled between 1986 and 2000 - attaining 180 million tons. Therefore the expectations are towards the railways to play a significant part in the modal shift needed to sustain the mobility, the environment and the competitiveness of the European economy. In order to absorb the growth announced and for railway companies to be able to propose adapted and competitive products it has become of vital importance that the sufficient infrastructure capacity is necessary. In terms of investment and to guarantee the railways’ share of the intermodal business on the EU level the measures to be taken by transport industry players (government authorities, railway companies, operators, infrastructure management) has been identified. In the current context of infrastructure saturation, in order to realize the modal shift towards rail advocated by the EU in its White Paper, several measures need to be taken. These measures range from investments in rail and terminal infrastructure, technical-operational improvements, to the fostering of the working procedures of all the stakeholders in combined transport rail-road. Assessing the major importance of combined transport in 1970 in Munich was founded the International Union for Combined Road-Rail Transport Companies (UIRR). The first member companies were ASG (S), Hucketrans (A), Hupac (CH), Kombiverkehr (D), Novatrans (F), Trailstar (NL) and TRW (B). The members organized rail transport using swap bodies and semi-trailers: at that time national traffic represented around 230,000 shipments, and there were only 17,000 international shipments. After having been formed as a de facto association in 1970, in April 1991 the UIRR was transformed into a co-operative limited company under Belgian law. All the tasks carried out by the UIRR are geared towards the development of road-rail combined transport in Europe at the same time as defense of the interests of its member companies, which in their respective markets take care of the organization and marketing of this intelligent system of routing goods. The exclusive mission of the Union is the promotion, by every possible means, of combined transport, mainly the combination between the road and rail modes. Today the UIRR consists of 20 member companies in 15 European countries. The share of these companies in the total volume of goods delivered in view of a combined transport type forwarding amounts to around 70%. The members are subdivided into two major groups (UIRR, 2008): 1). The operators offering a terminal-to-terminal service: to do this they acquire the necessary rail traction from the rail undertakings and, usually through the intervention of the latter, access to the rail infrastructure; when this is combined with transshipment operations in the terminals and the required transport documents, they are able to supply an integrated service to their clients—road haulers, freight forwarders and logistics companies—who themselves ensure transport of the loading unit to the point of departure and/or its collection at the point of arrival. This type is the most widespread among the UIRR operator members,

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

13

who express in this manner their concern of not wishing to compete with their road transport shareholders. 2) The operators offering the complete chain of transport from door to door, i.e., from the shipper to the final consignee: they thus also take care of the initial and/or final leg by road to the transshipment yards.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Table 3. Combined transport traffic by company Company 2000 International Traffic Consignments (thousand) Kombiverkehr (DE) 374.8 Hupac (CH) 231.1 Cemat (IT) 149.4 ICA (AT) n/a Okombi (AT) 148.8 T.R.W. (BE) 57.7 Ralpin (CN) n/a Hupac NV (NL) 26.4 Adria Combi (SL) 19.2 Conliner (DE) n/a Novatrans (FR) 77.3 Combiberia (ES) n/a Hungarocombi (HU) 61.8 Alpe Adria (IT) n/a Total 1146.5 National Traffic Consignments (thousand) Kombiverkehr (DE) 163.3 Cemat (IT) 194.3 Novatrans (FR) 176.1 Okombi (AT) 117.1 Naviland Cargo (FR) n/a ICA (AT) n/a Hupac (CH/DE/IT)) 16.2 Alpe Adria (IT) n/a Adria Combi (SL) n/a Total 667 Traffic of UIRR companies in 32.5 billion tonnes kilometers of which national 8.2 Traffic % of consignments Semitrailers 9 Rolling Road 23 Swap Bodies 68 Traffic of Intercontainer 961.7 Interfrogo - CT (thousand TEU)

2006

Change, %

375.8 346.1 266.5 154.8 117.5 98.5 80.9 57.1 53 45.3 43.7 29.5 27.3 27 1723.0

0.27 49.76 78.38 n/a -21.03 70.71 n/a 116.29 176.04 n/a -43.47 n/a -55.83 n/a 50.28

285.6 173.4 153.6 88.2 68.8 41.6 41 22.5 15.9 890.6

74.89 -10.76 -12.78 -24.68 n/a n/a 153.09 n/a n/a 33.52

45.4

39.69

9.8

19.51

9 16 76

0,00 -30.43 11.76

498.8

-48.13

Source: EUROSTAT (2007). Directorate-General for Energy and Transport, EU Energy and Transport in Figures Note: Consignment = an average road transport (2.3 twenty-foot equivalent unit (TEU)). TEU correspond to 10–12 tonnes.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

14

Christina Nikolova

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The analysis of current international combined transport (CT) traffic performed by the International Union of Combined Road-Rail Transport Companies shows that traffic volumes totaled 1146.5 thousand consignments in 2000 and increased to 1723 thousand consignments in 2006 (see table 3). There is a growth of 50.28% on account mainly of the increase in consignments carried by Hupac NV (NL), Adria Combi (SL) and Cemat (IT). The national traffic performed by the UIRR companies increased over the reviewed period from 667 thousand consignments to 890.6 thousand e.g. there exists 33.52% growth. The increase in the consignment carried by Hupac in Switzerland, Germany and Italy is at the root of this growth. The total traffic in tonne kilometers has increased by 39.69%, of which national traffic has increased by 19.51%. The figures show that the most important rule the combined transport has is at international carriages of goods. The most prevalent form of these carriages is the transport of single containers, swap bodies and semi-trailers, that is to say unaccompanied. For the UIRR companies, this represents about 80% of their traffic. A characteristic element of the unaccompanied CT is that the loading units are usually loaded vertically between the different modes of transport. This loading is carried out in the combined transport terminals by means of gantry cranes or mobile transshipment equipment. The transport undertaking or logistics company transports the intermodal shipping unit to the departure terminal and is responsible for it being collected at the destination terminal, from where it is then taken to its destination. Unaccompanied transport is more economic due to its better payload/deadweight ratio. The average distance covered by an unaccompanied combined transport consignment in Europe is between 700 and 800 kilometers. In principle CT can transport just about any type of goods that can be carried purely by road. Instead of fixed superstructures, the road vehicles must be equipped for this purpose with detachable superstructures. The following loading units are most often used in road-rail CT: • • •

standardized swap bodies, classes A (principally 13.6m) and C (7.15m, 7.45m or 7.82m) ISO containers of 20' (6.1m), 30' (9.15m) and 40' (12.2m) cranable semi-trailers (13.6m)

In unaccompanied CT, it is necessary to make the distinction between continental transport and maritime transport for the hinterland. For continental transport, it is a matter of connecting the important European economic regions by means of a network of rail links between terminals. All types of loading units are used in this case, in particular swap bodies in conformity with the European Committee for Standardization (CEN standards). In the framework of maritime transport it comes to providing high-quality road-rail CT services for all the large European ports by means of the transport of containers according to ISO standards between the large European ports and the continental hinterland. Almost 60% of the total European unaccompanied CT was generated by continental services and 40% by the hinterland transport of maritime containers. Given that, it is striking that in services between Central and Eastern European countries and the EU-15 member states maritime containers made up about 80% of the total volume, while continental shipments reached 20%.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

15

Accompanied Combined Transport is a convenient and rapid form of CT which consists of transporting whole trucks, road trains or articulated vehicles on special wagons. In this type of transport, the drivers of the road vehicles carry out the loading (called “horizontal loading”) themselves and accompany the railway shipment in a couchette carriage in order to effect delivery by road at the final destination. In accompanied combined transport, only a part of the total journey of the road vehicle is carried out by rail. Before and after being loaded onto the wagon, the vehicle is driven on the road. In general, rail transport allows avoidance of a geographical obstacle or of a route section involving weight or access restrictions. The distance covered by rail depends on the length of the “obstacles” on the road and on the required statutory night rests. In this manner the driver can rest during a section of the route or during the crossing of the Alps, for instance, and on arrival he can continue his trip completely refreshed. The main advantage of the rolling road is that any type of road vehicle can be transported by rail without any prior conditions. For the UIRR companies, accompanied CT represents about 23% of their traffic in 2000 and decrease to 16% in 2006. The main drawback is the importance of the deadweight because, besides the load, the whole truck must be carried by rail. In certain countries of the European Union, particularly in southern Europe and Great Britain, the railway gauge is not sufficient to transport the 4m-high trucks on rolling road wagons. In its final report prepared for International Union of Railways, the Combined Transport Group presents a forecast for international combined transport by 2015 (Kessel & Partner, 2004). According to this forecast, international combined transport on the Trans-European Network will increase from 54.5 million tonnes in 2002 to 116.0 million tonnes in 2015 (see table 4 below). Table 4. International combined transport 2002–2015

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Market segment Unaccompanied Accompanied Total

TEU (mio) 2002 2015 3.48 8.7 1.26 1.5 4.74 10.2

Net tonnage (mio tonnes) 2002 2015 Change, % 44.1 103.6 135 10.4 12.4 19 54.5 116 113

Source: Kessel & Partner (2004). Study on Infrastructure Capacity Reserves for Combined Transport by 2015. UIC - CTG

The forecast assumes that such a political framework is due to change and that both subsidies for rolling highway services and quota restrictions on road transport will be significantly reduced or eliminated. On the other hand more qualitative controls of road vehicles and a comprehensive road toll scheme will be enforced. According to above mentioned expertise this will lead to a considerable cut down of the number of accompanied CT services, which provide for the following features: •

Focus on high-frequency services, calculated as one departure every three hours, seven days both ways. • Services, which bring value to road operators, e.g., compliance with driving hours. However, international accompanied CT has a chance to survive. It could even grow to a volume of 652,000 trucks carrying 12.4 million tonnes, which is +19% compared to 2002.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

16

Christina Nikolova

International unaccompanied CT is expected to be the more dynamic market segment, with an increase by 2015 to almost 9 million TEU with a net load of 103.6 million tonnes. This corresponds to an average annual growth rate of 6.8%. Compared to 2002 international unaccompanied CT will have more than doubled by 2015. An increase of +135% over 13 years is not as extraordinary as it might appear at first glance especially if one looks at this forecast in the light of the results of the 2000-2006 period. Market experts concluded that CT was likely to witness a stronger growth than road due to major enhancements in rail and intermodal transport such as improved quality, efficiency, and interoperability, and, on the other hand, increased controls of road vehicles and charging of road infrastructure usage. In a second step specific aspects of the freight corridors were evaluated in terms of their likelihood to promote or impede CT development (transport policy, topography, etc.). The experts took into account recent research on the transportrelated effects of the EU enlargement. The report consist also an evaluation of rail network capacity by 2015. The determination of network capacities on a European level is done by using standardized capacity limits. From 1 different studies , a capacity limit is estimated on 144 movements (passenger and freight) per day and direction on a double tracked electrified line. This figure is based on movements and not on (theoretical) train paths. For the forecast horizon 2015 a growth of the average network capacity of 20% has been assumed. This is due to shorter block distances, improved operational/signaling systems and bi-directional traffic. This assumption, which is relatively optimistic, leads to the capacity limits of 173 trains per day and direction. With regard to the expected enhancement investments the funds needed to develop further the trans-European Network exceed EUR 110 billion for the major priority projects alone. This means that some projects had to be selected ahead of others. In the investment process public funding (Community funding, e.g., Structural Funds, Cohesion Fund and budget for trans-European network) is the main source so there is a necessity of greater involvement of private capital in infrastructure funding through public-private partnerships and contracts. The study shows that even if all planned infrastructure investments are realized by 2015, considerable bottlenecks (lack of capacity for operating daily trains) would remain. This would be exacerbated if capacity enhancements programmes regarding train and line capacity parameters were not achieved. In that case network bottlenecks would increase further. In conclusion, it becomes apparent that considerable efforts will be required until 2015 to cope with the increase in transport volumes. The capacity assessment of intermodal terminals has analyzed 34 transport areas on the 18 trans-European corridors, representative of the network of terminals for unaccompanied combined transport services. These areas cover 70 individual terminal sites representing some 85% of the total 2015 volume of international unaccompanied combined transport. The total transshipment volumes in these 34 transport areas is forecast to increase by 80% from 6.3 million intermodal load units (2002) to 11.4 million units (2015). Investigations into enlargement programmes proved that a large scope of investments is scheduled or already in progress, both extending existing sites and building new terminal sites. According to that, the nominal total transshipment capacity is due to rise from 9.6 million units (2002) by 39% to 13.3 million load units. Despite these ambitious enlargement programmes, capacity gaps are 1

Studies carried out by DB AG (DE), RFF (FR) and SNCF (FR).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

17

likely to arise in 20 out of 34 transport areas by 2015. As a consequence, on top of the investments scheduled another 13% of transshipment capacity enabling to handle 1.7 million units p.a. is required to meet the increasing demand for unaccompanied CT services, and to maintain a high quality of service towards intermodal customers. The study into international combined rail-road transport carried out from Combined Transport Group for International Union of Railways shows that combined transport market segment is likely to expand over the 2002–2015 period. In order for rail to be able to absorb the forecasted growth for international CT, increased capacity is required both in terms of the rail network and intermodal terminals. It is crucial that enlargement investment is taking place on time to avoid temporary capacity shortages: calculate sufficient time for planning, approval procedures and financing, construction and opening of the enlarged terminals and their access infrastructure. Being the interface between road and rail, the terminal is the most crucial part of the CT supply chain. Sufficient handling capacity is thus a prerequisite for ensuring high performance: allow capacity reserves to prevent the terminal from becoming the bottleneck. The most decisive factor is having qualified terminal management and staff. The “human factor” is probably the most important driver for an efficient use of infrastructure. Actions to optimize capacity utilization on intermodal terminals, e.g., by enhancements of process organization and operations (clear definition of roles and interfaces) supported by an IT terminal management system Combined transport is recognized as being the most dynamic market for the transport of goods in Europe, which will most surely enable the railways to participate in the growth of transport requirements essentially resulting from growing economic welfare and EU enlargement. Moreover, in comparison to road and maritime routings, road-rail combined transport makes it possible to limit the emission of pollutants and energy consumption.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The projects run on European level regarding intermodal transport are as follows: -

-

1996 CESAR Co-operative European System for Advanced Information Redistribution; 1999 QUALITY Developing a Quality Strategy for Combined Transport; 2000 EUTP Thematic Network on Freight Transfer Points and Terminals; 2001 CO2 CO2 Reduction through Combined Transport; 2004 BRAVO Brenner Rail Freight Action Strategy Aimed at Achieving a Sustainable Increase of Intermodal Transport Volume by Enhancing Quality, Efficiency and System Technologies; 2005 INSECTT Intermodal Security for Combined Transport Terminals; 2005 SINGER Slovenian INtermodal Gateway to European Rail; 2005 TREND Towards New Rail Freight Quality and Concepts in the European Network in Respect to Market Demand; 2006 COUNTERACT Cluster of User Networks in Transport and Energy Relating to Anti-terrorist ACTivities; 2006 DIOMIS Developing Infrastructure Use and Operating Models for Intermodal Shift.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

18

Christina Nikolova The ongoing projects are: -

2008 RoMo-Net Rolling Motorways Network Through Eastern Europe; 2008 SEEIS South East European Intermodal Services.

From the point of view of transport and environmental policy, the development of combined transport represents one of the main thrusts of the EU’s and its Member States’ strategy. Indeed, this transport system benefits from strong and deserved support at European level which consists of various promotional measures such as the elaboration and preservation of framework conditions to ensure it fair access to the transport market.

IV. European Legislation Review The first piece of major legislation goes back to 1991 and the adoption of directive 91/440/EEC on The Development of the Community’s Railways, by the Council of Ministers. This introduced a degree of liberalization into certain areas of rail transport, above all prompting the railways to concentrate more on competitiveness. Directive 91/440, required each national railway undertaking within the European Union to be established as an independent body, run on commercial management principles. It also enforced the financial restructuring of railway undertakings, to provide separate accounts for infrastructure management and rail operations (ESCAP, 2003). The directive required Member States of the EU:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

•

• •

•

to manage railway undertakings in such a way that these understand the need for competitiveness and sound financial management. Member States must thus, jointly with existing public railway operators, take steps to reduce the indebtedness of railway undertakings; to make railway undertakings independent by giving them a budget and system of accounts which are separate from those of the State; on specific terms, to guarantee rights of access for rail transport operators in Member States to international combined transport services. The aim here is to open up the Community markets in these sectors. It has also created the possibility to open the market for international freight and passenger services under certain conditions; and to have separate accounting for railway infrastructure (track and related equipment) and the operation of transport services. The aim here is greater transparency in the use of public funds, but also the ability to measure the actual performance of these two activities in a better way. It is with this requirement in mind that a number of Member States have in recent years set up bodies which manage the railway infrastructure but are separate from the railway companies, which continue to manage the carriages of passengers and freight.

Linked to this directive was Regulation 91/1893/EEC concerning the obligation inherent in the concept of public service in transport. This directive required that any public service obligations should normally be provided for in specific contracts. Directive 91/440 offered national railways open access rights to the main-line infrastructure of other EU Member

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Railway Infrastructure Market in Europe

19

States, but only for co-operative ventures involving train operators in the Member States at either end of a route or for combined transport freight operations. This Directive has not proved effective in encouraging new international rail services; rather, it has defined the minimum requirements for the financial and organizational restructuring of national railway companies. Directive 95/18 required national railway undertakings to hold a current operating license. It defined the criteria governing the award, retention, and international validity of such licenses, with the goal of ensuring consistent and nondiscriminatory conditions for market entry - particularly for companies seeking to exercise international open access rights. Directive 95/19 defined the basic principles and procedures governing the allocation of main-line infrastructure capacity between alternative users, including open access operators, in order to facilitate the development of new services. It also specified criteria for setting rail infrastructure charges, in the continuing absence of a coherent, multi-modal framework of user charges. Directives 95/18/EC and 95/19/EC strengthened the requirement for member states to permit competition amongst train service operators by promoting the allocation of railway capacity and the development of non-discriminatory charges for access to the infrastructure. The problem has been that their provisions are limited to international operations qualifying for the very limited access rights set out in Directive 91/440. On 26 February 2001, the Council of European Community adopted three directives to allow a further opening of the European market for international freight transport by rail (Directorate-General for Energy and Transport, 2008). These directives are also known as Rail Infrastructure Package or First Railway Package Directives (Directives 2001/122, 2001/133 and 2001/144). These measures aim at increasing international rail transport of freight by setting clear rules on the conditions under which railway undertakings can obtain licenses and safety certificates; the framework for the allocation and charging for the use of rail infrastructure capacity; the role and responsibility of Regulatory Bodies in the Member States and the separation of accounts for subsidized and non-subsidized activities. The package envisaged also monitoring the railway markets, which could provide an essential input to the Commission's report on the impact of the implementation in the Member States of the Infrastructure Package. On 1 January 2006 it provided the opening of the market for international freight transport by rail over the entire rail network in the enlarged European Union: any licensed railway undertaking having a safety certificate for the networks on which it wants to run transport operations can apply for capacity and offer international transport services by rail throughout the EU. On 18 October 2007 the European Commission adopted a Communication to the Council and the European Parliament on monitoring development of the rail market COM (2007) 609. The Communication focuses on:

2

Directive 2001/12/EC of the European Parliament and of the Council of 26 February 2001 amending Council Directive 91/440/EEC on the development of the Community's railways, Official Journal L 075 , 15/03/2001 P. 01 – 25. 3 Directive 2001/13/EC amending Directive 95/18/EC, Official Journal L 075 , 15/03/2001 p.26-28. 4 Directive 2001/14/EC of the European Parliament and of the Council of 26 February 2001 on the allocation of railway infrastructure capacity and the levying of charges for the use of railway infrastructure and safety certification, Official Journal L 075 , 15/03/2001 p.29-46

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

20

Christina Nikolova •

•

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

•

the regulatory and institutional framework created with a view to liberalizing the rail market and strengthening the position of railways as a safe and environmentally friendly mode of transport; development of the rail market in terms of freight and passenger transport performance, intermodal comparison and market-opening indicators; the financial performance of the sector, including information on the capacity, state of play and utilization rate of rail infrastructure and on development of the supply industry.

The Communication is accompanied by a Commission Staff Working Document SEC (2007)1323 presenting a number of quantitative data on the development of rail market in individual Member States and the EU as a whole. On the basis of the information requirements mentioned above, a proposal for a Rail Market Monitoring Scheme (RMMS) has been elaborated, which should enable the Commission and interested parties to monitor the implementation of the directives in the Member States and to assess the impact of the new market conditions on all actors. It consists of web pages with information on different aspects of the markets, such as passenger and freight rail transport developments, applicable legislation, the legal and administrative framework for the implementation of the community railway acquis, the available infrastructure, assets (rolling stock, locomotives), the number and names of the licensed railway undertakings and information on the labor market. As this information will only become available on a step-by-step basis and it will be regularly updated. In 2001, the EC published its White Paper on ‘European Transport Policy for 2010: time to decide’, in which for the first time the Commission placed the needs of users at the heart of its transport strategy by proposing 60 measures to refocus Europe’s transport policy on the needs of its citizens. The first of these measures is designed to shift the balance between modes of transport by 2010 by revitalizing the railways, promoting maritime and inland waterways transport and linking up the different modes of transport. For railways, the goal is to achieve by 2010 the same modal share as in 1998, thus reversing the decline of the last thirty years. Railway transport is therefore expected to grow significantly as the total transport demand in 2010 is expected to be 40 per cent higher than in 1998. The White Paper also paved the way for new Directives to improve access to the rail networks of Europe known as the ‘Second Infrastructure Package’. The new Directives extend the access rights provided by 2001/12/EC to include cabotage, i.e., loading and unloading international trains and adding and removing wagons within transit countries and to cover domestic freight markets. It will also set up a new agency and new procedures for harmonizing safety procedures and equipment specification, complementing 2001/16/EC in this respect. Earlier plans to require completely separate ownership of infrastructure management from train operations, and completely separate management of operations from key infrastructure allocation functions have been dropped for the time being in face of opposition from some national governments, and doubts from independent experts and thinkthanks. The EC has also set out plans to extend access rights to passenger services in future stages of reform. In essence, the EU perceives that the future development and efficient operation of the railway system will be made easier if a distinction is made between the provision of transport

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Railway Infrastructure Market in Europe

21

services and the operation of infrastructure. The EU recommends that these two activities are separately managed and requires that they have separate accounts, in order to have enhanced financial transparency and boost competition in railway service management in terms of improved comfort and greater quality of services. However, the EU recommends that Member States retain general responsibility for the development of the appropriate infrastructure. The EU requires that Member States must guarantee that railway undertakings are afforded a status of independent operators behaving in a commercial manner and adapting to market needs. Railway undertakings are expected to be able to adjust their activities to the market and to manage those activities under the responsibility of their management bodies, in a way that ensures efficient and appropriate services at the lowest possible cost for the quality of service required. Railway undertakings must be managed according to the principles which apply to commercial companies; this shall also apply to their public services obligations imposed by the State and to public services contracts which they conclude with the competent authorities of the Member State. EU Directives also require that railway undertakings shall determine their business plans, including their investment and financing programmes subject to achieving the undertakings’ financial equilibrium. The business plans should be developed in the context of the general policy guidelines determined by the State and take into account national plans and contracts (which may be multi-annual) including investment and financing plans. In the absence of common rules on the allocation of infrastructure costs, Member States must, after consulting the infrastructure manager, lay down rules providing for the payment by railway undertakings for the use of railway infrastructure, where such payments must comply with the principle of non-discrimination between railway undertakings. In conjunction with the existing public owned or controlled railway undertakings, Member States have been required to set up appropriate mechanisms to help reduce the indebtedness of such undertakings to a level which does not impede sound financial management or their ability to improve their financial situation. Further, to that end, Member States may take the necessary measures requiring a separate debt amortization unit to be set up within the accounting departments of such undertakings. The balance sheet of the unit may be charged, until they are extinguished, with all the loans raised by the undertaking both to finance investment and to cover excess operating expenditure resulting from the business of rail transport or from railway infrastructure management. Debts arising from subsidiaries’ operations may not be taken into account. The above policies have led EU Member States to institute a programme of organizational and financial restructuring of national railway undertakings. The degree and speed of change has varied widely: the railway undertakings of the Sweden and United Kingdom have, to date, been subject to the most radical restructuring—including changes of ownership; those of France, Belgium, Spain, Italy, Ireland and Greece have, so far, changed the least. But, as a minimum, railway infrastructure management units have now been established in most Member States—either as separate corporate bodies, or as business units within a holding company structure. And in parallel, user charges for the use of rail infrastructure have been calculated (albeit with significant international differences as regards methodology adopted, and the resultant level or structure of charges). Typically, the passenger and freight divisions of national railways have also been restructured into business units, with much greater financial accountability and stronger managerial control over rolling stock and human resources. European rail productivity, financial performance and service quality has improved as a result but principally on domestic rather than international rail

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

22

Christina Nikolova

networks. Whilst the Directives have had some impact in stimulating restructuring of railways, the very limited access rights which it provides have had minimal impact in stimulating competition in the provision of rail services or the development of open access or international rail services.

V. Railway Infrastructure Reformation and Management in the European Union It is possible to describe the extent of railway restructuring in individual Member States of the EU, according to how they have implemented policies embodied within the various Directives, in particular the extent of:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

- Legal independence; - The separation of infrastructure management from railway operations; - The financing of public service obligations; - Market access for new operators; and - Improvement in the finances of railway undertakings. The most of the infrastructure enterprises throughout Europe are state owned companies providing access to the railway networks to carriers (duly licensed and holding safety certificates). Main subjects of activity are: 1). provision of the use of railway infrastructure at equal conditions; 2). taking measures for the development, repair, maintenance and operation of the railway infrastructure; 3). collection of infrastructure charges; 4). making the train timetable, coordinated with the carriers, and for passenger transport – with the municipalities under public service obligations; 5). management of the train operations in view of the safety requirements; 6). acceptance and execution of all requests arising from public service obligations; 7). preparation and keeping of a register of the land and the railway infrastructure sites. The infrastructure companies do not deal in railway carriages and cannot have a stake in trade companies dealing in such activities. The funds necessary for the operation of railway infrastructures in individual memberstates are usually provided by: -

The state budgets; The user charges for railway infrastructure; Incomes from leased sites and other activities; Credits.

The collected funds provide for the costs for design, exploitation, construction, maintenance, development and operation of the railway infrastructure and for reimbursement of the credits. Tendering access to the rail infrastructure is a natural monopoly which requires that incentives have to be provided for the infrastructure managers to reduce the costs for maintenance and operation of the rail tracks and for efficient management of the infrastructure (Baldwin, R. & Cave, M., 1999). On the European level a long-term program

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

23

for the development of the railway infrastructure, the safety and reliable operation has been adopted, and the Ministries of Transport in the individual countries have endorsed annual programs for the construction, maintenance, repair, development and operation of certain national railway infrastructure. The second package of directives makes provision for the concessions to be granted to private undertakings under the control of public owners as an opportunity for management and building up of the existing and/or new railway infrastructures. The land remains in public ownership and interrelations are specified by long-run contracts. The relations between infrastructure managers and the carriers are regulated by contracts for access and use of the railway infrastructure signed for a period of 5 years (which is continued for another 5 years after review of the carrier’s licenses), and no re-assignment of routes to other carriers is allowed. The railway infrastructure managers are to grant all requests for capacities filed by the transport operators. Upon additional agreements and payments, the infrastructure managers may provide additional and accompanying services (provision of traction electric power, advanced heating of the passenger wagons, provision of fuel, conditions for shunting, contracts for delivery of dangerous goods and provision of access to the telecommunication network, additional information, technical control of the transport machinery, etc.). The infrastructure companies may allow discounts in the prices of these services down to the level of the actual costs made; they may introduce discount systems applicable for all users, or discount may be granted for special freight or passenger flows, for certain sections or services, for certain periods of time, in order to encourage the development of new passenger railway services, and also discounts stimulating the development of low-performance routs. The demand for the railway infrastructure can be characterized with the number of the existing rail operators and the capacities they have reserved and used.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Table 5. Number of railway operators in the individual EU Member States Operators BE BG CZ DK DE EE EL ES FR IT LV LT HU NL AT PL PT RO SL FI

SE UK

Freight

2

3

55 3

157 2

1

1

1

10 3

1

3

5

12 28 1

22 1

1

6

6

Passenger 1

1

5

150 3

1

1

1

4

1

2

2

3

9

1

6

25

7

4

4

2

1

Source: ECMT, Rail Transport: Railway Reform and Charges for the Use of Infrastructure, Report, 2006

Despite the fact that in most of the countries the railways are restructured and there exist great number of operators, in a majority of EU countries, national railways are organized as 100 per cent state-owned joint stock companies. However privately owned joint stock companies have been introduced in the United Kingdom – except Northern Ireland. In countries where railway infrastructure and operations have been separated organizationally only the United Kingdom has privatized both parts. Most governments have opted to retain infrastructure in public ownership with the creation of a state agency (non-ministerial state department) to manage it as in Denmark, Finland, France, Netherlands, Portugal and Sweden. Among the joint stock companies private capital is incorporated only in Belgium, less than 1 per cent of shares, and the United Kingdom. In Great Britain all passenger operations have been franchised (contracted for limited periods of time) to private companies and infrastructure, freight operations and ancillary businesses have all been sold to the private

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

24

Christina Nikolova

sector. In Germany, shares in rail companies, were sold after DB AG transformed into 5 companies under DB Holding AG in 1999. Only a minority share holding (up to 49.9 per cent) may be sold in the infrastructure company DB Netz AG (currently DB Fahrweg division). Uneconomic investment decisions have, historically, been imposed by governments on railways, and in most EU countries have been the main cause of accumulated debt. Insulating railway operators from such debts has been a central aim of the railway reforms supported by Directive 91/440/EEC and the second railway package. In some countries, such as the United Kingdom, this has been addressed by isolating non-commercial investments and non-commercial aspects of overall investment planning, making these the subject of specific grants from public funds. This has also been part of the strategy in many other countries where, for example, regional or local authorities will be expected to participate in investments in regional networks. The intervention of the Governments in the functioning of the railway infrastructure market is demonstrated in the following aspects:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

1) Enacting regulative legislation (Viscusi et al., 2005). As we noted above The EU Parliament has adopted the necessary legislative framework and the Ministries of Transport in individual countries – the respective secondary legislation regulating the functioning of the transport infrastructure market. Regarding the railway infrastructure these are: Directives 91/440/EEC and the first railway package and the respective regulations on their application in the different states. 2) Protecting public interests in the infrastructure market functioning. The individual states execute their rights of ownership on the infrastructure and offer access to it through the establishment of a specialized public enterprises. They participate in the funding of the activities of construction, maintenance, development and operation of the railway infrastructures. For that purpose a long-term contracts are signed for the amount of the funding between the state governments and infrastructure companies. The infrastructure managers participate in the funding of the activities related to the current maintenance of the rail tracks and their operation with theirs commercial activities incomes. 3) License regimes for the carriers – they regulate the access of the consumers to the market. This aspect of the state regulation of the railway infrastructure market arises from the non-competitiveness of the provided public service on the one hand, and on the other hand – from the possibility of exclusion of certain consumers (carriers) through application of certain criteria which they have to meet in order to be granted access. Thus the Governments ensure reliability of transport services and infrastructure tendering. The applicants for licenses have to be financially stable, professionally competent, with good reputation, and to have the financial means to insure their civil liability. According to the Directive 95/18/EC amended by Directive 2001/13/EC a financially stable candidate is the one who proves that it will be able to cover its actual or potential liabilities for a period of 12 months. In order to be assessed as professionally competent, the carrier must have internal rules for the transportation activities and its management organization. It must possess the knowledge and experience necessary to exercise safe and reliable operational control and supervision of the type of operations specified in the license, it must employ personnel

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

25

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

responsible for the safety, in particular drivers, who are fully qualified for their field of activity, it must have rolling stock and its organization could be able to ensure a high level of safety for the services to be provided. Good reputation is determined by the general formula that the management personnel of the applicant cannot have been sentenced or managed a company which has been declared insolvent and for which unsatisfied creditors have remained. Besides meeting all the above requirements, the licensed persons have to certify the suitability of the vehicles they possess for safe operation within the infrastructure and the compatibility of the qualifications of the personnel with a safety certificate. The application of regulatory measures for participation in the transport business through licensing aims at ensuring efficiency in the infrastructure use, offering transport services of high quality and achieving social wealth as a whole at cost efficiency and optimal provision of the public services. 4) Economic regulation of the railway infrastructures as natural monopoly. Key issues controlled by economic regulation are infrastructure management and charges for access to the railway infrastructure. This is done through control on the activity of infrastructure companies aiming at the more efficient use of the railway infrastructures. The regulatory bodies control the work of infrastructure managers and the carriers. They inspect the network statements, the procedures for the capacity allocation, the applications of the tariffs, the safety certificates and control the safety norms and rules. The establishment and existence of such bodies guarantee the efficient implementation of the EU policy in the field of railway transportation and the realization of the social and economic objectives under public regulation. The governmental interventions in the functioning of the railway infrastructure market are necessary due to the fact that a public service is provided. The latter has to be provided while meeting the requirements for effective distribution of the economic resources, restriction and regulation of the monopolistic activities of the infrastructure managers, correction of the market failures and its partial replacement. The state regulations does not eliminate private entrepreneurship and the market mechanisms (Brown & Jackson, 1998), on the contrary, their place and role are subordinated. The main objective is the partial correction of market failures (for example through provision of goods and services which are subject of common consumption, through protection of the state ownership interests, through regulation of the monopolistic activities of infrastructure managers and through taking into account externalities, etc.) and directing the market powers to the achievement of higher common wealth. Arrangements for imposing public service obligations (PSO) and providing compensation from public funds are critical to the commercial freedom of railway operators and to the financial viability of the sector. Most EU Member states have introduced contracts that meet the aims of the Regulation. The United Kingdom and Finland have moved furthest in making ex-ante negotiation of PSO packages as part of competitive tendering for passenger operations introducing ex-ante negotiation of the service levels required and the financial compensation to be paid. This measure is the key to the commercialization of rail operations. In most EU countries, tariffs for domestic passenger services are determined or at least approved by Government although in some of them this applies only to certain types of tickets (for example, to standard second class tickets in the United Kingdom). Tariff regulation usually takes the form either of multi-year agreements on trends in maximum

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

26

Christina Nikolova

tariffs or Government review of all tariff changes (for example in Bulgaria, Romania, Lithuania and Estonia). Government passenger tariff policies generally attempt to balance covering costs with concerns to keep the cost of rail travel relatively low for low income groups. The emphasis tends to vary over time and between countries. Negotiation is likely to be important in arriving at stable tariff structures. Where a disaggregated industry structure is adopted, an independent regulator offers the most transparent structure for tariff regulation. Freight services are subject to much less tariff regulation than passenger services in the EU. In accordance with provisions of the Directive 2001/12/EC, all EU member states have separated infrastructure from operations for accounting purposes. Accounting separation is intended to promote efficiency in infrastructure management and in railway operation by providing for a tighter focus on each of these two distinct activities and should provide for transparency in the use of public funds in the railways. Separation of accounts is also an essential step towards enabling new railway undertakings to exercise access rights by providing a basis for the development of infrastructure charges that are fair and nondiscriminatory. Many railways have gone further than the separation of rail operations and infrastructure management required by 91/440/EEC and made extensive use of contractual relations between different business units to improve the definition of management responsibilities. A step by step approach has been adopted, at the European level, for the liberalization of access to infrastructure. Directive 2001/12/EC provided for the promotion of international groupings of national rail companies by guaranteeing such groupings nondiscriminatory access to infrastructure for the provision of freight services and international passenger services. In all cases the directive covers only international services and only international combined transport operators enjoy full nondiscriminatory access to track in all EU Member States. However, individual countries are going further in liberalizing access to infrastructure, for example Germany has opened all rail markets to competition. The Netherlands has adopted an approach of allowing access rights with the process controlled by the Ministry of Transport. Priority is assigned to domestic passenger services and on main lines new operators may use spare capacity. In the United Kingdom, there is open access for freight, as there was before privatization. Access rights have been limited for passenger services by the Rail Regulator’s moderation of the competition provisions, designed to provide some initial protection to franchise operators both from new entrants and each other. There have been relatively few new international operators emerge to date but more progress has been made in introducing competition and greater efficiency for domestic services, perhaps because no intergovernmental agreement or reciprocity is required. For example, some local passenger services and short line operations for freight have been contracted to new entrants in Sweden. Germany has opened up the whole national network for access and it is possible for the DB AG network to be used by local passenger services and short line freight operators. In the Netherlands, Lover’s Rail has been allowed to compete with NS on one passenger service. In the United Kingdom, all rail passenger services have been franchised and all freight assets sold, each without associated track. European Conference of Ministers of Transport argues that although there is no evidence that the provisions of Directive 91/440/EEC have had any direct impact on freight operations so far, the mere existence of laws providing access rights may have acted as a catalyst for change. Also these reforms are ones which may be expected to take time. Not only must the national railways adapt to the new laws, but there is also a need for innovative firms to emerge to take advantage of the opportunities created by access rights. The two countries that

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

27

have so far undertaken the most radical reforms (Sweden and the United Kingdom) have demonstrated that infrastructure separation and even privatization can have substantially positive, but with some adverse effects on rail services. Both traffic and infrastructure conditions have improved in Sweden. In the United Kingdom, passenger and freight demand have improved substantially and rapidly. Private sector investment is substantially higher than under Government ownership. While certain aspects of the United Kingdom approach, for example the large number of private enterprises, might not be repeated elsewhere, others, such as the importance of correct access charges and the continuing need for government involvement, must receive careful review by all other governments. Despite the critical press coverage of a number of accidents, both freight and passenger traffic in the United Kingdom have grown faster than in any other EU country over the same time period. The concept for the functioning of a railway infrastructure market assumes characteristics of demand, supply and balance in this sector and also its state and economic regulation in order to estimate possibilities for achieving of the transport policy objectives as following: – –

–

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

–

optimum utilization of the railway infrastructure which will lead to reduction of transport costs for the society as a whole; application of appropriate schemes for levying infrastructure charges supplemented by the respective schemes for charging other infrastructures and competitive operators. That will lead to optimization of the balance between the different modes of transport; provision of incentives for minimization of the disruptions and enhanced utilization of the infrastructure; ensuring conditions for fair competition between companies offering railway transport services through application of charges for the use of infrastructure and capacity allocation.

The review of the present state of the railways in EU countries shows a developing market having opportunities for growth and success. The governments in individual member states take all required and eligible measures to ensure high quality of the infrastructure and to achieve main goals with respect to transport by rail. The European Parliament and the Council have provided legislative measures for establishing fair competition on the rail market, regulation of the infrastructure market as a natural monopoly and possibilities for development of public-private partnership in building up and managing railway infrastructures.

VI. Railways in South-Eastern Europe The processes of globalization and European integration called for amendments in the European transport model. These amendments covered prolonged period and they aimed at the adjustment of the transport systems to the new logistic necessities of the European economy. The fact that five of the Trans-European transport corridors pass through SouthEastern Europe is an additional incentive for discussing and solving the problems related with the perspective in front of the operators licensed for performing international freight and

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

28

Christina Nikolova

passenger carriages. The emphasis is mainly on the opportunities for collaboration in the field of transport among South-Eastern European countries (SEEC). The process of joining of Slovenia, Bulgaria and Romania to the EU imposed the harmonization of the transport legislation requirements and procedures with the Acquis Communautaire. As Croatia is a country candidate for EU membership and the other Balkan’s countries declare their intentions for the same it is obvious that a special attention should be given on this process. It is a prolonged process that continues after the admission of the countries in the Community and includes three main stages (Directorate General for Transport, 1999) as follows: •

• •

Introducing the Acquis Communitaire in the sphere of transport in country’s legislation by using the due national procedures and measures (laws, regulations, decrees etc); Applying the new transport legislation through setting up institutions and raising funds required for executing the laws and regulations; Control over the observation of legislation by inventing the due measures necessary for guaranteeing law appliance thoroughly.

And while the first two levels are limited in time and their accomplishment is prerequisite for carrying out the engagements taken owing to joining the EU, conducting the third stage of the process is a permanent one and ensures appliance of the European transport policy. The main goals of this policy could be set out in three spheres: •

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

•

•

Improving the quality of transport services through establishing the integrated transport systems based on the modern technologies and, thus, protecting environment and raising security of the services; Getting the functioning of the united transport market better, that is aiming at the efficiency soaring and expanding the customer choices. It is supposed to lead to a higher level of services’ quality while keeping pace with the social standards; Further development of the positive externalities through improving the quality of transport links between the EU and the third countries which gives the opportunity for penetrating in foreign markets.

The above mentioned purposes correspond with the requirements for giving open access, for secure and stable transport systems and they suppose development of the reliable concepts for improving the competitiveness of the transport companies as well as the infrastructure capacities. Collaborating and discussing the current problems of transport are, of course, necessary prerequisites for achieving the aims. It is apt to coordinate the measures for applying common approaches for better use of transport infrastructure and securing the rational development of European Transport System for international carriages. With regard to this the opportunities for achieving the aims by improving the relations between the countries in South-Eastern Europe and by collaboration in transport sector should be subject to clarification. These opportunities are supposed to be specified in the light on the development and elaboration of the transport technologies and organization of carriages.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

29

The characteristics of the different modes of transport outline different opportunities for cooperation in solving the transport problems as well as taking measures for improving the competitiveness of the transport companies on the European market.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

1. Development and State of the Railways in South-Eastern Europe The railways in SEEC are facing a crisis. The tendencies in the whole region are toward rail market contraction, increase in operational costs and need for greater governmental expenditures for subsidization of rail services (especially passenger ones). The financial implications of these tendencies are particularly severe. Despite the national economies gradually strengthen, operational costs are doubled in the railway transport. There is a loss of market share comparing to the other modes of transport. The market studies within PHARE Multi-country Transport Program revealed a consistent misunderstanding of the changing needs of both passenger and freight customers in the railway transport field (Thomson, 2001). However, most of the railways in the region do not carry out market studies for determining clients’ needs. Common complaints from the railways passengers concerned the low level of personal security, limited outlet for tickets sales which necessitating lengthy queues and long delay at international borders. The customers of the freight services are complaining about the lengthy and complex contract procedures, about the lack of tracking and monitoring systems and the lack of freight specialized wagons. In general, railways in South-Eastern European countries are failing to meet the challenges of the developing market economies. The customers have increased requirements and choice which are in favor of the other modes of transport. The situation being as it is, the urgent measures for improving the conditions and the quality of rail services are necessary. The main barriers to competitiveness of the railways in SEEC are said to be internal for these organization, without underestimate the external threats. Some of the rail enterprises in the region are still operating as centralized systems which delegate minimal responsibilities for their management and impose little commercial accountability. In 2001 the study of the Halcrow consultancy established an index of rail adaptability (Brown, 2001), defined below: A=P * C, where: А is the Index of Adaptability; Р – Power index – reflects the extent to which each railway is empowered to define its own organizational structure, appoint senior staff, set its own budget, raise finance and dictate the timetable and fares. С – Index of the Accountability – concerns the extent to which a railway is held responsible for its commercial performance (defined as the percentage of turnover subject to financial targets). A railway with an Adaptability of 1.0 would have a fully empowered management that is commercially accountable for all aspects of its business. SEEC railways had an average Index

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

30

Christina Nikolova 5

of Adaptability of 0.32, barely half that of EU railways (0.62) . Interestingly, the average accountabilities of SEEC and EU railways are similar, suggesting that attempts are being made to introduce some commercial focus. However, low power indices amongst SEEC railways indicate little progress in implementing a more empowered, commercial approach to management. Accountability and power has not been devolved within railway organizations. The Adaptability analysis was accompanied by a comprehensive benchmarking exercise. A variety of benchmarking methods were used, ranging from simple partial productivity measures to total factor productivity and cost frontier analysis (which measures the overall efficiency of the organizations). In the table 6 below are presented the values of the key factors for railways benchmarking in SEEC. The results reveal following main trends: • •

•

•

Only Slovenian railways which are advanced in restructuring have high comparative efficiency; All other rail enterprises in the region have negative indices of productivity and they confirm the theory that rising wages suppress the efficiency when the level of organizational reform is low; If all of the railways in the South-Eastern Europe achieve the level of most efficient one (the Slovenian Railway), then the prerequisites for saving operational costs and long-term savings will exist; The perspectives for salaries rising in the SEEC could lead to the growth in the operational costs of the railways.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Table 6. Key factors for railways benchmarking

Country

Railway

Total staff (railway and non-railway staff)/ gross tonne km

Bulgaria Macedonia Romania Albania Slovenia

BDZ CFARYM CFR HSh SZ

10 11 8 12 5

Wagons/ freight tonne km

Passenger coaches/ passenger km

Total factor productivity

Cost frontier analysis

10 11 9 12 6

1 11 2 10 8

5 6 9 n/a 10

3 7 10 9 11

Source: European Conference of Ministers of Transport, What Role for the Railways in Eastern Europe?, OECD, 2001. Legend: Low efficiency 12 11 10 9 8 7 6 5 4 3 2 1 High efficiency

Consequently, the main threat the railways in the SEEC face is related to the usage of the out-dated, product-led organizational structure, set within institutional framework which prevent management from the powers they need to conduct an effective commercial activity. In addition to the negative tendencies are a variety of operational and technical barriers as follows: 5

Source: Profitability of Rail Transport and Adaptability of Rail (PRORATA) prepared for DGVII of the European Commission, Halcrow Fox, February 1999.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe -

31

Lack of management and analytical information about the expenditures and lack of business-planning system in the most of the railways in the region; Low level of asset utilization as a result of the lack of management information systems; Excess of employees hired in all activities; Poor infrastructure and vehicle maintaining especially of those performing international carriages as well as terminals and tracking systems; Poor coordination in international carriages planning and management.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The study of the railways competitiveness has a significant meaning in the disclosure of the opportunities for improving their state. The main threats and barriers to competitiveness are not discrete issues. These are complex of external, institutional and technical problems. For example: a lack of commercial freedom, an absence of consistent business-planning and lack of management information systems. The obstacles to competitiveness of the railways in SEEC could be classified in accordance with their influence as follows: market, production and strategic. The responsibility for each barrier and hence the responsibility for its overcoming could be specified towards the owners/regulators, train operators and infrastructure owners (Table 7). It should be noted that the same or similar barriers appear in more than one category. This once again emphasizes the close interrelations and especially the way the strategic obstacles drive the market and production ones. Although on the production and market level the changes could be done, these changes are restricted in scope and in efficiency, if the strategic obstacles are not changed. If there exist relatively limited opportunities for modeling the interrelations between the railways and the governments, then it should be noted that they have a fundamental influence on the possibilities for changes on the other levels. This is extremely important when estimating the barriers to competitiveness of international transport services. The main conclusions that could be drawn from the analysis of the railways competitiveness in the SEEC are as follows: -

-

-

The railways need an explicit concept of the role and the objectives they follow. They should have freedom in the business management and to achieve their goals. Accountability and power should be delegated throughout the railways enterprises as part of matrix of responsibilities, targets and objectives; There is a lack of commercial focus. Few, if any, of the railways in the region are communicating and listen to their clients to find out what services they want. To a large extend this is due to the institutional framework within which the enterprises exist and to a contract with governments. In other words, the railways are not interested in studying their clients’ needs; Lack of management information systems and thorough business-analyses and the evaluation processes. This is linked also to the relations with governments and existing organizational structures in which the main responsibilities are delegated to the top-management and on the lower levels the necessary controlling information is not consigned. The managers do not have at their disposal sufficient data for adequate decisions and recourses’ conducting even when they are empowered for this. The full potential of management information systems could be revealed only provided that the organizational structures of the railways are such that could give enough power to managers to use these systems;

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest Ebook

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Table 7. Barriers to competitiveness Responsibility Production Owner/ −Lack of clarity and transparency in the relationship regulator between government and railways; (government) −Lack of objectivity and clarity within plans for restructuring, commercialization, privatization and access charges, etc. −Insufficient funding for infrastructure and rolling stock rehabilitation; −Insufficient subsidies for loss-making services; −Strong trade unions inhibiting reform and restructuring.

Market −Regulation/control of fares by government, below market levels; −There is a lack of flexibility in freight tariffs; −High levels of concessionary travel; −Government control of timetable; −Inadequate compensation for social obligations to operate loss making services; −Lack of integration with other modes.

Train operators −Inappropriate organizational structures and accountabilities; −Absence of freedom to manage; −Absence of business management structure, techniques, tools and skills; −Absence of suitable management information systems linked to analysis and evaluation processes; −Absence of market oriented strategies; −Absence of personnel training and development strategies; −Bureaucratic processes and resistance to cultural change; −Poor condition of locomotives and rolling stock; −Lack of equipment and spare parts for repairs and rehabilitation; −Poor productivity of resources; −Poor quality products and delivery; −Increasing operational costs for labor, materials and services; −Poor level of cost recovery and significant loss-making activities; −Inadequate revenue collection and protection systems; −Lack of common technical standards and operating procedures for international services.

−Poor ticketing and retailing methods; −Poor passenger information; −Lack of customer focus; −Poor marketing of services; −Lack of market research and communication with customers; −Poor quality station environment and facilities; −Poor quality rolling stock and facilities; −Passenger security; −Lack of integration with other modes, including car parking at stations; −Poor timetabling and low frequency services; −Overcrowding; −Inflexible and lengthy contract negotiations for freight transport; −Freight tariffs are complex and difficult to understand; −Lack of specialist freight wagons; −Speed and transit times are considered to be poor and unreliable; −Tracking of consignments; −Delays at border crossings; −Poor security of goods.

Strategic −Strong state involvement in management and policy; −Railway management has limited powers and accountability; −Lack of clarity in contract with government for social role of railway; −Slow rate of legal reforms and introduction of progressive legislation; −Poorly executed and ineffective privatization; −Limited privatization polices; −Limited implementation of EV directives. −Inappropriate organizational structure; −Lack of/ineffective management information systems; −Lack of business analysis ethos/techniques; −Growing financial deficits.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest Ebook

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Table 7. Continued Responsibility Production Infrastructure −Poor condition of infrastructure and significant backlog owner in maintenance and renewals; −Lack of equipment and spare parts for repairs and rehabilitation; −Poor resource productivity; −Increasing operational costs for labor, materials and services; −Route capacity constraints – single lines, level crossings, permanent speed restrictions, etc. −Excessive networks in need of rationalization.

Market −Poor interchange facilities; −Poor quality station facilities; −Low line speeds; −Inefficient freight terminals and poor freight handling facilities.

Strategic − Costs of operations and infrastructure are not made explicit and potential barriers to new entrants remain; −Excessive networks; −Lack of investment and poor condition of infrastructure.

Other/ common

−Rising customer expectations; −Liberalization of road freight carriages; −Rising levels of car ownership. −Low purchasing power of the population in some countries, particularly the Balkans.

−Low motivation to restructuring and reform within the rail industry; −Rising labor costs; −Growth in car ownership; −Growth of competing modes – car, road freight and in some countries efficient state airlines; −Regional conflict in the Balkans; −Poor economic performance and low purchasing power.

Source: European Conference of Ministers of Transport, What Role for the Railways in Eastern Europe?, OECD, 2001.

34

Christina Nikolova -

-

-

It is difficult to conduct any form of business planning and to choose eligible market strategies in view of the fact that the investors, respectively the government does not confide to the management of the enterprises; Rolling stock productivity is lower than that in Central and Western European Countries. This leads to higher costs and is related to a lack of adequate management information and business-estimation and monitoring systems; The hired staff is more than really needed, which imposes lower labor productivity than this in Western and Central European Countries. The reason is the absence of eligible management information systems but also it has a close link to the relations with government and the delegated powers of management;

The growing labor’s costs impose serious problems which to a large extend reduce the effect of staff reduction and rapidly drawing up the operational costs. At the same time the railways fail to achieve positive results from the restructuring; -

-

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

-

Collaborations among railways in the region, and between railways and their customers are on a low level. This leads to lengthy delays on borders, unreliable international services and a general lack of customer confidence. There is a lack of standard operating procedures and technical standards and a unitary body responsible for marketing and international services between countries and along TransEuropean Corridors; The investments in traction and rolling stocks are very low, that is why it is necessary to ensure that the investments are properly evaluated by railways to have maximum benefit. Failing to address most of the problems mentioned above will see lower level of funds and continued decline of the services offered; Customer expectation and requirements rising and the populations want to achieve the standard of live in Western and Central European Countries. The railways failed to keep pace with these changing requirements. When providing international freight services it couldn’t proceed from the literal space moving of goods between two points. Transport services should be considered as an integral element in the entire logistic chain. The product offered should rather be a service which enhances the goods flows than vehicles that perform carriages of goods. Freight services will become increasingly important but the railway transport systems of the SEEC fail to meet these requirements.

It should be recognized that in the railways in some of the countries in the region are established and developed good practices and institutional models. Bulgarian and Romanian railways are restructured and the freight traffic is stabilized and does not go down anymore. The comparisons amongst the railways efficiency in SEEC show that it is possible to achieve significant cost savings if all of the countries in the region reach the level of the most effective railways –Slovenian railways. The table 8 below presents the barriers to competitiveness of the railways classified by countries. Of course, there are always some exceptions but the goal is to focus on the common problems.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

35

Table 8. Common barriers to competitiveness Country Slovenia and Greece

Bulgaria and Romania

Albania, Bosnia and Herzegovina, Croatia and FYR Macedonia All

Production

Market

−Labor costs rising −Limited management powers −Large networks of branch lines −Poor rolling stock utilization −Low labor productivity −Maintenance backlog −Poor rolling stock utilization −Low labor productivity

−Train service not market led −Growing customer expectations −Poor retailing/ticketing −Low speeds −Basic marketing −Poor information −Poor logistical management systems

−Maintenance backlog −Single line working −Low resource and labor productivity

−Low speeds −Poor condition of coaches and wagons

−Passenger security −Theft of freight −Condition of rolling stock

−Bureaucracy at borders −No market research −Poor integration

Strategic -

−Growing deficits −Competition from other modes −Contract with government −Lack of investment −Balkan crisis

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

−Lack of management information systems −Lack of business evaluation and analysis systems and processes Source: European Conference of Ministers of Transport, What Role for the Railways in Eastern Europe?, OECD, 2001 and personal assessment

In spite of the reforms undertaken, the major part of the railways in the SEEC is still in decline. Special attention should be paid to the perspective of raising salaries in the region in connection with the economy growth. This will bring to a paradoxical rise of the risk for the railways in times of getting economy conditions better. Classifying railways of SEEC into three groups is based on their similar characteristics. Group one includes Slovenia and Greece, which have stable economy but have low cost efficiency. The second group consists of Bulgaria and Romania that find themselves in period of soaring economy, but have not yet accomplished thoroughly their reformation in railways. It needs to be emphasized that the main problems in the SEEC railways are not caused by lack of investments. In most cases problems derive from the fact that due procedures of correct estimation of the investments’ needs and for directing the funds for modernization and rehabilitation of the rolling stock and infrastructure have not been duly carried out. Something more, it is even possible raising investments to have little impact on the state of railways companies. This is typical for the third group of countries including Albania, Bosnia and Herzegovina, Croatia and FYR Macedonia. The reason here stems for the low level of the conducted institutional reforms. Besides, railways in all SEEC are not supposed to expect the market growth in all the segments to solve their financial difficulties if not restructured. At that the attitude and the approach to the customers should, by all means, be changed in order to preserve at least the existing market shares. The complexity of mentioned problems clearly shows why the situation in railways continues to get worse and the necessity of instant measures are required. Therefore, the

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

36

Christina Nikolova

priorities for the railways in the region should be implementation and conducting programs that will help to the restructuring and will focus on rising operational efficiency. If more commercial orientation is being imposed, then it will enable invention and carrying out more ambitious marketing programs.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

2. Measures for Improving the State and Efficiency of the Railways in South Eastern Europe and Opportunities for Partnership Multitude of research and recommendations for the railways development (Commission of European Communities, 2001) define different opportunities for improving their status, efficiency and competitiveness. Some of the opportunities are linked to supporting institutional reforms and commercial liberalization. This, in turn, requires change in the regulatory and institutional frame as a precondition for achieving the expected results. Such kind of changes will bring to a further progress from heavily regulated and controlled structures toward a business orientated companies and even private ones. Other measures are being implemented in order to get the efficiency better without carrying our implicitly structural reform. These may include changes in exploitation activities, asset management and marketing. Nevertheless, the competitiveness will go up as the institutional and organizational structure change, too. It should be taken into account that the index of adaptability of badly-run but liberalized railways in any case is higher than those of well-run, but with traditionally set structure. When the regulatory frame is made up in such a way as to allow changes in the company adaptability, it is possible to boost the efficiency in their activities, but it will not be evident at once. Most effectively working railways in Europe – those of Sweden, Germany and UK are still relatively less efficient and adaptable. However, at present, the aforesaid countries represent the most successful models in the development and could be singled out as a sample to any other railway. The railways of these three countries have achieved better efficiency through implementing a sequence of successful measures along with the development of their adaptability and as a result they have improved their status as a whole. On the other hand, the measures are strongly related to improving the power index. Bearing in mind the example with the railways of Sweden, Germany and UK, we can make out an action plan with priority measures. The potential financial benefits from their 1 implementation could run up to around EUR 4-5 billion a year . Major part of the measures aim at improving competitiveness; e.g.: developing organizational structures and accountancy systems, drawing out business plans and implementing new products could easily be applied by the very railways. What matters in this case is that the companies can take decisions by themselves. The external factor influence should not be underestimated as a means to rule the changes and yet this should be handled through partnership between the government and the railways. There should be undertaken regional measures that will incent the railways to action in increasing the market share. This cannot be reached only by offering lower prices or increasing frequency of the transport services. At first there should be undertaken a reform in the sector, more power of attorney to be delegated and the costs for control to be estimated. 1

Source: European Conference of Ministers of Transport, What Role for the Railways in Eastern Europe? OECD, 2001.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

37

International partnerships, respectively in South-Eastern Europe are of strategic importance for implementing new practices and technologies, e.g., through leasing and franchising. The international freights in the region offer the opportunity for including in the private sector as in the exploitation, so in ensuring specialized freight wagons. The following measures could be summarized by countries: •

•

•

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

•

Albania, Bosnia and Herzegovina and Macedonia: there exists necessity of rehabilitation of the network and rolling stocks, in order to have the normal functioning ensured. The outside reasons, including the political barriers, however, deter the progress. Yet the applying decent models is possible for setting up institutions, work organization and procedures from the other countries in the region that are ahead of the reforms in the sector. These measures are basic in achieving profit from the investments in the infrastructure. Bulgaria and Romania have both some good results in restructuring and institutional reforms in the railways. Measures, that the railways of these countries should take in these sectors for increasing the efficiency are supposed to be directed to the businessmanagement and exploitation issues. The approaches to improving financing and supplying are of major importance, too. For the railways in Slovenia and Greece the accent of the perspectives actions and measures for getting better the efficiency, could be put on the government (institutional questions and measures for increasing the responsibilities). In these countries the reforms are on their way for a long time and the first positive results are evident. The international freight and passenger transportation in the region are expected to peak considerably until 2015. Nevertheless, the forecast for the railways are rather to generate further market share losses to those of the road and air transport. Although some specific barriers exist on the way of the international transport in South Eastern Europe (e.g., delays in crossing borders, different technical standards and equipment, poor quality of the rolling stock), they can positively be overcome. What matters is that as the domestic so the international transport services in the region as a whole are product-oriented and do not respond to the expectations of the clients. The measures to bring change and redirecting the railways from exploitation to commercial control will add for improving the quality of the international railway freights. There exists an opportunity for penetrating of private companies and developing of innovative services in the field of the freights.

The EU has a clear role in fostering the implementation of the directives linked to ensuring an open access to the railway infrastructure and collecting the charges for its use. By doing this, further institutional reforms base frame is being ensured. If the access to the railways is open for sure and innovative international transport services are to be developed, then the next step is developing system for monitoring. Creating a European Railway Organizations System could become a model that will constantly be expanding and to include all the SEEC. The crisis in the railways in SEEC could not be prevailed only by investments in new vehicles and infrastructure. Finding solutions to the problems are in the competence of the

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

38

Christina Nikolova

railways by themselves and their owners – the governments of the countries. In order to rid of the causes of the low competitiveness and efficiency, organizational and institutional reforms are required. Management should become more independent to posses more rights and responsibilities. It is necessary that the present model of product-oriented management to be replaced with more commercial organization. Along with it standard business processes could be implemented. This would inevitably lead to considerable savings of exploitation costs. Reinvestment of funds made by the savings in commercially focused structures will bring to increase in the competitiveness of the railways which will be able to respond to the higher level of demand in the market. The transport market in the region has every opportunity to grow, yet it is required that the accelerating of the processes on realization of infrastructure projects in the field of transport and carrying out a common infrastructure policy is ensured and based on following principles: • • • •

European orientation, stemming from the integration of the most of the countries in the European economical and political processes; Developing the infrastructural policy within the frame of multilateral forums and organizations; Combining infrastructural projects on the side of the Black Sea and Danube partnership with common European orientation; Taking into account the relation and opportunities for development of the transport and telecommunications in direction to the Middle and Far East, Asia and North Africa.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

VII. Infratsructure Charges in EU 1. Basic Principles of Charging The development of tariffs for the use of infrastructure varies greatly between EU Member countries - most of them have identified procedures for setting charges, and a number have laid down precise rules for the structure and level of tariffs. In others, infrastructure companies, or divisions, are responsible for setting charges. The objective of most governments that have set rules for infrastructure charges is to cover costs, differentiating charge elements to reflect such factors as type of service, wear on track, distance of run, routing, etc. The European Commission in its documents (CEC, 1995 & 1998) recommends the marginal social cost – based pricing as the start point for charging for the use of infrastructure. This approach could be improved in recognizing budgetary funding and thus to achieve second best ‘MSC based pricing’. There is a number of endorsements considered as sufficient for applying these principles (Nash, 2003), notwithstanding some difficulties in the process of their implementation in transport infrastructure charging system and some objections against them (Rothengatter, 2003). The caveats of the marginal social costs pricing models, currently in use in infrastructure charging systems over EU countries, could be rebutted when the model envisages also theory of regulation of natural monopoly and theory of public sector economics. Of course, there exist also other pricing principles applicable to

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

39

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

the transport infrastructure, such as the Average Cots Pricing, Development Costs Pricing and Long Run Marginal Social Cost Pricing (Quinet, 2005). Some of them have certainly been involved in the Infrastructure Charging Systems across European countries, but there has also been a strong argument in favor of starting with and estimating marginal social cost principles as the most appropriate approach. Defining the transport infrastructure as a public service requires picking and choosing an adequate methodology for comparison and measuring costs for the railway infrastructure operation and, respectively, taking into consideration the external costs caused by the use of infrastructure for transport services. Such a choice should be based on the comprehension that an infrastructure manager that provides public services has to be compensated for the incurred costs, as it is usually the state that possesses and functions under the rules of public ownership and welfare economics theory. On the other hand, the attainment of the high level of cost recovery through the involvement of trade-offs would ensure the accomplishment of the main objectives related to the efficient use and provision of infrastructure, encouragements for competition in railway transport market and consistency with government objectives. Therefore, it is important for the infrastructure managers that operate on the market of railway transport infrastructure to assess marginal social costs and to figure out an appropriate level of user charges. Applying an adequate infrastructure charging system is a prerequisite in providing access to all transport operators at equal and fair conditions, for tendering services at required quality and managing infrastructure demand in the right way (Baumstark, 1998). Furthermore, thus society could show its values for the output of provided services are higher than opportunity costs for provision of other goods (Dodgson, 1998). The present section makes review of the methodology for estimating marginal social costs for railway infrastructure and explores the opportunities for improvement of cost planning and management. The reviewed methodologies are based on the background of research on estimating marginal infrastructure costs for transport accomplished in EU member-states and, respectively, adopted approaches to determining infrastructure charges.

2. Existing Approaches for Measuring Marginal Social Costs of the Railway Infrastructure in the EU Countries In the EU a number of studies have been, carried out, setting up the basis for invention of two main streams of approaches in estimating the marginal social costs for transport infrastructure (Nash & Matthews, 2006). These are described below: -

-

Top-down approaches: estimate the costs of maintaining and renewing the infrastructure and their variation with traffic. This group includes three approaches implemented in the EU - econometric techniques for estimating the infrastructure cost function, engineering methods and cost allocation method (the accounting approach). Bottom-up approaches: aiming at measuring marginal wear and tear costs of infrastructure incurred by an additional train on rail tracks. Usually these approaches are used in the railway transport only for the allocation of variable costs. Their application considers the estimation of first derivative of marginal cost function and concerns econometric techniques, as well as engineering ones.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

40

Christina Nikolova

The application of one of the above mentioned approaches depends on the characteristics of rail network, considered cost elements, adopted maintenance and renewal strategies and available data for analysis. According to the marginal social costs pricing of infrastructure (CEC, 1995), only the cost incurred by an additional unit of traffic (one train) is to be taken into account for the purposes of pricing. In the short term, the fixed costs (cost of the existing infrastructure which does not depend on the freight volume) are defined as sunk costs. As far as the efficiency is concerned, these costs are irrelevant to the process of pricing. The main issue here is to clearly distinguish between the fixed and variable costs. Provided that the infrastructure capacity is considered to be constant, the cost of scarcity or of the capacity enhancement in case of congestions should not be considered as preliminary cost. If the capacity is really enhanced, this cost will be included in the following capital investment expenditures, which are not short run costs. It is important to note the double reading problem which could arise when calculating the infrastructure cost and the cost for dealing with congestions and environmental pollution. The assessment of the marginal infrastructure cost is very important when accounting for the costs. In this context, the assessment of the fixed and variable cost under the top-down approach, as well as the methods for their allocation by type of train, in accordance with the causal factors, deserves particular interest. The research and studies about these approaches and the empirical estimations held in the EU countries show that three main methods are used for achieving the objective. These are: econometrics and engineering -based approaches (Link et al., 2002) and cost allocation - accountancy based approach. The differences among the three methods are in principles and in data used, as well as the econometrics approach estimates rather cost function but the engineering one measures the relationships between use of the infrastructure and caused damages. The accounting approach for cost allocation is more often used in the industry for determining variable elements of infrastructure charges. The three main groups of costs estimated as the most important for measuring the marginal social costs of railway transport infrastructure (Nash & Mathews 2005) are set out below: -

additional maintenance and renewal costs incurred by an additional traffic unit to the existing infrastructure; costs imposed on other users of the railway infrastructure, i.e., costs for delays, congestions, scarcities and accidents; costs imposed outside the transport system as well as some costs which are not recovered by users, i.e. environmental costs.

In this section the emphasis is laid upon the costs for maintenance and renewal as basis for determining certain charges. The fundamental directions for connecting the elements of costs with the derivative short run marginal costs for maintenance and renewals concerning the existing railway network are as follows: -

maintenance and renewals; running traction costs; signaling cost;

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

-

41

train planning cost; costs for terminals; interruption cost caused in other railway services.

On applying top-down approach for cost estimation, different studies show that these costs take a relative share of 5-10 % of the total railway infrastructure costs (Rothengatter, 2003). It is not clear, though, what exactly empiric data these numbers come from. DB AG (DE), for example, defines as variable about 15-20 per cent of its average cost, Finland – 20 % of the variable maintenance and renewal costs etc (ECMT, 2005). However, the fact that this information quite often is not published must be taken into account when calculating the access charges and developing cost schemes, based on the average variable costs. Moreover, the institutional structures become even more independent from the governments working on the basis that this is confidential commercial information and they are not supposed to disclose any on a large scale. The econometric methods applicable under bottom-up approach test the marginal costs function and describe costs’ behavior and relationships. The most frequently applied model under these methods is the translog model (Link & Nilsson, 2005). In the econometric methods for estimating the marginal costs of infrastructure use, it is necessary to measure traffic volume as an infrastructure output. The second stage is to distinguish between the different impacts of certain vehicle types on infrastructure damages. The application of this model requires vast data for individual track section related to traffic volumes, km run by different trains, gross tonne and axle-load km for each type of trains and, of course, data related to different cost categories (classified in accordance with input prices) or big time series. The main advantage of these methods is providing sufficient evidence on dependence of variable costs on certain measures (variables, indices), but on the other hand, this approach is data-demanding and provides only aggregates measures (Betancor et al., 2005). The application of engineering methods when using bottom-up approach for measuring marginal costs for the railway infrastructure is related with the use of concrete simulation techniques. In view of the fact that these methods derive marginal costs as consequence of the estimation of maintenance and renewal needs, it could be concluded that they are more theoretical and suggest perfect compliance between the real infrastructure costs and estimated costs on the base on above mentioned needs. This is not always possible, so the chance for overcharging exists. Applying the accounting method for the allocation of infrastructure costs is based on average variable costs pricing. If there is a lack of appropriate data for applying some of the other methods, it could be considered as a closest approximation to the marginal costs pricing for the infrastructure use. The cost allocation for the railways in this case could follow the concept used in the road transport: first, calculating the fixed and variable cost and the cost for providing the necessary capacity; second, dividing that cost by the freight volume or the type of trains, using different spreading metrics. Along with this method, econometric analysis of allocation factors could also be applied (Betancor et al., 2005).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

42

Christina Nikolova

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

3. Charging Approaches The variety of problems described in Europe by the diversity of approaches to infrastructure charging could not be treated separately. That is why a gradual and progressive harmonization of charging principles in all major commercial modes of transport is required according to EU White Paper, 1998. The charging system must be based on the user-pays principle i.e., all users of transport infrastructure should pay the costs, including environmental and other external impacts they impose, or as close as possible to the point of use. It is important to note that changes of charges are not automatically reflected in final transport prices, as transportation companies may adjust their use to reduce their costs. Moreover, the state will continue to support the provision of public services through subsidies to transport operators using the infrastructure and to infrastructure managers as well, thereby off-setting effects on prices paid by final consumers (railway operators, for example). It can also directly compensate infrastructure managers for wider benefits (e.g., improved land use) that the provision of infrastructure generates for non-users. The development principles do not impose a centralized state charging scheme. Rather, they provide a framework within which the infrastructure operators would be free to set charging levels. The "marginal social cost” charging principles should enhance both the efficiency and the sustainability of the transport system (CEC, 1995). Having already set up this framework, it allows addressing specific problems related to transparency, fairness and discrimination effectively. The approach based on common principles for setting the charges is phased in gradually in order to give transport users and providers time to adjust. On the other hand, the need for charging of external costs when setting the total charging levels leads to a better covering of costs but the total charges could be raised to a considerably higher level especially on road transport. Marginal social costs pricing requires a charging system based on the marginal costs. As for the infrastructure costs, the most appropriate are considered to be the pricing schemes with different prices for the various types of infrastructure and transportation means. Since the rest of the social cost of transport (e.g. traffic congestions and environmental pollution costs, accident and scarcity costs) can be similarly integrated in the charges, the base price for using the infrastructure is the one that covers the marginal infrastructure cost. The estimated marginal costs under one of approaches described in section 2 are a basis for determining railway infrastructure charges. In relation to this, across EU countries three main approaches are tended to be followed (ECMT, 2005). These are: -

-

Short run marginal cost pricing (SMC)—it is based on measured short run marginal social costs and it permits efficient use of infrastructure. It takes into account both marginal private and marginal external costs. The first group includes wear and tear costs (maintenance and renewal costs), train planning and operation costs, congestion or scarcity costs borne by the infrastructure companies. The second group includes congestion costs borne by third parties, e.g. environmental and accident costs. Marginal cost pricing with mark-ups (MC+)—could be described as departure from SMC - includes charges for covering at least maintenance and renewal costs and where it is possible a part of management and investment costs as well. The main principle in applying this approach is to minimize the distortions in competition and

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

43

the loss of traffic with high elasticity towards charges. With regard to this mark-ups should vary by certain sections of infrastructure or certain market segments. This approach is consistent with achieved goals of FC approach. These two approaches proceed from the economics theory’s concept that marginal social cost pricing is socially effective and fair. They impose state budget share in covering full social costs for the use of infrastructure. Different approach is: -

Full financial costs pricing (recovering) (FC-)—main presumption is cost recovery after receipt of budgetary grants for commercial organizations which are infrastructure enterprises. Applying this approach allows protecting financial results of infrastructure managers, but does not hold out enough incentives for increasing quality of services and efficient use of the railway infrastructure.

In the EU countries different approaches are applied depending on the level of implementation of infrastructure charging system, stated objectives in infrastructure charging and annual budget funding available (see table 9). Table 9. Adopted charging approaches in individual EU countries Approach

Countries

MSC

NL, SE

MC+

AT, CZ, DK, FR, FI, PT, RO, UK

FC-

BE, BG, EE, DE, HU, IT, LV, LT, PL, SL, SK

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Source: ECMT, Rail Transport: Railway Reform and Charges for the Use of Infrastructure, Report, 2006

In implementing certain approaches for charging for the use of the railway infrastructure across Europe, following tariff tools are used: -

-

Simple tariffs—the charges vary directly by the use of infrastructure and they are measured as for example gross tonne km or train km. The main advantage is simple application and low price for the implementation. This tariff is the most appropriate for less complex networks and less complex use of the capacity; Two-part tariffs—include a part variable with use and a fixed part based on expected use of capacity as train path or train path km. This tariff ensures covering of the part of fixed costs. It is more appropriate for mixed-use networks and in conditions of higher requirements for cost recovery. Such a tariff is less distortable for train operators than mark-ups on variable charges, but it could yields discrimination between users, especially in the case of international freight carriages and on-track competition (Link, 2004).

Regarding the currently applied charges in the EU countries, it should be considered that the actual level of charges differs widely among states (see Figure 9). The differences shown may partly reflect the overall cost levels in the different countries, and the different levels of efficiency with which rail infrastructure is constructed and maintained. It also reflects

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

44

Christina Nikolova

differences in local circumstances and different objectives concerning the government contribution to infrastructure costs. Differences in the level of charges can also reflect excess costs for some railways when the network is over-dimensioned for current demand. Severe cases of this developed in Central and Eastern Europe with the collapse of traffic in the late 1980s and early 1990s. Rationalization of the network is underway in some cases, for example Romania; in some other countries the problem persists. All these differences condition the need for the revision of currently applied charging regimes, as well as decisions on the application of mark-ups to passenger, freight and in particular international freight traffic. 9 F reight tra in s P a s se ng er tra in s

8 7 6

A rrow s in dic a t e C E E

1 0 0 0 g ro s s to nn e fr e ig h t t ra in 5 0 0 g r oss to n ne in te rc ity pa s s e ng e r tra in 1 4 0 g r oss to n ne s u b u rb an p a ss en g er tr ain (c h a rg e s h o w n fo r p as se n ge rs is w e ig h te d a v er a g e o f inte r c ity a n d s ub u rba n )

5 4 3 2 1 0 S

N

NL

F

B

P

CH

I

SI

SF DK

A

UK

CZ

D

BG RO

H

EE

LT

LV

PL SK

Source: ECMT, Rail Transport: Railway Reform and Charges for the Use of Infrastructure – Report, 2006

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 9. Average rail infrastructure charges (€/train-km)

Mark-ups come essentially in two forms: fixed charges as part of a two-part tariff and mark-ups on variable charges. The latter will be a charge per train-km (or vehicle or gross tonne-km) varying with the nature of the traffic the train carries. Great Britain, Italy, France, Bulgaria. Hungary, Lithuania and Romania have two-part tariffs. Where the ‘fixed’ part of the tariff is in fact a charge per path or per path-km, it may of course simply reflect marginal costs of train planning or of scarcity, and a mark-up applied to this is similar to a mark-up per train-km. A true fixed charge is a lump sum for access to the infrastructure (possibly related to the route length accessed, as with the two-part tariff that used to exist in Germany). Fixed charges are attractive inasmuch as they permit mark-ups without distorting the incentives to train operators regarding the number and types of trains to run. Unless the fixed element is designed carefully it may create distortions by preventing some operators from accessing the system at all and by biasing the terms of competition between large and small operators. In Britain, the fixed element is charged only to passenger franchisees, and covers their avoidable costs (i.e. not just variable costs but also any fixed costs that would be avoided if the particular set of services were no longer running) plus a share of all remaining joint and fixed costs. This charge is simply reflected in the payment the franchisee receives for operating the franchise, and therefore there is no need for differentiation according to ability

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Railway Infrastructure Market in Europe

45

to pay at the level of individual trains or train types. The franchise system allows fixed costs to be passed on in a fixed charge without any distortion of competition. In Italy the fixed charge varies according to the characteristics of the route used, being higher for higher quality infrastructure. It has to be paid by all operators and may deter some operators from entering the market. In France, the current pricing system collects 55 % of its income from train path reservation charges, and another 16 % from station stop charges, a total of 71 % of all charges. By comparison, only 4 % is collected from track access charges and another 14 % from running charges. Traffic in France has actually been stagnant or slightly shrinking since 1990. It is not immediately clear why so much of the access revenues should be generated by prospective charges for the use of infrastructure, which can be regarded as a form of mark-up. The disadvantages of two part charges can be outweighed by their advantages for allocative efficiency in situations where competing train operators are roughly equally matched in terms of market share or control of access to essential facilities. Currently in Europe, however, competition is mainly developing between relatively small new train operators and large government owned rail companies, sometimes part of a holding company that also owns infrastructure. In these circumstances the fixed component of two part charges can be highly discriminatory. This argues for two part charges to be avoided in European markets where policy seeks to promote competition on the same tracks (as opposed to the urban commuter case where the objective is competition for the market). Elsewhere (e.g. Germany, Belgium, Italy etc.) mark-ups take the form of variable charges and are generally related to the type of traffic. In some cases (in Finland, Sweden, Denmark) mark-ups are used on new routes to contribute to their financing costs. It is doubtful whether mark-ups on specific routes to help fund capital costs are efficient; there is no reason to suppose that elasticities are systematically lower on those routes than elsewhere, although some of these routes involve bridges where there is a toll on road traffic too, and there a mark-up may be feasible without intermodal distortion. Determined under some of these approaches and tariff tools, railway infrastructure charges are different in the different EU countries. The reasons for the variation are related with the specific features such as different cost levels, construction efficiency and maintenance practices as well. On the other hand, adopted approaches for railway infrastructure charging reflect also the local conditions and the amounts of state subsidies for the infrastructure. This variety of approaches raises the complexity in specifying the most effective and non-distortive charging system. But the application of the most appropriate charging regime for certain conditions and level of desired cost recovery could lead to a good approximation to the pursued level of effectiveness in the use of railway infrastructure.

4. Assessment of the Approaches The reviewed studies and the research across EU countries and especially the results of the projects IMPRINT, UNITE and CATRIN cover the acquaintance with the knowledge in the area of measuring marginal social costs for infrastructure use. Moreover, this review permits indication of the most appropriate approach for measuring marginal infrastructure costs for the individual EU countries.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

46

Christina Nikolova

The currently used charging systems are not based clearly on the marginal social cost principles. The acceptability of MSC pricing is obvious. It enables precise calculation of infrastructure costs and thus the basis for further estimation of marginal external costs could be established. Following this sequence, the prerequisite for the implementation of much more precise charging will be achieved. For the optimal charges to be set, the required revenue must be assessed also. There is obviously a necessity of applying the principle of transparency in cost’s accounts, which will lead to evidence that there is a pricing orientated towards the infrastructure. In the cases of lack of reliable infrastructure data for econometric analysis, engineering approach is a good variant for estimating the infrastructure marginal costs of the railway infrastructure. It is based on traffic and infrastructure data and number of vehicle axles. This is a top-down approach which estimates the percentage of the costs varying by the level of traffic (Heike et al., 2002). The model for marginal cost estimation in this case could be based on existing accountancy’s distinction between fixed and variable costs as it has been done in Booze Allen & Hamilton Study (1999) on track access charges in Britain. For supporting the further flexibility of infrastructure charging system in railway transport it is clear that there is a necessity of specifying the relationships between costs drivers and development of infrastructure costs. Therefore, the next stage is to allocate variable infrastructure costs on different vehicle types. Hence, the obtained results are respective marginal infrastructure costs. They could serve as a basis for further application of MC+ approach to the infrastructure charging in order to achieve higher level of infrastructure costs recovery. As main disadvantages to the afore-said approach, however, it could be mentioned the high level of relative assumptions, which could mislead some of conclusions. The requirements for precise estimation and measuring of marginal social costs for the use of railway infrastructure impose the applying of econometric approaches. Latter have the main strength of producing accurate evidence on the relationships between costs and transport infrastructure output. There exists a variety of econometrics methods which could be used for this purpose. The exact choice of one of these methods should be based on the estimation of data eligibility. Revising the application of linear, Cobb-Douglas and translog function in later years, it could be concluded that flexible specifications are most exercised in transport costs studies. So this method is well known and adaptable and could provide appropriate measure for determining infrastructure charges.

5. Requirements for Data Provision According to EU legislation infrastructure managers are obliged to submit financial reports with data related to infrastructure costs and information about infrastructure usage and quality as well. This empirical data are not enough as a basis for building up above mentioned approach. The experience of the EU member-states shows that the greatest part of required data either has not been submitted or it has been submitted only partly. Hence, for research purposes it is necessary to get access to unusually detailed data which is a matter of some difficulties due to confidentiality constrains. The application of the econometric approach suggests that the above mentioned data should be disaggregated by track sections. The following data are indispensable for cost function estimation:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe -

47

Technical information related to certain types of vehicle; Costs causes—for example: tracks construction standards (number of sleepers), geographical region, (mountain or environment sensitive regions);

For the purpose of econometric approach railway infrastructure should be subdivided into sections having closest costs magnitudes. Then it could be possible to collect data related to length of each section, number of tracks, type of power supply, type of safety control, maximum allowed speed. If submission of detailed data is impossible, then as a minimal requirement should be accepted reporting data for different track categories. Some other requisite data are: -

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

-

Cost classified as follows (High Level Group on Transport Infrastructure Charging, 1999): costs for rail track’s construction (including land purchase, construction of new lines, upgrading/enlargement of existing lines, major repairs, renewals), costs for construction maintenance (minor repairs, ballast cleaning etc), costs for ongoing maintenance and operation (including winter maintenance, cleaning, check of facilities conditions, servicing bridge bedding, signaling, telecommunication facilities, operating of these facilities, traction current), administration costs(overhead, police, time tabling and train planning); Gross tonne km and train km; Subdividing of transport services depending on train-types (gross weight, axle weight etc).

Providing this large database it becomes possible to ensure reliable measuring and estimating of marginal costs for the use of railway infrastructure. If it is impossible to provide collecting of this huge database, then for research purposes the transfer of methodology suggested by Maarten van den Bosshe et al. (2005) could be applicable. Under the conception of transferability of econometric methods for measuring and estimating marginal social costs for the railway infrastructure it is possible to support this method by cost allocation (accounting) method and farther transferring cost elasticities from econometric studies elsewhere. For instance, assuming that cost elasticity could be measured as ratio of marginal to average costs, then it is possible to combine data on average costs for measuring certain marginal costs. Main problem in such an approach is to choose correct magnitude of the cost elasticities of infrastructure costs with respect to traffic volumes. The choice should be made regarding to available econometric studies and results in countries with similar level and conditions of traffic. By applying in this way the suggested transferability tools, the appropriate and reasonable approach for infrastructure charging, could be applied.

VIII. Opportunities for Internalisation of the External Costs In certain modes of transport there exist different procedures for costs’ identification and recovery. With regard to ensuring efficient and sustainable use of railway infrastructure it is necessary to ensure higher level of transparency and consistency in applying “user is to pay” principle in infrastructure charging. Through developing relative principles for common

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

48

Christina Nikolova

approach to charging it is possible to solve many of the problems in railway transport and to encourage transport service providing. The measures for improving the use of existing infrastructure related to shift towards more ecological modes of transport are necessary, as well as using economic instruments for decreasing oil consumption and pollution. The main objective is to increase the efficiency and sustainability of transport system and to encourage competitiveness of the European economy. The necessity of Common framework for public transport is defined on European level (Commission of European Communities (CEC), 1998). This necessity is imposed because of the fact that infrastructure charges influence the competitive conditions on the internal market, they are related to the provision of market access and they work on the development of the international transport. The only approach responding to the requirements is the marginal social costs pricing, e.g. the users are to pay the costs – internal and external – they cause in the process of using the infrastructure. The application of this approach will give incentives to the users for decreasing total costs while individual profit will be maximized. Thus, the economic and social welfare will be maximized. The following basic principles should be considered in applying the common approach: -

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

-

Developing common requirements in determining infrastructure charges; Infrastructure charges should be based on “user is to pay” principle according to which all users should bear the costs they cause; The charges should reflect the level and the character of the infrastructure use. They should be related directly to the costs and caused externalities (social and ecological); In order to lessen the distortions in competition different charges for one and the same infrastructure could be applied only when there exist significant differences in infrastructure costs and discrimination to the certain users is not allowed; The infrastructure charging system in transport should provide high efficiency and ecologically acceptable level of the use of infrastructure.

The application of these principles in railway transport is related to the estimation of marginal social costs for the use of railway infrastructure. The main accent is laid on the identifying of costs. With regard to this the costs drivers should be determined and subsequently they should be examined and categorized with the purpose of their further allocation. As a result the amount of infrastructure costs per one tonne kilometer respectively per transport unit could be estimated and thus infrastructure charges could be determined. Except the costs reflected in the charges (internal costs), there exist costs that are not paid by the users but which influence the external sides for the process of transportation. A part of these costs are marginal ones. The procedures for cost allocation depend on the estimation approach. The costs allocation is usually carried out through theoretical and empirical analysis with the use of different factors of influence (LIBERAIL, 2001). For example: - the caused costs because of delays could be included in specified forfeits. For the railway infrastructure the scarcity of capacities is possible. In such a case it is necessary to study the capacity demand by network sections and to define measures for capacity enhancement. It is possible also the approach “speed-traffic volume” to be applied. Within

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

49

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

this approach for different sections there exist different travel times depending on traffic volume; - with regard to environmental costs—the quantity of harmful emissions and their concentration should be defined. Thus, the level of pollution could be determined and after that the proportion “dose-costs”. The main purpose is to specify the influence of every polluter. In practice this proportion could be estimated for certain vehicle categories transporting goods or passengers on certain lines. Results could be generalized. This effect could further be transformed and estimated as a loss of wealth for the society (Brown & Jackson, 1998). - with regard to transport accident costs—it should be considered that they are a function of the accident risk and caused material and non-material damages. Consequently, cost drivers such as vehicle flows, speed, infrastructure category, vehicle parameters and skills of the driver are used for risk’ and costs’ allocation. These indices could be combined in order to find out the marginal costs for different types of vehicles. The differentiation of charges which counts for the external costs will create an opportunity for balancing modal split while considering for externalities. Such a measure will lead to more efficient impact of charges on the use of the railway infrastructure. In this relation it is necessary charges for pollution, scarcities and congestions to be introduced simultaneously in all modes of transport. From the financial point of view, more efficient use of the railway infrastructure will lead to reduction in budget expenses for infrastructure, healthcare and environment as well as to direct financial benefits for lower taxes. The net effect in commercial sector will be positive and direct effect from higher infrastructure charges will be neutralized by decreasing costs for congestions, accidents and by all possible reductions in taxes given by the governments. It is possible to have a slight contraction in transport intensive sectors but the whole increase in transport prices will be very slight and the companies will adapt their supply of provisions to the new requirements. The transport operators should pay different charges for different destinations and hours in order to reflect the different level of demand by hours, type of traffic and destinations. The purpose is to account for the scarcity and to ensure the efficient use of the capacity. This is a line along which the efficiency in infrastructure use could be increased. In determining infrastructure charges the accent should be put on the cost drivers. The infrastructure costs, the environmental and congestion costs could be referred to traffic volume directly. With regard to accident costs this approach is unacceptable. On the other hand, the approaches based on taxation and specific transport charging schemes are not very accurate as they are not based on particular costs caused as a consequence of accidents. Moreover, these taxes and charges do not give signals for users to correct their behavior in compliance with the occurrence of accidents. The use of insurance policy is a partial measure for accident costs recovery. The insurances could enhance the connection between the users of the infrastructure and caused costs. The variety of applied insurance premiums and “bonus-maltus” systems could provide clear signals to transport operators and could reflect appropriately occurred costs. Consequently, the application of insurances is the most suitable mechanism as insurance companies have improved in details the criteria for estimation of the damages. The purpose is to reflect the individual characteristics and skills of the drivers, travelled distances and caused

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

50

Christina Nikolova

accidents. Deregulation of the insurance market and application of the individual criteria for estimating the accident risk gives an opportunity for precise reflection of certain costs. The internalization of the other external costs will lead to the provision of additional revenues. From the point of view of the fairness, it is desirable to use accumulated financial resources for compensation of persons suffered from accidents or for financing the measures for restraining the negative impacts in the future. If, instead of this, the finances accumulated are allocated for achieving common goals in infrastructure development, then higher levels of cost recovery could be attained. The analyses made within research in EU give the reason for conclusion that the total revenues for the transport system as whole will exceed infrastructure costs. The principles mentioned above require infrastructure charges to be defined on the base of marginal social costs for transport, i.e., on the base of variable and fixed costs for the use of infrastructure. With that end in view, it is important to specify the approaches for estimation of different external costs (High Level Group on Infrastructure Charging, Working Group 2, 1999). That being the case, the calculation of the infrastructure charges is related to defining marginal external costs. In congestion costs estimation three main indices are usually used. These are valuation of travel time, proportion “travel time – service demand” and demand function (Doll & Jansson, 2005). Valuation of travel time could be achieved in several ways. First of all, it is possible to think about the value of time for the society as whole. In such a case the distinction between the work time and leisure time spent for transportation purposes should be made. The wage rate per hour could reflect the output produced for the time in question. Regarding leisure time, study of willingness to pay for actual or hypothetical choices (stated preference) is usual approach. The value of private travels could be estimated in usage of neoclassical model for utility maximization under budget restrictions. For business travels particular aspects of productivity per person are considered, too. In this field the additional research is necessary in order to clarify the driving factors of travel time values. The optimal level of congestion charges is fixed by the cross point of the user costs’ curve and demand curve. It is defined in this way because the demand for access to the railway infrastructure will react to internalization of these costs and, therefore, the costs will fall. Like this demand function defines the proportion between actual external costs and equilibrium charges for congestions. The speed-flow function describes a physical relation between the number of vehicles competing for railway capacity in a particular time segment and the resulting travel speed. The quotient “value of time – the travel speed” could be described through different functions. In this instance the most important thing is to account for the share of demand, which is choked off from the congested link and will be realized at a different time or on an alternative route. The major driving factors of traffic congestion are the costs of individual users which go up as more users are on track. Therefore, if user charges are set equal to actual external congestion costs, demand would fall and with it the external congestion costs. On the European level a number of case studies regarding congestion costs have been carried out mostly on the road transport (Quinet & Vickerman, 2004). However, in railway transport there is not enough research. Something more, it is not clear whether these costs should be considered as external when the question is about scheduled carriages. For example, congestion costs caused from one operator to another are external but it is

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

51

controversial point whether delays in presence of only one operator should be deemed as considerable for pricing or we could accept that they are included in prices. Taking into account this, as well as on the base of the actual use of railway infrastructure it could be concluded that congestion costs should be taken in consideration in the process of determining infrastructure charges. Consequently, a further in-depth study is necessary to be conducted for the purposes of their internalization. Environmental costs are related to elimination of negative effects such as noise, air and soil pollution, greenhouse gases, etc. The fact that the question is not about goods or services, hampers the quantification of these costs and therefore, different methods for quantifications take place as for instance: -

-

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

-

Surrogate market – the costs for travel, when users take advantage of leisure facilities, are used for the purpose of cost estimation. Other opportunity is to use certain goods depending on their exposure to pollution or noise. Thus, the estimation of these goods presents their ecological value. Environmental costs could be measured by estimating the amount which the members of society are ready to pay for the negative effects to be diminished or eliminated. There are some difficulties regarding the application of the last mentioned approach related to the sensitivity of the individuals to environmental factors; Direct methods – these are related to the users’ willingness to pay for negative effects to be diminished or their willingness to accept compensations. In this case difficulties are in developing questionnaires and receiving reliable answers as well as psychological prejudices regarding lower level of willingness to pay for elimination of negative effects than for acceptance of compensations for continuing bearing them; Indirect methods – they are applicable towards environmental and damage costs. Two stages could be outlined for these approaches. The first one is technical and it 2 aims at quantifying negative environmental results . Second one aims at valuation of damages through market prices of damaged goods as well as through costs for 3 4 elimination of the damages or through other subjective measures . The application of these methods is hampered by the lack of data and uncertainties related to the value of environmental damages.

Assessing negative effects is difficult, yet but there exist some case studies which were carried out in this field (Bickel et al., 2003). They attempt to provide quantification of the environmental costs. For example, the noise could be measured depending on its duration and the sensitivity of human ears. Usually the noise impacts are estimated through the decrease of building prices in noisy regions. The noise influences also the human’s health as per its impact on heart system and sleep, which are elements of a total state of the human’s health. The air pollution could be measured by the quantities of emissions from vehicles. However, it is more difficult to figure out their impact on the human beings, on animals and on buildings. Nevertheless, there exist significant discrepancies in determining the long term effect on the 2

For example: in the case of air pollution, the frequency and the significance of the impacts on the human health and building damages should be estimated. 3 These are costs for medical treatment of sufferers etc.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

52

Christina Nikolova

human’s health. The valuation of environmental costs include the direct costs for medical treatment, costs for eliminating negative effects of the emissions and estimation of the users’ willingness to pay for avoiding damages (Working Group 2, 1999). The external accident costs are inherent and ones of the highest in the transport sector. Their value depends first and foremost on the number of victims in accidents as well as on the value of human life and caused damages. The value of human life could be measured through the estimation of human capital and accounting for decrease in output because of damages suffered. It is possible also to estimate the willingness to pay for carriages performed under higher risk. The research in this field takes into consideration not only the value of human life but also the value of damages suffered by humans and their properties as well as losses in output as a consequence of the absences of medical certificates. The value of damages includes direct costs (e.g., costs for medical treatment, transportation of sufferers etc), indirect costs (loss of output) and subjective costs (estimation of pain and suffering). Therefore, it is not possible to define to which extent the external accident costs are covered by transport insurances. Besides, there still are some hesitations over the extent to which these external costs are related to the traffic volume respectively vehicle flow and speed (Working Group 3, 1999). All this gives reasons to outline further needs of research in the field of interrelation between the congestions and the number of accidents. The opportunities for internalization of external costs are used in limits within individual countries throughout European Union. Congestion charges in the big cities and fuel taxes are introduced, they are designed to cover caused congestion and environmental costs but there still exist necessity for in-depth study of the opportunities for internalization, as well as for assessment of reasonable approaches for measuring external costs. The revenues of these charges are meant to be used for funding future investments. Within the EU a common approach for measuring marginal external costs is not fixed but an issue of great importance is to define the way of proper application of different scientific research. Depending on the type of the cost function the opportunities for internalization of external costs could be estimated. If the cost function is declining, as it is in railway transport, then the marginal social costs have not reached their minimum and there are opportunities for progressive economy of scale, e.g., at 1 % increase in traffic volume there is less than 1 %t increase in costs. Consequently, it is possible to use mark-ups and to internalize external costs without significant decrease in revenues. And vice versa, if the cost function is growing, e.g., for 1 % rise in traffic volume there is more than 1 % increase in costs and then there will be regressive economy of scale. In this case, internalizing external costs will induce significant increase in charges and will influence the use of transport infrastructure. Railway transport has lower external costs than the other modes of transport, so differentiation of charges accounting for the externalities will not distort the market. Such a measure will lead to more purposeful influence of charges on the use of infrastructure, as well as on the users of different modes of transport. In this connection, the introduction of environmental charges is necessary simultaneously with certain charges in road transport. From the economic point of view, the long run effect of the application of such charges will have slight or indirect influence on the GDP growth, but it will enable receiving subsidiary benefits through increase in revenues. Thus, more accurate base for comparing 4

For instance: value of life.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

53

returns of investments will be achieved and the conditions for private investment will be improved. The elaboration of direct infrastructure charges will support the accurate estimation of the costs and the revenues related to every type of services. In this way the services performed by different modes of transport will provide an economic profit. On the other hand, the internalization of the environmental costs will improve the ecological efficiency, i.e., when charges reflect the environmental costs, then the level of harmful emissions would decline. Therefore, from the point of view of social efficiency, the welfare of the society, as a whole, will increase unlike the traffic volume. From the financial point of view, more effective use of railway transport system will bring about decrease in governmental expenditures for infrastructure, healthcare and environment. The net effect in commercial sector will be positive, as well as the direct effect from higher charges will be neutralized by the decrease in congestion costs and by all general reductions in taxes. There could be a slight shrinkage in transport consuming sectors where the share of transport costs is higher, but the whole increase in transport costs will be very slow and the companies will have enough time to react by adapting their provisions.

IX. Key Issues for Sustainable Development of Railways in EU

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The future development of the railway infrastructure aims to favor the development of services tendered by the railway operators in domestic and international markets and it is mainly oriented to the revival and improvement of the technical and operational parameters of the railway tracks, catenary, telecommunication and signaling equipments along the crossborder transport axes according to the Trans-European Network (TEN) policy; Agreement for High Speed railway lines (AGC); Agreement for Railway Lines for Combined Transport (AGTC); Trans - European Railway Project (TER). The future railway infrastructure improvement will be focused on: • • • • • • • •

achievement of the EU quality standards and ensuring interoperability through the planned investments and especially speed restrictions removal; increasing of the infrastructure productivity in order to reduce the infrastructure costs (respectively reductions in the infrastructure charges); introduction of the best practices in the field of contracting with the operators as to develop well adapted time-tables (especially in the new member states); improvement of the railway links with the neighbor countries; reduction of the time needed for border crossings; improvements in the railway stations attractiveness; introduction of appropriate infrastructures and provision of conditions for combined transport; and keeping high quality standard related to the railway network safety.

Competition between companies (licensed operators) can be promoted by ensuring that the routes for which they are granted overlap sufficiently to encourage competition for patronage on common sections of route. This form of competition makes it possible to some degree to organize supply, and limits anti-competitive or chaotic operating practices, so long

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

54

Christina Nikolova

as there are regulatory bodies to prevent such cases. Competition between modes can be effective where demand is dense and varied, as exemplified by the role of privately operated services. Some flexibility towards the introduction of new categories of services at higher prices may be a means to reconciling the maintenance of a basic low fare with the provision of adequate total capacity and a sufficiently varied range of price/quality combinations to meet demand. Within regulated systems, such as railways, this could arise either by design or by default.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

1. Sustainability and Transport Policy To be effective, transport policy must satisfy three main requirements (Munasinghe, 1993). First, it has to ensure that a continuing capability exists to support an improved standard of services. This corresponds to the concept of economic and financial sustainability. Second, it must generate the greatest possible improvement in the general quality of life, and not merely an increase in traded goods. This relates to the concept of environmental and ecological sustainability. Third, the benefits that transport produces must be equitably shared by all sections of the community. This is a social sustainability. Economic, social and environmental sustainability are often mutually reinforcing. Railway transport system that falls into disrepair because it is economically unsustainable fails to serve the needs of the clients and often have environmentally damaging consequences. Hence, there are some policy instruments which serve all of the dimensions of sustainability in a synergic way, generating “win-win” solutions. These include measures aiming to improve asset maintenance, charging for external effects, technical efficiency of supply, safety, contract design, and administration. Increased mobility in the EU counties, particularly private motorized mobility, typically increases measured GDP but damages the environment. Global sourcing of manufacturing industry and just-in-time logistics reduce product costs. However, expenditures on railway transport tend to increase as many more goods are transported over longer distances. Shifts to movement by faster modes (air) or in smaller batches with greater flexibility in frequency of schedule and variety of routes (road) also have potentially adverse environmental implications. More efficient production of transport services in a competitive framework may involve loss of jobs, imposing some social costs and restructuring of prices and services which may hurt some users.

1.1. Economic Sustainability—Creating Incentives for Efficient Response to Needs Establishing a sound economic base is fundamental for sustainability. Investments in railway infrastructure should thus continue to be subject to rigorous cost-benefit analysis, expediently expanded to include environmental externalities. The need of economic justification applies not only to infrastructure, which typically accounts for only between 25 % and 50 % of the value of total capital stock employed in transport and contributes only about 5 % to the total cost of provision of transport services, but also to decisions on vehicle fleet purchase and use, whether in the public or private sector. Ensuring the long-term sustainability of facilities requires an adequate maintenance of capital assets. In infrastructure, this is hampered by inadequate maintenance budgeting and follow up, accentuated by governments taking the “soft” option of deferring maintenance during a debt crisis. In railway transport service supply, regulated prices have often been set at levels that are too low to

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe

55

provide for the adequate maintenance of the equipment. The costs of maintaining excessively railway network and the subsidized operations of poorly managed public infrastructure company also frequently impose unsustainable fiscal burdens. This is not all that is implied by economic sustainability. Changes in the global economy have altered the nature of the demand that economic development makes on railway transport. To take advantage of the benefits of global trade in manufacturing goods, railways must be capable of providing freight services that are fast, reliable and, above all, flexible in response to the user needs.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

1.2. Environmental Sustainability—Promoting More Livable Settlements and Reducing External Effects Demands for greater transport flexibility have increased dependence on the road transport, which tends to raise aggregate energy consumption and generate air pollution and to have other adverse effects on the environment which, though not always cumulative and irreversible, are nevertheless not sustainable in the sense that they do not represent chosen outcomes. In practice, however, these adverse environmental (and social) impacts are very difficult to reverse once activity locations and personal lifestyles have been arranged to accommodate a high level of road transport dependency. The challenge is to devise a transport policy ensuring that the actual outcomes are chosen, rather than being the unintended and unforeseen consequences of the policies adopted. Viewing transport within the general perspective of sustainable development yields some immediate insights on this process. The weight placed on the various components of the general quality of life varies, of course, with the country and every country must ultimately define its own path of development. Whatever the preferred balance, increasing economic sustainability can always advance environmentally sustainable development, but does not necessarily do so. Failing to incorporate environmental considerations in the assessment of projects and policies is what creates the “sustainability gap.” The policy challenge is to recognize the trade-offs and to devise instruments that will prevent the sustainability gap from developing. With regard to this on the European level the Marco Polo program is a means for achieving the main goals related to modal shift. Marco Polo is the European Union's funding programme for projects which shift freight transport from the road to sea, rail and inland waterways. This means fewer trucks on the road and thus less congestion, less pollution, and more reliable and efficient transport of goods.

1.3. Social Sustainability—Reducing Poverty In rural areas, the poor are mainly dependent for their livelihood on their ability to produce and market agricultural products. Increasing access to traded inputs (for example, fertilizers and equipment) and making it possible to transport agricultural products to distant markets are means whereby cash cropping can replace subsistence farming. This transformation will also facilitate the development of non-agricultural activities in rural areas. Inadequate provision for vehicles can also be very costly. In urban areas, the principal resource of the poor is their labor. Adequate and affordable transport to work is, therefore, crucial. In both, urban and rural areas, anything that limits

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

56

Christina Nikolova

“basic” public transport provision, or makes it more expensive, will be particularly damaging to the poor. The ultimate damage that can be done to the poor is the elimination of either the home (resettlement) or the job (redundancy), which can also happen as a by-product of transport change. Central to these problems is the failure to provide or maintain those activities and services that are most critical in ensuring that people have access to markets, employment and social facilities.

2. Sustainability Requires Policy and Institutional Reform The challenge is to define the strategy that governments need to adopt to increase sustainability of railway transport. Generally, it is recognized that the problems are very diverse. Infrastructure and basic accessibility deficiency tends to be a more dominant problem in railway subsector, while service quality deficiency tends to be in the background. The policy and institutional changes that are necessary to enhance economic, environmental and social sustainability should be considered. The new focus does not vitiate the continuing importance of efficient transport to trade, mobility and, hence, to economic growth, or its contribution to the achievement of environmental and social objectives. However, it does highlight the fact that the traditional emphasis on public sector operation and regulation has often failed to make that contribution in a continuing, sustainable way. This has been partly a human resource problem, as governments do not possess adequate skills for carrying out the planning and control tasks required, and partly an institutional problem so far as governments continue to rely on mechanisms that make unrealistic demands on human resources and motivations. Therefore, we consider the new demands that the emphasis on sustainability place on governments.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

2.1. Criteria for Strategic Development Most transport improvements are designed to reduce transport costs. This can often have a beneficial impact on the environment, for example, by reducing fuel consumption and air pollution. Almost all transport projects are subject to at least a partial assessment. The increasingly stringent application of these procedures has encouraged the design of projects that are sensitive to environmental concerns and that mitigate any directly adverse environmental impact. Furthermore, transport projects, or project components, are increasingly going beyond avoiding direct harm and focusing more positively on environmental improvements. These projects address the immediate and direct effects of transport. In such cases the fundamental question is what should be the main accent in transport policy. Strategies for railway transport are, in principle, based on a full cost-benefit analysis incorporating both transport and environmental objectives (World Bank, 2008). They include short-term management and pricing instruments and long-term strategic instruments as the context for identifying investment actions. Still, further research will be required to identify critical environmental effects and to determine the efficacy of different interventions. In the interim, an appropriate framework would include:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Infrastructure Market in Europe •

•

57

Strategic and structural actions, including the creation of economic incentives for environmentally sensitive behavior, would be of a high priority. Although their effects may be slow to come to fruition and may be politically and administratively difficult to be implemented, they are the critical and pervasive basis for environmentally sustainable transport development. Within that strategic framework, some priority problems can be identified where the benefits of making improvements are judged to be very high, particularly because they are seriously life and health threatening. The most appropriate technology should be selected on the basis of relative costeffectiveness, in the context of the main problems in the railway transport, taking into account what the companies can afford and effectively implement. This often means that the actions with the highest priority are not those attempting to impose “state-of-the-art” standards or technologies but those that make more immediate, implementable changes in the way in which existing equipment is used.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

X. Conclusion Delivering sustainable railways in Europe is the most positive statement about the growth and development of the EU railways. Passengers demand railways that are reliable and represent value for money, that are comfortable, accessible and easy to use. The public as a whole wants railways that contribute to economic growth and help people to meet the environmental challenges ahead. The railways can only meet these goals if they have the capacity to carry the passengers who want to use them. It should be considered also that the railways can never afford to take safety for granted. The railways have consistently to improve their safety records. Yet, there are lines in Europe on which reliability is unacceptable. And there are still too many delays caused by basic problems with track and signals within some of the EU member states. On some lines, provision of additional capacity will require new infrastructure. These increases in capacity make a good start on tackling crowding for some of the busiest services, while enabling the railways to accommodate growth. And further improvements will follow – this is a long-term strategy, not one that stops after just few years. For regional and rural lines, the focus is also on growth. Regarding the quality of services it is not enough simply to provide railways of sufficient size – they must also be accessible and easy to use. The fares system should be simplified, so passengers have greater confidence that they are being sold the right ticket for them. Measures to be undertaken are: to cut queues at ticket offices, to assist passengers and to enhance their sense of security. The future investments should be directed to medium-sized stations that are run down or lack basic facilities and to address the areas that the railways have neglected for too long. The main measure is to improve access to stations. The next important measure to be undertaken is delivering rail’s environmental potential. The biggest contribution railways can make to the environment is to expand their capacity to accommodate those who want to make ‘green’ travel choices and to provide the quality of service necessary to retain their customers. This is true for freight as well as for passengers. Railways must also reduce their own carbon emissions. But train operators should

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

58

Christina Nikolova

undertake efficiency measures that pay for themselves via reduced fuel bills. Train operators must take responsibility themselves. Sustainability considerations presuppose stable funding. Ambitions of this scale require sustained investment to match. The governments in the EU are providing this, supported by the European funds. The development of a long-term strategy for sustainable railway transport in Europe is a challenge because trains ordered now will still be in service in 30 years’ time, and other assets will last even longer. But it is also challenging because it is impossible to accurately forecast demand that far into the future. Some cities and regions will grow faster than others. People and firms are likely to respond to the challenge of global warming by changing travel patterns and ways of working. However, the pace of technological change is unpredictable. Forecasts proved to be wrong before, and any strategy that tried to build a rigid investment programme based on fixed long-term forecasts would inevitably turn out to be wrong again. To overcome this challenge, the guiding principles in a long-term strategy should be: •

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

•

To invest where there are challenges now, in ways that offer the flexibility to cope with an uncertain future; and To put in hand the right preparatory work so that, as the future becomes clearer, the necessary investments can be made at the right time.

Sustainability demands a broader look at priorities for the railways alongside other modes, to find the best balance between the needs of the economy, society and the environment. In this regard, the main ambitions are for railways to contribute to the economic success of the entire EU by enabling more people to travel in a way that minimizes the environmental impact; railways that are flexible enough to adapt and respond to social changes, protecting the networks and improving their ability to operate for longer in the day and more consistently over the work week; and railways that are easy and accessible to use and that are built on stable foundations of safety, reliability and sound finance. These, namely, should be the main points in the strategy: to deliver a sustainable, modern railway transport.

References Baldwin, R. & Cave, M. (1999). Understanding Regulation. Oxford: University Press. Baumstark, L. (1998). France: Introductory report. In: ECMT Round Table 107. User charges for railway infrastructure. Paris: OESD Publications Service, pp.47-100. Betancor, O. et al. (2005). Operating costs. In: C Nash and B. Matthews (eds). Measuring the Marginal Social Costs of Transport. Amsterdam, Oxford: Elsevier, pp. 85-123. Bickel, P. et al. (2003). Environmental Marginal Cost Case Studies. Deliverable 11, Unification of accounts and marginal costs for Transport Efficiency, University of Leeds. Bitzan, J. (2003). Railroad Costs and Competition: The Implications of Introducing Competition to Railroad Network. Journal of Transport Economics and Policy [Online], 37(2), [Accessed 2nd July 2008], pp. 201-225. Available from World Wide Web: http://0www.ingentaconnect.com.wam.leeds.ac.uk/content/lse/jtep/2003/

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Railway Infrastructure Market in Europe

59

Bosshe, M. et al., (2005). Measuring Marginal Social Cost: Methods, Transferability. In: C Nash and B. Matthews (eds.). Measuring the Marginal Social Costs of Transport. Amsterdam, Oxford: Elsevier, pp. 287-310. Brown, C. and Jackson P. (1998). Public sector economics. 4th edition. Oxford: Blackwell. Brown, M.(2001). The Future of Railways in Central and Eastern Europe. In: What Role for the Railways in Eastern Europe? European Conference of Ministers of Transport, OECD. Commission of European Communities (1995). Towards Fair and Efficient Pricing in Transport [Online].[Accessed on 28th June 2008]. Brussels: CEC. Available from World Wide Web: http://europa.eu.int/en/record/green/gp003en.pdf Commission of European Communities (1996). White paper: A Strategy for Revitalizing Community’s Railways [Online]. [Accessed 27th June 2008]. Available from World Wide Web: http://europa.eu.int/en/record/white/rail967/wp9607en.pdf Commission of European Communities (1998). Fair Payment for Infrastructure Use: A Phased Approach to a Common Transport Infrastructure Charging Framework in the EU [Online]. [Accessed on 28th June 2008]. Brussels: CEC. Available from World Wide Web: < http://europa.eu.int/comm/transport/infr-charging/library/lb98-en.pdf Commission of European Communities (2001). White Paper: European Transport Policy for 2010: Time to Decide [Online]. [Accessed 3rd July 2008]. B8russels: CEC Available from World Wide Web: Kessel & Partner (2004). Study of Infrastructure Capacity Reserves for Combined Transport by 2015 [Online], [Accessed 3rd July 2008]. Available from World Wide Web: 0.5σ (ON intervals) varied from 1 to 10– 12s. The shortest ON intervals were the most probable ones. Our results did not reveal any special differences in intermittency characteristics between EL (both AC and DC) and EMU. Then, we can conclude that there are common features in magnetic field encountered both in DC- and AC-powered railway systems. These magnetic field features might be responsible for similar adverse health effects, in particular in the cardiovascular system (increase of myocardial infarctions, etc.) found in locomotive engineers, powered both by DC and AC currents. Our special animal experiment [57] clarified plausible biological pathways for such health effects. The common and peculiar features of magnetic field spectra encountered in Swiss AC and Russian DC railways helped to develop “typical railway” patterns. Such “railway” magnetic field patterns have been simulated in laboratory studies on biological responses in mice. Found results indicate that magnetic field with highly variable complex spectra, typical for railway magnetic environments, can be one of the factors increasing the loading for regulatory mechanisms of cardiovascular system. This increased loading could lead to the development of classical pathogenesis states such as myocardial ischemia, arrhythmia and others, which lead to increased risk of cardio-vascular catastrophes [57]. Recognition of the main spectral features and other peculiarities of magnetic field recorded in Russian DC and Swiss AC trains connected with different route conditions can be useful in the identification of the sources of these specific features. It will allow developing design-related preventive measures to diminish the health-hazardous potential of magnetic field. In particular, the similar geometrical regularity of polarization structure (predominantly circular polarization in vertical planes) of magnetic fields, found in Swiss and Russian electric locomotives, could be indicative of common peculiar constructing features responsible for this kind of polarization. In the case of further supporting evidence for potential

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway-Generated Magnetic Field: Environmental Aspects

127

cardiovascular risks of circularly polarized magnetic fields, it could be possible to elaborate cost-effective preventive measures to diminish this component “at the source”.

Appendix A.1. Electromotive Forces Ψ Produced by the Time Change of Magnetic Field Intensity within the Human Body Let us consider in more detail the first term in the right-hand side of (3.1.4):

dB ⎛∂ B ∂ B G⎞ = −S cos α ⎜ + G V⎟ = ⎝∂ t ∂r ⎠ dt , ∂ Bc G ∂ Bv G ⎛ ∂ Bv ⎞ −S cos α v ⎜ ⎟ − S cos α c G V − S cos α v G V ⎝ ∂ t ⎠ ∂r ∂r

Ψ1 = −S cos α

(A.1.1)

G where V is the velocity of the closed loop (which in most cases coincides with the velocity of the body) relative to the system of coordinates in which magnetic field (MF) is considered. The first term in the right-hand side of (A.1.1) reflects the influence of variable part of MF:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

⎛ ∂ Bmv ⎞ ⎛ ∂ Bnv ⎞ ⎛ ∂ Bv ⎞ Ψ11 = −S cos α⎜ ⎟= ⎟ − S cos α m ⎜ ⎟ = −S cos α n ⎜ ⎝ ∂ t ⎠ ⎝ ∂ t ⎠ ⎝ ∂ t⎠ l

p

i =1

k =1

(A.1.2)

S∑ ω ni Bnisin(ω ni t − ϕ ni ) cos α ni + S ∑ ω mk Bmk sin(ω mk t − ϕ mk ) cos α mk where we assumed that l

Bnv = ∑ Bni cos(ω ni t − ϕ ni ) i =1

, Bmv =

p

∑ Bmi cos(ω mi t − ϕ mi )

,

(A.1.3)

k =1

and considered that the angle α can be different for different harmonics: α ni for natural geomagnetic field and α mk for man-made MF. Let us consider the second term of (A.1.1), i.e. the case in which ψ is produced by the movement of the loop in presence of a space gradient in constant part of MF:

Ψ12 = −S cos α

∂ Bc G ∂ Bnc G ∂ Bmc G G V = −S cos α n G V − S cos α m G V ∂r ∂r ∂r

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

(A.1.4)

128

N. G. Ptitsyna, G. Villoresi and Y. A. Kopytenko In this case the contribution of geomagnetic field Bnc , even in the case of a body

moving with great velocity (for instance 8.103 ms-1 of spacecrafts) is very small due to the G very small value of ∂ Bnc ∂ r , while the contribution of man-made MF in (A.1.4) can be much more important. The characteristic frequency of this Ψ will be ν mc ≈ V / rmc ,

where rmc is the characteristic distance for man-made MF to change by a factor e. For example, if a person runs with V ≈ 4 m / s and rmc ≈ 2 m , then ν mc ≈ 2 Hz which can be near to self frequencies in the human body. This is the case in which the resonance interaction can be realized. Let us assume that x ∂ Bc G ∂ Bmc G ≈ = V V G G ∑ bcr cos(ω cr t − ϕ cr ) ∂r ∂r r =1

(A.1.5)

and introduce this value in (A.1.4), thus obtaining for the influence of about constant part of MF: x

Ψ12 = −S ∑ b cr cos α cr cos(ω cr t − ϕ cr )

(A.1.6)

r =1

in which it is taken into account that the angles α cr can be different for different harmonics

ω cr .

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Now we consider the last term in (A.1.1):

∂B G ∂ Bnv G ∂ Bmv G Ψ13 = −S cos α Gv V = −S cos α n G V − S cos α m G V ∂r ∂r ∂r

(A.1.7)

which is determined by the movement of the closed loop in presence of a gradient in the variable part of MF. Also in this case the contribution of natural geomagnetic field can be neglected, then the only relevant effect is due to the gradient of variable part of man-made MF on moving body. Assuming that, as in (A.1.3), p Bv ≈ Bmv = Bmk cos(ω mk t − ϕ mk ) and that, as in A.1.5), k =1

∑

(

)

q ∂ Bmk G G V = ∑ b mkj cos ω mkjt − ϕ mkj , ∂r j=1 we obtain:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

(A.1.8)

Railway-Generated Magnetic Field: Environmental Aspects p q

(

129

)

Ψ13v = −S ∑ ∑ b mkj cos α kj cos(ω mk t − ϕ mk ) cos ω mkjt − ϕ mkj = k =1 j =1

p q

⎧

−S ∑ ∑ b mkj cos α kj ⎨cos⎡⎢⎛⎜⎝ ω mk + ω mkj ⎞⎟⎠ t − ⎛⎜⎝ ϕ mk + ϕ mkj ⎞⎟⎠ ⎤⎥ + ⎩

k = 1 j =1

[(

⎣

) (

cos ω mk − ω mkj t − ϕ mk − ϕ mkj

⎦

(A.1.9)

)]}.

Eq. (A.1.9) shows that in this case there is interference of harmonics contained in Bmv , according to (A.1.3), and harmonics arising from the body’s movement, according to (A.1.8). As a result of this interference the induced Ψ in the closed loop is formed by harmonics at frequencies ω mk ± ω mkj and amplitudes Sb mkj cosα kj .

A.2. Induced Ψ Caused by Changes in Cross-Section of Closed Loops in the Presence of MF of Natural and Man-Made Origin We consider here the second term in (3.1.4):

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Ψ2 = − B cos α

dS ∂S ∂S G = − B cos α − B cos α G V . dt ∂t ∂r

(A.2.1)

The first part reflects the ψ produced by the time variations of the effective cross section of closed loops (for example the rhythmical changes in breathing, in heart beating, etc.) in natural and man-made MF. The second part reflects the situation when the change in S is caused by the change in position of closed loop with velocity V (e.g. the case of a moving body). In the first part of (A.2.1), due to the change of S with time,

G G G ∂S G ∂S Ψ21 = − Bnc + Bmc cos α − Bnv + Bmv cos α . ∂t ∂t

(A.2.2)

the contribution of both constant MF, of natural and man-made origin, are present and the Ψ .strength is proportional to the total intensity of natural and man-made MF. Supposing that the change of S with time can be described by q

(

)

S( t ) = ∑ S j cos ω sjt − ϕ sj , j =1

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

(A.2.3)

130

N. G. Ptitsyna, G. Villoresi and Y. A. Kopytenko

the contribution of the constant MF to Ψ21 will be q G G Ψ21c = Bnc + Bmc cosα ∑ ω sjS jsin ω sjt − ϕ sj .

(

j =1

)

(A.2.4)

The about constant part of man-made MF in specific places, for instance inside public transport, can be much bigger than natural field and it can generate much bigger Ψ in the rhythmically changing closed loops according to (A.2.4). Let us assume that, for the effect of changes of S with time in presence of variable MF, the variable MF can be represented as in (A.1.3) and the change of S by (A.2.3). In this case we obtain: q

l

(

)

Ψ21v = ∑ ∑ Bniω sjS j cos α ij cos(ω ni t − ϕ ni )sin ω sjt − ϕ sj + i =1 j =1

p q

∑ ∑ Bmk ω sjS j cos α kj cos(ω mk t − ϕ mk )sin(ω sjt − ϕ sj ) =

k =1 j =1

q

∑ ∑ Bniω sjS j cos α ij{sin[(ω ni + ω sj )t − (ϕ ni + ϕ sj )] + l

i =1 j=1

[(

) (

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

sin ω sj − ω ni t − ϕ sj − ϕ ni

)]} +

{ [(

p q

(A.2.5)

) (

)]

+ ∑ ∑ Bmk ω sjS j cos α kj sin ω mk + ω sj t − ϕ mk + ϕ sj + k = 1 j =1

[(

) (

sin ω sj − ω mk t − ϕ sj − ϕ mk

)]}.

Eq. (A.2.5) reflects the nonlinear interference of different harmonics of both variable natural and man-made MF, with harmonics describing the change in time of closed loops. This interference will produce new harmonics at frequencies ω ni ± ω sj and ω mk ± ω sj with amplitudes BniωsjS j cosαij and

BmkωsjS j cosαkj .

We consider now the second term of the right-hand side of (A.2.1), which reflects the situation when the change of S is caused by the change in position of closed loop in space

G

G

with velocity V = dr / dt . This change of S can be rhythmical as, for example, during walking, running and exercising:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway-Generated Magnetic Field: Environmental Aspects

131

G G ∂S G ∂S G Ψ22 = − Bnc + Bmc cos α G V − Bnv + Bmv cos α G V . ∂r ∂r G ∂S Assuming that G V can be represented as ∂r x ∂S G G V = ∑ s j cos ω vjt − ϕ vj ∂r j =1

(

(A.2.6)

)

(A.2.7)

in the presence of about constant parts of natural (Bnc) and man-made (Bmc) MF, the contribution of Bc to Ψ22 can be written as x

(

x

)

(

)

Ψ22c = −Bnc cos α nc ∑ s j cos ω vjt − ϕ vj − Bmc cos α mc ∑ s j cos ω vjt − ϕ vj . (A.2.8) j=1

j=1

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Here we considered that the angles α nc , α mc between the normal to the closed loop and the direction of MF can be different for natural and man-made fields. In (A.2.8) the first term in the right-hand side reflects mostly a natural situation, to which people are adjusted by longterm evolution. The second term, which reflects the influence of about constant part of manmade MF, can be more important, especially in urban and technological areas, including public transport. We consider now the second term in the right-hand side of (A.2.6), i.e. the influence of the variable part of MF on a changing S caused by the movement of the loop, by taking into account (A.2.7) and (A.1.3) for the variable part of MF. We obtain: l

q

(

)

Ψ22 v = − ∑ ∑ Bni s j cos α ni cos(ω ni t − ϕ ni ) cos ω vjt − ϕ vj − i =1 j =1

p q

∑ ∑ Bmk s j cos α mk cos(ω mk t − ϕ mk ) cos(ω vjt − ϕvj) =

k =1 j =1

l

q

{cos[(ω )t − ( − )]} +

(1 / 2)∑ ∑ Bnis j cos α ni

[(

i =1 j=1

cos ω vj − ω ni

ϕ vj

) (

ni + ω vj t − ϕ ni + ϕ vj

)] +

ϕ ni

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

(A.2.9)

132

N. G. Ptitsyna, G. Villoresi and Y. A. Kopytenko l

q

{ [( )t − (ϕ − ϕ )]}.

) (

)]

+(1 / 2) ∑ ∑ Bmk s j cos α mk cos ω mk + ω vj t − ϕ mk + ϕ vj +

[(

k = 1 j =1

cos ω vj − ω mk

vj

mk

where the two parts reflect the influence of variable natural (n) and man-made (m) MF, respectively. Eq. (A.2.9) shows that instead of harmonics with frequencies ω nk , ω νj and

ω mk , ω νj we obtain, after the interaction of the variable MF with the rhythmical movement of body, a Ψ in moving closed loops characterized by harmonics with frequencies

ω ni ± ω vj and amplitudes (1 2) Bni s j cos α ni for natural MF and ω mk ± ω vj and

(1 2)Bmk s j cos α mk

for man-made MF.

A.3. Induced Ψ FEs in Closed Loops Caused by Changes of the Angle between the Normal to the Loop and MF Direction

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

We consider here the third term in (3.1.4), which reflects the generation of induced Ψ in closed loops caused by changes of the angle α between the normal to the loop and MF direction:

Ψ3 = − BS

∂ cos α ∂ cos α G d cos α = − BS − BS G V. ∂t ∂r dt

(A.3.1)

The first part in the right-hand side reflects the generation of induced Ψ in closed loops caused by changes with time of the angle α , and the second part the induced Ψ caused by the body movement accompanied by changes in α . We consider the first term in the right-hand side of (A.3.1):

Ψ31 = − BS

G G G G ∂ cos α ∂α ∂α = S Bnc + Bmc sinα + S Bnv + Bmv sinα . ∂t ∂t ∂t

(A.3.2)

Let suppose that the body rotates in some relatively short time period and this rotation can be characterized by frequency ω r and phase ϕ r : α = ω r t − ϕ r . In this case the first part in the right-hand side of (A.3.2) will be

G G Ψ31c = S Bnc + Bmc ω r sin(ω r t − ϕ r ).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

(A.3.3)

Railway-Generated Magnetic Field: Environmental Aspects

133

For the part in (A.3.2), which reflects the role of variable part of MF of natural and manmade origin, if we suppose again that the body in some relatively short time period rotates with frequency ω r characterized by phase ϕ r , we obtain for variable MF described by (A.1.3): l

Ψ31v = S∑ ω r Bni cos(ω ni t − ϕ ni )sin(ω r t − ϕ nr ) + i =1

(A.3.4)

p

S ∑ ω r Bmk cos(ω mk t − ϕ mk )sin(ω r t − ϕ mr ) , k =1

which can be written as

⎧⎪ l Ψ31v = Sωr 2 ⎨∑ Bni sin( (ωr + ωni ) t − (ϕnr + ϕni )) + sin( (ωr − ωni ) t − (ϕnr − ϕni )) + ⎪⎩i =1 p ⎫⎪ ∑ Bmk sin((ωr + ωmk )t − (ϕmr + ϕmk )) + sin( (ωr − ωmk )t − (ϕmr − ϕmk )) ⎬. ⎪⎭ k =1

[

]

[

(A.3.5)

]

Also in this case harmonics of induced Ψ will be generated with amplitudes

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

(1 2)Sω r Bni and frequencies ω r ± ω ni for the variable natural MF and with amplitudes (1 2)Sω r Bmk and frequencies ω r ± ω mk for the variable man-made MF.

We consider now the second term in the right-hand side of (A.3.1) which described the induced Ψ in closed loops caused by the moving of the human body with velocity

G G V = dr dt together with changing in α :

G G ∂ cos α G G G ∂ cos α G Ψ32 = −S Bnc + Bnv G V − S Bmc + Bmv G V . ∂r ∂r

(A.3.6)

The first term describes the role of about constant and variable MF of natural origin, whose gradient on the Earth’s surface is very small. This term is expected to be negligible ( Ψ32 n ≈ 0 ) even for great velocities of body movements (for instance by plane or by spacecraft). The second term shows the role of about constant and variable MF of man-made origin, that can be often characterized by rather big gradients. In such a case, if we can write: y x ∂ cos α G V A cos ω t ϕ = − + G ∑ mcj ∑ A mvj cos ω mvjt − ϕ mvj , mcj mcj ∂ r j =1 j =1

(

)

(

)

then we obtain

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

(A.3.7)

134

N. G. Ptitsyna, G. Villoresi and Y. A. Kopytenko x

(

)

Ψ32 mc = −SBmc ∑ A mcj cos ω mcjt − ϕ mcj , j =1

Ψ32 mv = −S

p

y

k =1

j=1

(A.3.8)

∑ Bmk cos(ω mk t − ϕ mk ). ∑ A mvj cos(ω mvjt − ϕ mvj ) =

{ [(

p y

) (

)]

−(1 2)S ∑ ∑ Bmk A mvj cos ω mk − ω mvj t − ϕ mk + ϕ mvj +

[(

k = 1 j =1

) (

cos ω mk − ω mvj t − ϕ mk − ϕ mvj

(A.3.9)

)]},

where for the variable part of man-made MF we took into account (A.1.3). Eq. (A.3.8) shows that, for a body moving in large MF gradient, the expected harmonics of Ψ for almost constant man-made MF, will have amplitudes

SBmcA mcj

with frequencies ω mcj and

from variable man-made MF, according to (A.3.9), will have amplitudes SBmk A mvj

with

frequencies ω mvj ± ω mk .

A.4. On the Resonance Interaction of Self Electromotive Forces Ψ in Closed Loops with Ψ Induced by MF of Natural and Man-Made Origin

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Let us suppose that in the closed loop S there are self electromotive forces Ψ as

Ψs = Ψso sin(ω s t − ϕ s ,

(A.4.1)

Expression (A.4.1) is the solution of the equation:

 = −ω 2 Ψ Ψ s s

(A.4.2)

with initial conditions Ψs ( t = 0) = − Ψso sin( ϕ s ),

 = Ψ cos(ϕ ) . Eq.(A.4.2) is Ψ s so s  , so analogous to the equation of self mechanical oscillations F = − aX, where F = mX

that

 = −ω 2 X, X s where ω

2 S

(A.4.3)

= a m . Let us remember that if, in the presence of mechanical oscillations, there

2 are external forces Fe = − A eω e sin(ω e t − ϕ e ) , then Eq. (A.4.3) will be transformed in

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway-Generated Magnetic Field: Environmental Aspects

 = −ω 2s X − A eω 2e sin(ω e t − ϕ e ), X

135 (A.4.4)

with solution for forced oscillations

X = A eω 2e sin(ω e t − ϕ e ) ω 2e − ω 2s .

(A.4.5)

If we take into account that each self frequency will have some half width Γs , the solution (A.4.5) will be transformed in

X = A eω 2e sin(ω e t − ϕ e )

(ω

)

2 2 2 e − ω s + Γs .

(A.4.6)

The resonance interaction will be important for all cases considered in previous sections. For example, let us consider the resonance interaction of self Ψ in closed loops with Ψ induced by the variable part of natural and man-made MF (see Section A.1). For this it is necessary to add to the right-hand side of Eq. (A.4.2) additional terms analogous to (A.4.4) on the basis of (A.1.2): l

d 2 Ψ11 dt 2 = −ω 2s Ψ11 − S∑ ω 3ni Bni sin(ω ni t − ϕ ni ) cos α ni + i =1

(A.4.7)

p

−S ∑ ω 3mk Bmk sin(ω mk t − ϕ mk ) cos α mk . k =1

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The solution of (A.4.7), by taking into account (A.4.6), will be l

Bnisin(ω ni t − ϕ ni ) cos α ni

i =1

ω 2ni − ω s2 + Γs2

Ψ11 = S∑ ω 3ni p

Bmk sin(ω mk t − ϕ mk ) cos α mk

k =1

ω 2mk − ω s2 + Γs2

S ∑ ω 3mk

+ (A.4.8)

.

The solution (A.4.8) differs from (A.1.2) only in the frequency region near the self frequency ω s , for which the background spectrum of induced Ψ will increase by a factor ≈ ω 2s Γ s2 , due to the resonance effect. For example, if Γs ≈ 01 . ω s , then the increase will be by about 100 times; it means that the induced Ψ at this frequency will be amplified by about 100 times by the resonance effect. If some closed loops have several self frequencies ω s1 , ω s2 ,... ω sn with half widths Γs1 , Γs2 ,... Γsn , then there will be several resonance 2 2 2 2 2 2 increases in electromotive forces Ψ , by ω s1 Γs1 , ω s2 Γs2 ,... ω sn Γsn

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

times,

136

N. G. Ptitsyna, G. Villoresi and Y. A. Kopytenko

induced by the variable parts of natural and man-made MF at self frequencies ω s1 , ω s2 ,... ω sn respectively.

Acknowledgments This research was partly supported by the European Commission (contract IC15-CT960303).

References [1] [2] [3]

[4]

[5]

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

[6] [7] [8] [9] [10]

[11]

[12]

[13]

Foster K.R.1992. Health effects of low-level electromagnetic fields: phantom or not so phantom risk? Health Phys. 62, 429-435. Baris D., Armstrong B., Deadman J. 1996. A mortality study of electrical utility workers in Quebec. Occup Environ Med. 53, 25-31. NIEHS Working Group Report, Assessment of Health Effects from Exposure to PowerLine Frequency Electric and Magnetic Fields; Editors, C.J. Portier, M.S. Wolfe; Research Triangle Park, NC 27709, NIH Publ. No. 98-3981, 1998. Van Wijngaarden E., Savitz D.A., Cai J., Loomis D., Kleckner R.C. 2000. Evidence of association between EMFs and exposed worker suicide. University of North Carolina News Release No. 147. http://www.unc.edu/news/newsserv/archives/mar00/ savitz2031500.htm. Savitz D. A., Loomis D. P. 1995. Magnetic field exposure in relation to leukemia and brain mortality among electric utility workers. Am J Epidemiol. 141, 123-134. Kavet R.I. 1996. EMF and current cancer concepts. Bioelectromagnetics. 17, 339-357. Alfredsson L., Hammar N., Karlehagen S. 1996. Cancer incidence among male railway engine-drivers and conductors in Sweden. Cancer Causes Control. 7, 377-381. Minder C.E., Pfluger D.H. 2001. Leukemia, brain tumors and exposure to ELF electromagnetic fields in Swiss railway employees. Am J Epidemiol. 153, 825-835. Floderus B., Tornqvist S., Stenlund C. 1994. Incidence of selected cancers in Swedish railway workers. Cancer Causes Control. 5, 189-194. Pfluger D., Villoresi G., Ptitsyna N., Temuriantz N., Martinyuk V., Kopytenko Y., Rasson J., Iucci N. 2000. Increased mortality for myocardial infarction in Swiss railway workers, Proc Millenium Workshop on Biological Effects of Electromagnetic Fields. Crete, Greece. 558-566. Ptitsyna N.G., Villoresi G., Kopytenko Y.A., Kudrin V.A., Tyasto M.I., Kopytenko E.A., Iucci N., Voronov P.M., Zaitsev D.B. 1996. Coronary heart diseases: an assesment of risk associated with work exposure to ultra-low frequency magnetic fields. Bioelectromagnetic.17, 436-444. Ptitsyna N.G., Y.A. Kopytenko, G. Villoresi, D.H. Pfluger, V. Ismaguilov, N. Iucci, E.A. Kopytenko, D.B. Zaitzev, P.M. Voronov, M.I Tyasto. 2003. Waveform Magnetic Field Survey in Russian DC- and Swiss AC-powered Trains: a Basis for Biologically Relevant Exposure Assessment. Bioelectromagnetics. 24, 546-556. Ptitsyna N.G., Villoresi G., Kopytenko Y.A., Tyasto M.I., Kopytenko E.A., Iucci N., Voronov P.M., Zaizev D.M. Magnetic field monitoring on DC electrified transport in

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway-Generated Magnetic Field: Environmental Aspects

[14]

[15]

[16] [17]

[18]

[19]

[20]

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

[21]

[22]

[23]

[24] [25] [26] [27]

137

Russia. In Electricity and Magnetism in Biology and Medicine. Ed. F. Bersani, Plenum Publ. Corporation, NY, 1999, 773-776. Villoresi G., Kopytenko Y., Pfluger D., Ismagyilov V., Kopytenko E., Zaitsev D., Voronov P., Ptitsyna N., Tyasto M., Iucci N., Temuriantz N.,.Martinyuk V., Rasson J. 2000. Magnetic fields generated by electric trains. Proc Millennium International Workshop on Biological Effects of Electromagnetic Fields. Crete, Greece. 95-105. Ptitsyna N., Villoresi G., Kopytenko Y., Pfluger D., Ismagyilov V., Iucci N., Tyasto M., Kopytenko E., Temuriantz N., Martinyuk V. 2000. Assessment of biologically plausible exposure parameters for complex-spectra transport magnetic fields. Proc Millennium International Workshop on Biological Effects of Electromagnetic Fields. Crete, Greece. 109-114. Tynes T., Anderson L. 1990. Electromagnetic fields and male breast cancer. Lancet. 336, 8730-8737. Balli-Antunes M., Pfluger D.H., Minder C.E. 1990. The mortality from malignancies of haematopoietic and lymphatic systems among railway engine drivers Environmetrics. 1, 121-130. Minder Ch.E., Pfluger D.H. 1993. Extremely low frequency electromagnetic field measurements (ELF-EMF) in Swiss railway engines. Radiation Protection Dosimetry. 48, 351-354. Savitz D.A., Liao D., Sastre A., Kleckner R.C., Kavet R. 1999. Magnetic Field Exposure and Cardiovascular Disease Mortality among Electric Utility Workers. Am J Epidemiol. 149, 135-142. Villoresi G., Ptitsyna N.G., Kudrin V.A., Iucci N., Health effect among engine drivers: association with occupational exposure to magnetic fields from DC electrified transport. In Electricity and Magnetism in Biology and Medicine. Ed. F. Bersani. Plenum Publ. Corporation, NY, 1999, 777-780. Whittington C.J., Podd J.V., Rapley B.R. 1996. Acute effects of 50 Hz magnetic field exposure on human visual task and cardiovascular performance. Bioelectromagnetics. 17, 131-137. Graham C., Cohen H., Cook M., Phelps J., Cercovich M., Fotopoulos S. A double blind evaluation of 60-Hz effects on human performance, physiology and subjective state. In Interaction of Biological Systems with Static and ELF Electric and Magnetic Field. Eds. L. Anderson, R. Weigel, B. Kelman., NTIS, 23rd Annual Hanford Life Sciences Symp, 1987, Springfield, VA, 471-486. Pfluger D.H., Kopytenko Y.A., Villoresi G., Ptitsyna N.G. 1999. Mortality from myocardial attacks in Swiss Federal Railways employees. In Proc 2nd Intern Conf Electromagnetic Fields and Human Health. Moscow, Russia, 312-313. Nishida A. Geomagnetic Diagnosis of the Magnetospher; Springer Verlag: NY, Heidelberg, Berlin, 1978. Bennett W.R. Jr. 1994. Cancer and power lines. Phys Today. 47, 23-29. Bracken T.D. 1993. Exposure assessment for power frequency electric and magnetic fields. Am Ind Hyg Assoc J. 54, 165-177. Kavet R.I., Banks R.S. 1986. Emerging issues in extremely low frequency electric and magnetic field health research. Environ Res. 39, 386-404.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

138

N. G. Ptitsyna, G. Villoresi and Y. A. Kopytenko

[28] Ho A.M., Fraser-Smith A.C., Villard Jr. O. 1979. Large-amplitude ULF magnetic fields produced by a rapid transit system: close-range measurements. Radio Sci. 14, 1011-1015. [29] Kazuo U. 1977. Magnetic field disturbances produced by electric railway, in Memories of the Kakioka Magnetic Field Observatory (Suppl) 7, 17-34. [30] Ptitsyna N.G., Villoresi G., Dorman L.I., Iucci N., Tyasto M.I. 1998. Natural and manmade low-frequency magnetic fields as a health hazard, Physics Uspekhi (Advances in Physics Sciences). 41, 687-709. [31] Ptitsyna N.G., Villoresi G., Kopytenko Y.A., Tyasto M.I., Kopytenko E.A., Iucci N., Voronov P.M., Zaizev D.M. Magnetic field environment in ULF range (0-10 Hz) in urban areas: man-made and natural fields. In Electricity and Magnetism in Biology and Medicine. Ed. F. Bersani. Plenum Publ. Corporation, NY, 1999. 279-282. [32] Paul H.U., Hirsch F., Studinger H., Hadrian W. In Interference of Monitors by Magnetic Fields of Power Frequency Installations and Traction Lines. 1990. CIGRE. Paris, France, 36-42. [33] Allen S.G., Blackwell R.P., Chadwick P.J., Driscoll C.M.H., Pearson A.J., Unsworth C.C., Whillock M.J. Review of Occupational Exposure to Optical Radiation and Electric and Magnetic Fields with Regard to the Proposed CEC Physical Agents Directive. National Radiological Protection Board. Chilton, Didcot, Oxon, 1988, OX11 ORQ. [34] Nakagava M., Koana T., EMF issues with Maglev in Japan. In Electricity and Magnetism in Biology and Medicine. Ed. M. Blank. Press Inc. San Francisco, Ca.1993, 264-266. [35] Dietrich F.M., Feero W.E., Papas P.N., Steiner G.A. Magnetic and Electric Field Testing of the Amtrak Northeast Corridor and New Jersey Transit/North Jersey Coast Line Rail Systes. US Department of Transportation, DOT-VNTSC-FRA-93-4.1, 1993. [36] Wilson B.W., Reiter R.J., Pilla A.A. Review of EMF bioeffects Literature Relative to Maglew-generated magnetic Fields. In Electricity and Magnetism in Biology and Medicine. Ed. M. Blank. Press Inc. San Francisco, CA.1993, 251-254. [37] Dietrich F.M., Steiner G.A., Robertson D.C., Feero W.E. Magnetic field testing of TR07 Maglev system. In Electricity and Magnetism in Biology and Medicine. Ed. M. Blank. Press Inc. San Francisco, CA.1993, 267-270. [38] Kopytenko Y.A., Kopytenko E.A., Amosov L.G., Zaitsev D.B., Voronov P.M., Timoshenkov Y.P. 1994. Magnetovariation complex MVC-2. Proc VI Workshop on Geomagnetic Observatory Instruments, Data Acquisition and Processin. Dourbes, Belgium, 24-31. [39] Hilemann B. 1993. Health effects of electromagnetic fields remain unresolved. Chem Eng News. 71, 15-29. [40] Neurath P.W. Simple theoretical models for magnetic intraction waith biological units. In Biological effects of magnetic fields. Ed M.F. Barnothy. Plenum Press, NY, 1, 1964, 25-32. [41] Frankel R.B. Biological effects of static magnetic fields. In CRC Handbook of biological effects of electromagnetaic fields. Ed. Polk C., Postow E. Press. Inc. Boca Raton, FL. 1986, 180-181. [42] Kirshvink J. L., Kobayashi-Kirshvink A., Woodford B. 1992. Magnetite biomineralization in the human brain. Proc Nat Acad Sci USA. 89, 7683-7698.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Railway-Generated Magnetic Field: Environmental Aspects

139

[43] Kirschvink J. L. Rock magnetism linked to human brain magnetite. 1994. EOS Trans. Amer. Geophys Union. 75(15), 178-182. [44] Beason R. C., Nichols J. E. 1984. Magnetic orientation and magnetically sensitive material in a transequatorial migratory bird. Nature 309 151-153. [45] Towne W. F., Gould J. L. Magnetic field sensitivity in honebees. In Magnetic Biomineralization and Magnetoreception in Organisms. Eds J. L. Kirschvink, D. S. Jones, B. J. MacFadden. Plenum Press. NY, 1985, 385-406. [46] Korchevsky E M., Marochnik L. S. 1965. Magnetohydrodynamic version of movement of blood. Biofizika 10 371-374 (Engl. transl: Biophysics 10, 411-414). [47] Morgan M.G., Nair I. 1992. Alternative functional relationships between ELF field exposure and possible health effects: report on an expert workshop. Bioelectromagnetics 13, 335-350. [48] Cleary S.F.A. 1993. Review of in Vitro Studies: Low-frequency Electromagnetic fields. Am Ind Hyg Assoc J. 54, 178-185. [49] Leal J., Shamsaifar K., Trillo M.A., Ubeda A. 1989. Embryonic development and weak changes of the geomagnetic field. Boielectricity. 7, 141-152. [50] Kavaliers M., Eckel L.A., Ossenkopp K.P. 1993. Brief exposure to 60 Hz magnetic fields improves sexually dimorphic spatial learning in the meadow vole, Microtuspennsylvanicus. J Comp Physiol A. 173, 241-248. [51] Lyskov E.B., Juutilainen J., Jousmaki V., Partanen J., Medvedev S., Hanninen O. 1993. Effects of EMF on the human brain activity. Bioelectromagnetics. 14, 87-95. [52] Kavet R.I. A brief perspective on biological effects from intermittent EMF exposure. In Future Epidemiologic Studies of Health Effects of Electric and Magnetic Fields. Electric Power Research Institute EPRI report, TR-101175. Palo Alto, CA, 1992, A47A53. [53] Sait M.L., Wood A.W., Sadafi H.A. 1999. A study of heart rate and heart rate variability in human subjects exposed to occupational levels of 50 Hz circularly polarized magnetic fields. Med Eng Phys. 21 (5), 361-369. [54] SastreA., Cook M.R., Graham C., 1998. Nocturnal exposure to intermittent 60 Hz magnetic fields alters human cardiac rhythm. Bioelectromagnetics. 19 (2), 98-106. [55] Gavalas-Medici R., Day-Magdaleno S. 1976. ELF weak electric field effects scheduled-controlled behavior of monkeys. Nature. 261, 256-259. [56] Makeev V.B., Temuryantz N.A. 1982. Research of frequency dependence on biological activity of magnetic field in geomagnetic pulsations range 0.01-100 Hz In Problemi kosmicheskoy biologii [Problems of Cosmobiology] Nauka, Moscow, Russia, 43, 116129. [57] Maresh C.M., Cook M.R., Cohen H.D., Graham C. 1988. Exercise in the evaluation of human responses to powerline frequency fields. Aviat Space Envirot Med. 59, 11391145. [58] Temuriantz N.A., Martinyuk V.S., Ptitsyna N.G., Villoresi G., Iucci N., Tyasto M.I., Dorman L.I. 2007. Complex-spectrum magnetic environment enhances and/or modifies bioeffects of hypokinetic stress condition. An animal study. Adv Space Res. 40(11), 1758-1763.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

140

N. G. Ptitsyna, G. Villoresi and Y. A. Kopytenko

Reviewed by:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Dr. Kassinsky V.V., Institute of Railway Transportation Engineering, Irkutsk, Russia.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

In: Railway Transportation Editor: Nicholas P. Scott

ISBN: 978-1-60692-863-9 c 2009 Nova Science Publishers, Inc.

Chapter 4

M ARGINAL C OST P RICING OF N OISE IN R AILWAY I NFRASTRUCTURE∗ Henrik Andersson Department of Transport Economics, Swedish National Road & Transport Research Institute (VTI), Sweden ¨ Mikael Ogren Department of Environment and Traffic Analysis, VTI, Sweden

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Abstract In order to mitigate negative effects from traffic it has been decided that infrastructure charges in the European Union (EU) should be based on short run marginal costs. The Swedish Parliament has legislated that operators in the Swedish railway infrastructure must pay charges based on short run marginal social costs in order to mitigate externalities in railway infrastructure. Internalization of the social cost of noise is of particular interest, since it is the only environmental problem perceived as more troublesome today than in the early 1990s. Inclusion of a noise component in rail infrastructure charges raises two issues: (i) the monetary evaluation of noise abatement, since noise is a non-market good, and (ii) the estimation of the effect on the noise level that one extra train will create. Regarding the latter, we are interested in the marginal noise, since infrastructure charges based on the short-run marginal cost principle should be based on the effect from the marginal train, not the noise level itself. Using already existing knowledge, this study shows that it is possible to implement a noise component in the rail infrastructure charges. The values that are used today to estimate the social cost of noise exposure in cost benefit analysis can also be used to calculate the marginal cost. We recommend, however, that further research be carried out in order to get more robust estimates based on railway traffic. We also show that the existing noise estimation models can easily be modified to estimate the marginal noise. Noise infrastructure charges give the operators incentives to reduce their noise ∗ The authors would like to thank Mats Andersson, Karin Blidberg, Gunnar Lindberg, Rikard Nilsson, Ulf Sandberg, and Kjell Str¨ommer for valuable comments on earlier drafts of this chapter, and express their gratitude to Rune Karlsson for the constrained least squares polynomial fit of the marginal cost function. Financial support from Banverket is gratefully acknowledged. The authors are solely responsible for the results presented and views expressed in this chapter.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

¨ Henrik Andersson and Mikael Ogren

142

emissions. We believe that this kind of charge can be used to reduce overall emission levels to an optimal social level, but that it is important for the charge to be based on monetary estimates for rail-traffic and not road-traffic.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

1.

Introduction

Increased environmental consciousness among the general public has led to many of society’s environmental problems becoming stabilized or reduced through political influence and consumer power. Noise is one of the exceptions, being the only environmental effect ¨ over which public dissatisfaction has increased since 1992 ( Ohrstr¨ om et al., 2005). The transport sector is a major contributor to society’s increasing noise problem, due to increased traffic volumes and urbanization exposing more people to noise (Boverket, 2003; Nijland et al., 2003; Bluhm and Nordling, 2005; Kihlman, 2005a). Road traffic is admittedly the largest individual noise source in the transport sector, but other transport modes such as aircraft and railways are also responsible for considerable noise emissions (Kalivoda et al., 2003; Lundstr¨om et al., 2003). This chapter focuses on railway noise, a noise that largely differs from road traffic noise in that an essential part of the disturbance effect comes from individual noise peaks. Noise does not cause any direct environmental damage but entails costs for society in the form of disturbances for the individual (sleep, conversation, recreation, etc.), worsened health and lost productivity. A weakness, from an efficiency perspective, in the transport sector is that the whole noise cost is not borne by those who cause the noise, a so-called external effect. According to the Swedish Parliament’s transport policy decision of 1998 (Prop. 1997/98:56, 1998), operators using the state railways must pay a charge equivalent to the short run marginal social costs (SRMC) in order to mitigate the problem of externalities. Infrastructure charges based on SRMC internalize the external effects on the rest of the society and within the transport sector (e.g. congestion). Charges based on SRMC, thus, give operators incentives to contribute to a more efficient allocation of resources, and have the potential to result in an optimal traffic volume and use of technology. Based on the Swedish legislation, Banverket (Swedish Rail Administration) and SIKA (Swedish Institute for Transport and Communications Analysis) were given a directive to propose a system of charges that would (SIKA, 2002a): • better reflect the railway traffic’s short-run marginal costs on a practical level, at the same time as the operators are given plenty of signals about the desired adaptation of the activity and • comply with EU legislation. 1 The marginal costs for the transport sector on both national and EU levels are the short-run costs; that is, those that directly concern additional traffic. 2 These are divided into five cost groups in the ”White Paper” (Rothengatter, 2003, p. 123); (i) operating costs, (ii) costs 1 EU legislation requires pricing to be based on marginal cost principles, but permits “mark ups” as a means of financing infrastructure projects (European Commission, 1998). 2 For the short-run marginal costs the infrastructure is assumed to be constant. A calculation of the long-run marginal costs also includes the infrastructure investments required as a result of an increase in traffic.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Marginal Cost Pricing of Noise in Railway Infrastructure

143

for the use and wear and tear of the infrastructure, (iii) crowding and scarcity costs, (iv) environmental costs and (v) accident costs. When reporting to the government, SIKA and Banverket came to somewhat different conclusions on noise and marginal costs (SIKA, 2002a). While SIKA was in favor of including marginal costs for noise in the system of railway infrastructure charges, Banverket’s contention was that noise costs should not be included. The latter was based on two standpoints. The first was that there was no basis for including noise costs since marginal cost calculations were lacking and that the average noise costs were not good indicators of marginal costs. The other reason was that charges would not be particularly effective means of influencing the noise from railway traffic. Banverket has given VTI the task of investigating the prerequisites for pricing, with railway noise as one of the components in the charge to be paid by the railway operators. Thus, this chapter comprises three part-aims; (i) ascertain whether marginal noise for the railways exists and how it may be priced, (ii) analyse and make recommendations for how future noise evaluation studies should be formulated and what should be considered in them and (iii) make a sketch of how pricing, based on the estimation of marginal noise and noise value, may be designed. This article is structured as follows. The next section is an outline of marginal cost pricing and externalities. Section 3. on noise first briefly discusses railway noise and thereafter how the marginal noise of trains may be estimated. Section 4. discusses different evaluation methods of noise exposure, followed by section 5. which goes through estimations of the pricing of railway noise. Finally the last two sections (6. and 7.) proposes a structure for how a noise charge for railways may be constructed and contains the conclusions.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

2.

Marginal Cost Pricing

This section provides an outline of the economic argument for why railway infrastructure charges should be based on marginal cost pricing. We will not discuss the conflict between the marginal social cost principle, and long-run incremental costs and full cost recovery. Instead we refer to other literature for discussions on charging principles or financing problems, e.g. Nash (2005), Rothengatter (2003), or Sansom et al. (2001).

2.1.

Externalities and Economic Efficiency

An allocation is socio-economically optimal when the value of the last produced unit is equal to its marginal cost. If the value is higher than the marginal cost, more resources should be allocated to the production of the good at the cost of alternative production. If the value is lower than the marginal cost, the resources should be reallocated to other production alternatives. This relationship is illustrated in Figure 1. The marginal cost function is represented by MCp and the demand function by D. The optimal consumption level is at the point of intersection of MCp and D, that is, at quantity Q p . The optimal quantity Q p is the result of the direct marginal cost. However, consumption often gives rise to external effects both positive and negative. Noise is an example of an externality normally considered negative. This negative externality may be termed an indirect cost and thus society’s social marginal cost consists of the sum of the direct and

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

¨ Henrik Andersson and Mikael Ogren

144

indirect marginal costs. The social marginal cost is given in Figure 1 by MCe and optimal consumption then is Q∗e . The welfare loss of consuming the quantity Q p instead of Q∗e is shown by b, since the values of the last units produced are lower than their marginal costs. P 6

τ

! MCe !! PP ! MCp ! PP PP !! !! ! ! PP! a! b !!! !PP P ! ! c !! c !!! PPP ! PP P !! !! ! D ! ! ! ! -

Qτ

Q∗e

Qp

Q

Figure 1. Marginal cost pricing and economic efficiency.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Assume now that price and quantity are not determined by the similarity between D and MCp or MCe, but that there is a fixed charge equal to τ . Since τ is higher than the price that would otherwise prevail, a lower quantity will be demanded, Qτ . The welfare loss of this too high charge, compared to the case where equilibrium is given by MCp = D, is denoted by the areas a and c in Figure 1. The welfare loss is not as large when equilibrium is determined by the social marginal cost MCe , in which case the welfare loss is denoted by a.3

3.

Railway Noise and Marginal Noise

The noise levels that people are subjected to normally originate from more than one source. ¨ Ohrstr¨ om et al. (2005) investigated the level of disturbance when the noise came from both road and railway sources and found that people were more disturbed by a combined sound than two separate ones, and that their respondents were more disturbed by railway noise than road noise, which runs counter to what other have found. 4 Bluhm and Nordling (2005) found a clear connection between exposure and the level of disturbance. They also found that sleep disturbance, both the falling asleep problem and disturbed night sleep, increased with exposure. Another interesting result of Bluhm and Nordling (2005) was that the disturbance level of railway noise was more correlated with the equivalent level than with the maximum level. The maximum level is the most commonly used indicator in regulations against sleep disturbance. 3 See e.g. S¨ ¨ oderstr¨om (2002) (in Swedish) for a discussion on the Oresund Bridge and the difference between the current bridge toll and the marginal cost of a crossing. 4 The authors suspected that the respondents’ answers to questions on railway noise may have been influ¨ enced by the plans for two more lines in their residential areas ( Ohrstr¨ om et al., 2005).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Marginal Cost Pricing of Noise in Railway Infrastructure

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

3.1.

145

Noise from Railway Traffic

The main source of railway noise is what is known as the rolling noise. Unevenness of the wheels and rails creates vibrations that are then emitted as sound. For low frequencies the rails are the principal noise source, while high frequencies are dominated by wheel sounds. Other sources that contribute are aerodynamic sound at high speeds, noise from fans in the engine and carriages, curve squeal and sounds from braking (see Hartung, 2000). A special problem is the unevenness caused by the brake pads of mainly freight trains, which leads to higher noise levels even when the brakes are not being used (Petersson, 1999). UIC (International Union of Railways) has taken an initiative in this area and shown that an improvement of about 8-10 dBA is attainable if the brake-pad system is replaced using state of the art technology (UIC, 2005). There have been joint Nordic noise estimation methods for railway traffic dating back to 1984. The current method, called NMT96 (Naturv˚ardsverket, 1996), has been in use since 1996. The basis of the method is source parameters, combined with the speed, that describe the noise effect per meter for a given train type. The total train length that passes a given point gives the equivalent level, and the maximum level is determined by the maximum train length for the noisiest train type. Moreover, the method takes into account distance from the track, whether the ground is soft or hard, whether there is screening and several other parameters. A prerequisite for the usability of the method is that new source data is published when new train types are introduced. New and improved methods have been developed since 1996. The so-called Nord 2000 (Plovsing and Kragh, 2000a,b) method has been developed in the Nordic countries and work on new joint methods has been ongoing in the EU with the projects HARMONOISE and IMAGINE (de Vos et al., 2005; Imagine, 2005) for all types of community noise. The largest changes compared to the previous generation of methods concern the accuracy for long distances, effect of the weather and screening. When these methods become operative, they will constitute a solid foundation for all types of community noise calculations in the EU.

3.2.

Two Examples of Marginal Noise Calculation

Marginal noise is the increase of the sound level caused by an extra train set along a given stretch. This change can then be used to calculate the marginal cost of this train type if the relation between the sound level and the corresponding cost is known. Two Swedish examples are given here to illustrate how marginal noise can be calculated, one with a high traffic load (Floda, western mainline) and one with a low load (Hind˚as, Bor˚as line). In a 24-hour period 190 and 39 trains pass through Floda and Hind˚as, respectively. Traffic data is summarized in Table 1. Three examples are used to estimate the increase in noise level for these two places, a commuter train (Swedish notation X14), a high speed train (X2) and a freight train, see Table 2. In both cases the traffic is somewhat simplified and the land around the track at the time of estimation is assumed to be flat and acoustically hard. Hence, the sound levels are overestimated somewhat compared to places with soft ground. It is also important to include the effects of noise screens wherever they exist. The total level is affected by both screens and soft ground, while noise changes are affected to a much lower extent. This

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

¨ Henrik Andersson and Mikael Ogren

146

Table 1. Rail traffic at example areas. Site Traffica Commuterb High speed b Freight Others High traffic 190 76 28 41 45 load (Floda) Low traffic 39 24 0 7 8 load (Hind˚as) a: Number of trains per 24 h. b: The Swedish notations for the commuter and high speed trains are X14 and X2.

means, in principle, that there is no obstacle to including both the screening and ground effects in the estimations. Indeed, both are included in NMT96 and the future models HARMONOISE/IMAGINE. Table 2. Train sets used for estimating the marginal cost.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Train Commuter (X14) High speed (X2) Freight

Length [m] 50 200 650

Speed [km/h] Floda Hind˚as 120 100 135 100 100 100

Figure 2 shows the estimated increase of the equivalent A-weighted sound level for a 24-hour period (LA,Eq,24h ).5 The largest increase (0.9 dBA) is for freight trains and the increases are larger at the low-traffic site since fewer trains contribute to the total level. The lowest effect is for passenger trains in the high-traffic site, 0.003 dBA. It is relatively easy to estimate the change in equivalent level for a train set, but estimating the effect on the maximum level is a more difficult matter. According to the methods available, the noisiest among all the train sets in traffic on the investigated line determines the maximum level. Hence, only one train set affects the maximum level, often the longest freight train in traffic, while the rest do not contribute to the maximum level at all. In other words, the maximum level is independent of the traffic volume in the estimation model, and the change in the maximum level is zero for all trains except the noisiest. In reality the probability that an extra noisy train will pass if the traffic is high is increased. It may also be that the maximum level is determined by different train types on different days depending on random factors such as wheel wear, number of freight wagons and so on. The maximum level is used mainly as an indicator of sleep disturbances. Another indicator that is becoming increasingly more important in the EU is Lden , “level day, evening night”. This is an equivalent level (that is, all trains contribute) but traffic in the evening and at night is punished with 5 and 10 dB penalty according to the formula (European 5 A-weighting means

that the sound pressure at different frequencies is transformed into a value that likens

the sensitivity of the ear. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Marginal Cost Pricing of Noise in Railway Infrastructure

147

Commission, 2000a): Lden = 10 log



 12 0.1 Ld 4 0.1 (Le+5) 8 0.1 (Ln+10) 10 + 10 + 10 , 24 24 24

(1)

where Ld , Le and Ln denote the equivalent noise levels for the day, evening and night period, respectively. The intention is to create a unit value for noise disturbance that will represent both general disturbance, which is described as the equivalent level, and sleep disturbance, which is described as the maximum level at night. Lden is the noise indicator that will apply in the EU in future. In the example above goods traffic will dominate even more if we calculate Lden instead of LA,Eq,24h . Since Lden is only a recalculation of the equivalent levels for three periods, it is easy to adapt NMT96 to it if the traffic distribution over a 24-hour period is known. 1 0.9

0.7

0.5 0.4 0.3 0.2

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Low traffic load (Hindås)

0.6 High traffic load (Floda)

Change in sound level [dB]

0.8

0.1 0 X14

X2

Freight

Figure 2. Change in A-weighted equivalent sound pressure level due to one extra train passage per 24 h, empty bar is for high traffic load (Floda) and filled bar for low traffic load (Hind˚as).

4. 4.1.

Preference Elicitation Evaluation of Noise

As noise itself is a non-market good, there are no direct observable market prices for the value of reducing noise levels. For so-called non-market goods, apart from noise, e.g., safety and clean air, methods have therefore been developed to estimate individuals’ preferences. These methods are usually divided into two main groups depending on what information is used. Preference estimations based on market data are called “revealed preferences”

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

148

¨ Henrik Andersson and Mikael Ogren

(RP) because they are based on observed behavior. The alternative is to estimate so-called “stated preferences” (SP) in which it is clear that individuals have given their preferences in hypothetical market situations, e.g. answers to questionnaires and interview questions. An overwhelming majority of noise evaluation studies use the RP approach and the method employed is the hedonic regression technique (Navrud, 2004). The hedonic method was formalized by Sherwin Rosen (1974) and has its point of departure in the fact that the prices of goods depend on their utility-bearing attributes. By studying how house prices are affected by exposure to noise, at the same time as effects on the price of other attributes are controlled for, it is possible to estimate the marginal willingness to pay (WTP) to reduce the noise level. The “noise sensitivity depreciation index” (NSDI), which is a measure of the percentage change in price resulting from a unit change in noise level, is used to compare the results of different hedonic price studies of noise evaluation. 6 Bateman et al. (2001) reviewed the estimated NSDI values of various studies. The values varied from 0.08 to 2.22 with a mean of 0.55. The mean value implies that a dBA increase would reduce property values by just over half a percent. One of the strengths of hedonic price studies is that they are based on individuals’ actual behavior and thereby their actual WTP. 7 A drawback, though, is that it can be difficult if not impossible to estimate the values that are of interest. SP methods, since they enable the analyst to tailor the survey/experiment, make it possible to elicit individuals’ preferences for specific noise circumstances. For instance, using a sample of residents near Bromma airport outside Stockholm, Carlsson et al. (2004) carried out an SP study in the form of “choice experiments” with the aim of estimating individuals’ preferences for aircraft noise at different times of the day and on weekdays and holidays. The results indicate a higher WTP to reduce noise level on weekday evenings and weekends, holidays and working days alike. The SP method that is used most extensively for evaluating noise is the “contingent valuation method” (CVM) (Navrud, 2004). There are several different ways of designing CVM studies, but common for all of them is that the respondents directly state their WTP for the good (Bateman et al., 2002), here a reduction of the noise level. The advantage of the CVM and other SP methods is, as mentioned above, that the person conducting the study is allowed to construct it and can thus ask the questions he/she wants answers to and control for how different factors, e.g., study design, may have affected the results. The drawback is that the answers are to hypothetical questions and that there is no guarantee that the respondents will actually act as they have answered. This is a generally known phenomenon and several factors may explain why people answer like they do. One that is often mentioned is “hypothetical bias”, which refers to an overestimation of the WTP since the respondent is aware that he/she does not really have to pay the amount in question. Other problems frequently linked to CVM studies are insensitivity to the quantity of the 6 Let P(A) and dBA denote the house price as a function of the attribute vector A and the noise level (which is included in A); then NSDI is given by (Bateman et al., 2001): % change in P(A) change in P(A) due to noise exposure 100 × = NSDI = P(A) change in dBA . change in dBA 7 For example, the price of a house is affected not only by exposure to noise but also by the number of rooms,

view and so on. The hedonic regression method takes this into consideration in estimating the price function. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Marginal Cost Pricing of Noise in Railway Infrastructure

149

good, strategic answers, and poorly constructed studies (see e.g. Kahneman and Knetsch, 1992; Carson et al., 2001; Bateman et al., 2002). 8 The official Swedish monetary values that are currently in use as a basis for the estimation of noise costs for both roads and railways (SIKA, 2002b) rest heavily on the results from a hedonic pricing study which investigated how road noise affected house prices in ¨ Angby outside Stockholm (Wilhelmsson, 1997). The recommended values were estimated by means of linear regression and an interaction variable for visual exposure to road and noise levels was used to estimate a progressive connection between the marginal WTP for a noise reduction and noise level. The marginal WTP consisted of two linear segments with a break at 68 dBA. Wilhelmsson’s study resulted in NSDI values that varied between 0.5 and 5.0.9

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

4.2.

Benefit Transfer

Benefit transfer (BT) refers to the technique of using estimated values of one population for other populations. One example is to use the values estimated for one country to calculate the utility of a measure in another country. The weakness of this kind of transfer is that the estimated values are dependent on how they are extracted, and on the consumption alternatives of the population. Estimated values are thus dependent on the context in which they are estimated. For example, a Swedish hedonic price study of car safety showed substantial differences compared to corresponding American studies (Atkinson and Halvorsen, 1990; Dreyfus and Viscusi, 1995; Andersson, 2005). The difficulties in using noise values from hedonic price studies of BT are that the values show the marginal WTP for just one market (Day, 2001) and that the estimates can be sensitive to model assumptions such as functional form and the included variables (Andersson, 2005). Noise evaluation indeed show a variation in estimates of NSDI between studies and countries (Bateman et al., 2001). CVM and other SP methods used in BT have the advantage that the values are estimated for hypothetical situations, which suggests that it is possible to establish how the study and selection have affected the results. However, a problem with BT is that the estimations in an SP study are dependent on just that specific study’s design (Carson et al., 2001). This, in combination with the difficulty of estimating individual preferences in SP studies, means that we cannot recommend the one over the other for use in BT.

4.3.

Future Evaluation

In a recent report from the Swedish National Road Administration (Wall, 2005) the recommendation is that noise values should be estimated with SP methods. Indeed, the use of 8 Many

of the environmental and health goods that are often of interest in SP studies are abstract for the general public because people are not accustomed to “dealing” in them. These goods are not easy to make tangible, and an important aspect of research in environmental and health evaluation deals with just the question of how these goods should be presented in evaluation studies. 9 A result worth noting in Wilhelmsson’s study is that he found that the marginal willingness to pay varied over time. Cross-section data from a short time period therefore runs the risk of over or underestimating the actual willingness to pay. Wilhelmsson chose to use data from a five-year period (1990-95) in his estimations.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

150

¨ Henrik Andersson and Mikael Ogren

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

research funds for an hedonic noise evaluation is frowned on. St˚ale Navrud strongly advocates that values estimated for disturbance, and not dBA should be used in BT and that they should be estimated with SP methods (Navrud, 2004, p. 30). An argument Navrud uses is that the value of reducing the level of disturbance can be estimated in SP studies and that this estimate is more suited to BT than the estimated value per dBA. The EU-project HEATCO recently conducted CVM studies in several European countries, including Sweden, with the aim of estimating the WTP for reducing the disturbance of road and railway noise and to estimate the value of time HEATCO. However, the Swedish questionnaire asked respondents about their WTP to reduce the noise level from road traffic alone. The results indicate a relatively large difference in WTP between countries. For Sweden the finding is that WTP increases with the level of disturbance. There were indications of method problems, for example the proportion that agreed to pay a certain amount of money did not diminish monotonically with the offer and a large proportion (45%) were not willing to pay anything at all. We are not as convinced that a certain method or a certain WTP measure (per disturbance level or dBA) is superior. As mentioned above, both RP and SP methods have their strengths and weaknesses. A well constructed study controls for all the accessible attributes that can be assumed to affect house prices, and thereby sorts out the WTP that is of interest. The most serious weakness of the HP method, as we see it, is the problem of BT, which does not depend on the study’s quality. A well conducted CVM study can solve many of the difficulties mentioned above. A problem in estimating WTP for reducing the level of disturbance, apart from the usual difficulties in hypothetical studies, is that the estimates build on the assumptions of those carrying out the studies on how reduction of the level of disturbance can be transformed into actual noise level changes. We are therefore not convinced that the level of disturbance is to be preferred to dBA.

5.

Noise and Marginal Costs

Disturbance as a result of noise is not the only effect of exposure to railway noise (and other sources of noise), exposure may also lead to reduced health status and loss of production. The latter may be due to worsened health resulting in absence from work or reduced working capacity, or be caused by e.g., disturbed night sleep leading to the worker in question becoming less productive. The social costs of the results of noise exposure may be divided into three groups (Metroeconomica, 2001): 1. Resource costs in the form of medical care and attention. Includes both costs financed by taxes and direct payments by the individual. 2. Opportunity costs in the form of loss of production. Also covers “non-market services” performed in the home and lost leisure time. 3. Welfare losses in the form of other negative effects resulting from exposure to noise. Disturbances of various forms and increased concern for the consequences of exposure are two examples of possible welfare losses. Resource costs and Opportunity costs can be estimated with existing market prices and the sum of both is usually called “cost of illness” (COI) in health economics literature. For the

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Marginal Cost Pricing of Noise in Railway Infrastructure

151

last part-cost, Welfare losses, for which there are no market prices, WTP has to be used to estimate prices, that is, with RP or SP methods. Official values recommended by ASEK (The Committee on Cost-Benefit Analysis) of the cost caused by a certain noise level (equivalent level) are available for both road and rail traffic (SIKA, 2008). Noise costs (BV ) for railways are recommended to be calculated with the following formula,   0.88 BV = 6.9(70 + t)1.1 e0.18(N−45) − 1 , (2) which gives a value per person and year and where t is the number of trains per 24-hour period and N the maximum level indoors in dBA (SIKA, 2008). 10 Equation (2) is based on estimated WTP for road traffic (Wilhelmsson, 1997), but is revised to reflect differences in noise profile between road and rail traffic. For N > 45 it can be shown that BV is progressively increasing in t and N and increasing more in t for high N and vice versa.11 The official noise cost model in use today shows that there is a positive connection between the number of trains and maximum levels and thus that a marginal cost is assumed to exist (cf. the discussion in section 1.).

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

5.1.

Marginal Noise Cost

A calculation of marginal costs requires knowledge of the connection between the actual marginal train’s noise and the costs caused by noise. Section 3.2. describes how the marginal noise from one more train can be calculated. Society’s marginal evaluation of an increased noise level is estimated aptly with WTP, as shown in the section 4.. Not having access to direct estimates of individuals’ marginal WTP for railway noise, we use the values recommended by ASEK for road noise (SIKA, 2008). An alternative would be to use Eq. (2) to calculate an indirect value. However, we have chosen to use the monetary values for road noise (SIKA, 2008) due to our lack of success in tracing the origin of Eq. (2) and the problem with the N-variable (see below).12 The values for the monetary social cost of road noise is only available in table format, and by adapting a polynomial to the difference in table values with the help of regression analysis, see Figure 3, we have derived a marginal cost function, f (L) = 0.0617(L − 62)3 + 1.28(L − 62)2 + 7.11(L − 62) + 39.1,

(3)

where L is the equivalent A-weighted sound level and f (L) gives the marginal cost per person and year. Note that this is a very simple function that illustrates the principle and that there are many alternative approaches, e.g., “splines”. Figure 4 shows the estimated yearly marginal cost for one person at a distance of 50 m from the railway line for the two case studies presented in section 3.2.. The figure includes 10 Equation (2) applies to t ∈ [1, 150]. For t > 150 BV is multiplied by the multiplication factor (M), M = 1 + t−150 1050 (Banverket and Naturv˚ardsverket, 2002). 11 ∂ BV > 0, ∂ 2 BV > 0, ∂ BV > 0, ∂ 2 BV > 0 (∀ N > 45.76) and ∂ 2 BV > 0. ∂t ∂ t2 ∂N ∂ N2 ∂ t∂ N 12 If we had access to the cost function that is the basis for the table values that have been used, the derivative of this in respect of the noise level would reflect the marginal evaluation. The marginal cost could then be calculated by multiplying the derivative of the cost function with the change of sound level. Despite making enquiries, we have not managed to find the original function.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

¨ Henrik Andersson and Mikael Ogren

152 500

ASEK m*10^(k*L) Polynomial

Marginal cost [EUR/yr./person]

450 400 350 300 250 200 150 100 50 0 50

55

60 65 Equivalent A−weighted SPL (24h) [dB]

70

75

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 3. Exponential and polynomial regression fits to the marginal cost function as a function of the sound pressure level. The cost function is taken from SIKA (2008).

two approaches with which to derive this cost: (i) directly from the estimation of noise level via Figure 3 and (ii) indirectly via Eq. (2). The parameter N in Eq. (2) is the maximum level, but to be able to make a comparison it is assumed to be the equivalent level ( LA,Eq,24h ) with and without the train at the point of reception. It is more difficult to work with the maximum level since it does not change if the train is not the noisiest along the line and is not affected by the volume of traffic. Since the maximum level is higher than the equivalent level, the marginal cost is underestimated when using this approach. The marginal cost is calculated at a distance of 50 m from the track in a position without significant screening for the three different trains shown in Table 2. The change in A-weighted equivalent noise level is calculated in Figure 2. While this change is not attributable to distance from the track, the absolute level is, which is why we assume a distance of 50 m. Marginal cost calculated with Eq. (2) becomes 50 − 100% higher than with the direct method. Note too that the marginal cost per person and year 50 m from the track becomes the same in size for both high traffic (Floda) and low traffic (Hind˚as). This is due to the fact that a train makes smaller changes to a noise level that is already high because of other traffic, but the smaller change incurs a higher monetary value since the noise level is high. Put another way; since the change in cost increases substantially for higher sound levels (see Figure 3), even a small change in effect on the equivalent noise level in a noisy place will incur a cost.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Marginal Cost Pricing of Noise in Railway Infrastructure

153

45 40

15 10

BV − low traffic load (Hindås)

20

Low traffic load (Hindås)

25

BV − high traffic load (Floda)

30 High traffic load (Floda)

Marginal cost [EUR]

35

5 0 X14

X2

Freight

Figure 4. Calculated marginal cost per year for one inhabitant at a distance of 50 m from the railway.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

5.2.

Marginal Cost and Infrastructure Charge

The social cost of a marginal noise increase can be calculated by combining the estimated marginal cost with the number of exposed. Let g(L) denote the distribution function for the number of exposed at different noise levels. The marginal cost (MC) for noise at a given noise level Li can thereby be written as; MC(Li ) = f (Li )g(Li).

(4)

Society’s total marginal noise cost ( T MC) can then be calculated as T MC =

Z ∞ π

f (L)g(L)dL,

(5)

where π is the lower limit under which noise is not expected to cause annoyance. Table 3 below gives a calculation example of the marginal of driving one extra train of type X14 (commuter train), X2 (high speed) and freight train through the area. The link is entirely fictitious and consists of, apart from departure and arrival points, a small urban area between two sparsely populated areas (denoted rural). The total number of exposed in each noise level category is estimated from calculations using NMT96 and assuming an even population distribution throughout the affected areas. The number of exposed is fewer for the low traffic load example since the total noise source strength is lower causing lower noise level at the same distance from the railway compared to the high traffic load area, which is illustrated in Figure 5. The results of the highly simplified model show, in line with the results in Figure 2, significant differences in marginal noise between the X14, X2 and freight trains and between

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

¨ Henrik Andersson and Mikael Ogren

154

high traffic load

low traffic load

55 dB 60 dB railway line

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5. Sketch of the effect of changing the traffic volume in the example. The noise contours move outward to include more dwellings when the traffic is increased.

low and high traffic loads. The table also shows the effect of noise on the marginal cost calculation. The speed is assumed to be the same for X14 and freight trains, which is not realistic, but since the effect of speed on the estimations is not significant we have, for the sake of simplicity, chosen to let the speed be the same. 13 Table 3 also shows estimations of differentiated noise charges in /km (2006 price level). The large difference between noise charges for commuter and freight trains can be traced to the difference in marginal noise addition. What is also interesting is that the charge through the low traffic load area is higher than through the high traffic load area. The effect of higher marginal noise at a lower noise level is thus stronger than at a higher noise level. It is also worth noting that the charges are in line with the ranking of disturbance level for passenger and freight trains; that is, people feel more disturbed by noise from freight trains. The marginal cost is directly proportional to the number of exposed. Doubling the population in the area will double the marginal cost per km through it (under the assumption that the added population is evenly distributed). Note however that the marginal cost is not proportional to the total traffic. In the example above increasing the traffic from 39 trains per 24h to 190 (almost a 400% increase) lowers the marginal cost by approximately 65%, so the estimated marginal costs are less sensitive to changes in the total traffic volume.

5.3.

Impact Pathway Approach

The Impact Pathway Approach (IPA) is a method that has been suggested for estimating the cost of noise (Metroeconomica, 2001; Navrud, 2004). 14 The method has been developed for energy externalities (ExternE, 2005) and is intuitive and easy to understand. In short, the method follows a “bottom-up approach”; that is, its starting point is the source of emission, and it estimates the propagation and final effects of the emission. These effects are then given monetary values and the social cost can be established. IPA has been used in a several studies to estimate noise costs (Metroeconomica, 2001; Schmid and Friedrich, 2002; Bickel et al., 2002). The method is also advocated by St˚ale Navrud in his “state-of-the-art”-study of the evaluation and cost calculation of noise 13 This is not to imply that we are suggesting that speed should not be considered in future implementations of noise charges. 14 Most often termed “impact pathway approach”, and sometimes “damage function approach”, in the literature. For a run-through of the method see e.g., Friedrich and Bickel (2001).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Marginal Cost Pricing of Noise in Railway Infrastructure

155

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Table 3. Example of noise charges for three train types Rural Urban Rural Total Length [km] 8 2 8 18 High traffic load Number of exposed 50-54 dB 600 1600 600 2800 55-59 dB 250 600 250 1100 60-64 dB 100 250 100 450 65-69 dB 40 80 40 160 70-74 dB 2 4 2 8 Marginal costa Commuter 0.02 0.24 0.02 0.05 High speed 0.09 0.94 0.09 0.19 Freight train 0.74 7.41 0.74 1.48 Low traffic load Number of exposed 50-54 dB 100 240 100 440 55-59 dB 40 100 40 180 60-64 dB 16 32 16 64 65-69 dB 0 0 0 0 70-74 dB 0 0 0 0 Marginal costa Commuter 0.03 0.33 0.03 0.07 High speed 0.11 1.04 0.11 0.21 Freight train 1.49 13.99 1.49 2.88 a: /km, 2006 price level.

(Navrud, 2004). Another reason for using the method is that it cannot be assumed that individuals are aware of all the effects of noise exposure (Navrud, 2004), e.g., increased blood pressure. The monetary value that individuals either disclose through their decisions (RP) or state in questionnaires and interviews (SP) is therefore expected to underestimate the actual social noise cost. Lindberg (2003) compared the Swedish efforts to estimate social noise cost with the results from Schmid and Friedrich (2002) and found only small differences. However, Lindberg was of the opinion that this was probably due to “chance” (Lindberg, 2003). Even if IPA is intuitively appealing, there are several well known problems. The method builds on emission being measured with precision in terms of its spread and eventual effects. Data for the specific effects (e.g., health status) and preference estimations for the goods or populations is also necessary. Since this data is often lacking, the cost calculations will depend on assumptions regarding health data and how well BT functions in the specific case. This, together with the unreliability associated with the calculation of “dose-response”, results in great uncertainty in the estimations (Metroeconomica, 2001). The assumption that individuals are not aware of the various health risks can also be an issue. If there is general

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

¨ Henrik Andersson and Mikael Ogren

156

awareness of the risk of, e.g., increased blood pressure, IPA will mean a double counting of the social utility from reducing the noise level. It is therefore important to establish what individuals take into account when they state their WTP in practical decisions or as answers to hypothetical questions, something that has to be done empirically. A general problem with IPA is the evaluation of the risk of death. The measure “value per statistical life-year” (VSLY), which is based on estimations of “value of a statistical life” (VSL), is used in IPA. 15 When using VSLY it is assumed that VSL diminishes with age (Hammitt, 2000; Alberini and Krupnick, 2002). A document from the European Commission on recommended values for reducing risk states that there is strong theoretical and empirical evidence that VSL diminishes with age (European Commission, 2000b), but Johansson (2002) showed that there is no strong theoretical evidence of VSL diminishing with age. Empirical support for the same also seems to be lacking (Aldy and Viscusi, 2007; Krupnick, 2007), something that users of IPA are often aware of (Bickel et al., 2002, p. 16).

6.

Suggestions for a Charging Model

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

A noise charge for railway operators should be differentiated and reflect the actual differences in marginal costs for different trains, how the trains are driven, at what time of the day and on what stretches and so on. The desired differentiation is not always possible when implementing the charges according to the marginal principle, since it would be prohibitively costly to obtain the correct bases for exact differentiation. Therefore, there is a need to introduce a system that is implementable, given the existing knowledge level of the various marginal costs and taking into account the balance between the benefit of including the externality and the cost of calculating it. In order to establish the size of the marginal cost of the noise to be included in the railway infrastructure charge, the following bits of information are needed; 1. the noise situation before the change, 2. how the noise situation is changing and 3. connection between the noise level and cost. It is natural to use the available noise calculation methods for noise from railway traffic (Harmonoise, NMT96) for the first two points. At best there is information on the noise situation before the change which will already be accessible in the form of noise maps where the levels are given as contours or colored areas. The change in the new trains can probably be calculated with good precision once and then be valid for the whole area. If, for example, the equivalent level increases by 0.1 dBA, then it will be applicable both near and far from the track and in screened areas as well. It would also be desirable to be able to distinguish between these in terms of their actual noise emission, e.g., in the form of a certification system. Operators who have invested in noise reducing technology should be charged less in such a system compared to those who have not invested in similar technology. δ denote theR discount rate of interest; then the following relation applies (Alberini and Krupnick, T (1 + δ )−t . 2002); V SL = V SLY t=1 15 Let

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Marginal Cost Pricing of Noise in Railway Infrastructure

157

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The monetary evaluation of changed noise levels is also needed to estimate the noise component of the railway infrastructure charge. As seen in the calculation example in the Table 3, estimates for road traffic are used. Research shows that people are disturbed to varying degrees by road and rail noise and it can therefore be desirable to evaluate railway noise. Given the values that reflect the social marginal cost for railway noise, information on the noise levels at the exposed individuals’ houses is needed for calculation of the actual marginal cost. In principle, the sound level depends on the distance from the track plus any screening and land effects. In other words, points 1 and 3 above require geographic information (where houses are located, distance, possible noise screens, land quality and so on) and noise evaluation. For point 2 details are required for the train for which the railway infrastructure charge is to be calculated. If it is a known train type then a measuring method is already prescribed for example in de Vos et al. (2005). How the necessary input is connected is illustrated in the Figure 6.

Figure 6. Overview of information needed to estimate the noise component of the rail access charge (the marginal cost of noise).

An alternative that will reduce the amount of input data required for the calculation is to use the rule of thumb for how the population is distributed at various distances from the track. Noise calculation can then be made for a number of scenarios, e.g., built-up area with noise screens, countryside with soft arable land and the like. A suitable basis for the development of such rules of thumb would be to make accurate case studies with noise maps and investigation of housing locations in several places. With these case studies as a starting point, one can compare the effects of different assumptions and rules of thumb and thus acquire an understanding of the loss of precision in the noise charges from these assumptions and rules of thumb.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

¨ Henrik Andersson and Mikael Ogren

158

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

7.

Conclusion

This chapter makes suggestions for how marginal noise from trains can be calculated, discusses issues associated with the evaluation of noise, proposes a sketch of what is needed to include a noise component in the railway infrastructure charge and provides a calculation example with a noise charge. The estimated values in this chapter are only numerical examples and we would advise against their use in noise charges. 16 Considering the sketch for the railway infrastructure charge in this chapter, we do not see the declaration procedure as a basis for collection of the railway infrastructure charges, as a problem. The information that rail authorities need from the operator is the stretch, train type, number of carriages and speed. The noise maps that are presently being produced as required by the Environmental Noise Directive of the European Commission (European Commission, 2002) will give the authorities the necessary information on noise level and the number of exposed. We consider that the greatest obstacle to an implementation of a noise charge in the near future is the lack of an estimated monetary value for railway noise. This value can be estimated, however, and is not an acceptable reason for rejecting the marginal cost principle. Criticism has been aimed at the slowness of the work on reducing noise levels, especially on the emission side (see e.g., Kihlman, 2005b). Noise charges based on the marginal cost principle are a possible way of providing an incentive for those who cause the noise to reduce the “emission”. Even if a socially optimal noise level, and not a reduced noise level, is the primary objective of marginal cost pricing, today’s “zero price” of noise means that the marginal cost principle will make it costly to spread the noise and thereby provide an incentive to reduce the noise level. We, therefore, believe that a charge that reflects noise costs would be an effective means of reducing noise levels. Bluhm and Nordling (2005) found no connection between railway noise and increased blood pressure, but several studies show a connection between noise disturbance and health levels (see Babisch et al., 2005). Therefore, there is support for the fact that a marginal noise increase leads to worsened health and a resulting social cost. Whether the individual takes this into consideration when evaluating noise (actual or hypothetical) is an issue of an empirical nature that must be investigated. As we do not believe that it is possible, in the short run, to estimate differentiated values for housing, workplace, school, recreation and so on, estimations should be based on housing in the future as well. The same applies to the possible differentiation between passenger and freight trains. Differentiated estimations would probably be far too expensive today. If rail authorities unilaterally internalized the railways’ externalities and based the railway infrastructure charge on the marginal cost principle, the resulting problem would be a possible distortion in the competitive climate. Notwithstanding, this is not a sustainable argument for excluding noise from the charge while e.g., wear and tear are included. We consider it better for all the externalities to be included and for the problem of the distortion of competition to be dealt with in another way. If the railways could solve the implementation issue and introduce a system based on the marginal cost principle, we believe it would probably result in political pressure for other transport modes to be embraced by the principle as well. 16 For

¨ estimates of the SRMC based on a Swedish case study, see Andersson and Ogren (2007).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Marginal Cost Pricing of Noise in Railway Infrastructure

159

References Alberini, A. and A. Krupnick: 2002, The International Yearbook of Environmental and Resource Economics 2002/2003 , Chapt. Valuing the Health Effects of Pollution, pp. 233– 277. Cheltenham, UK: Edward Elgar. Aldy, J. E. and W. K. Viscusi: 2007, ‘Age Differences in the Value of Statistical Life: Revealed Preference Evidence’. Review of Environmental Economics and Policy 1(2), 241–260. ¨ Andersson, H. and M. Ogren: 2007, ‘Noise Charges in Rail Infrastructure: A Pricing Schedule Based on the Marginal Cost Principle’. Transport Policy 14(3), 204–213. Andersson, H.: 2005, ‘The Value of Safety as Revealed in the Swedish Car Market: An Application of the Hedonic Pricing Approach’. Journal of Risk and Uncertainty 30(3), 211–239. Atkinson, S. E. and R. Halvorsen: 1990, ‘The Valuation of Risks to Life: Evidence from the Market for Automobiles’. Review of Economics and Statistics 72(1), 133–136. Babisch, W., B. Beule, M. Schust, N. Kersten, and H. Ising: 2005, ‘Traffic Noise and Risk of Myocardial Infarction’. Epidemiology 16(1), 33–40.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Banverket and Naturv˚ardsverket: 2002, ‘Buller och vibrationer fr˚an sp˚arbunden linjetrafik: Riktliner och till¨ampning’. Banverket (National Rail Administration) and Naturv˚ardsverket (The Swedish Environmental Protection Agency). Bateman, I. J., R. T. Carson, B. Day, M. Hanemann, N. Hanley, T. Hett, M. Jones-Lee, G. ¨ Loomes, S. Mourato, Ozdemiro¯ glu, D. W. Pearce, R. Sugden, and J. Swanson: 2002, Economic Valuation with Stated Preference Techniques: A Manual . Cheltenham, UK: Edward Elgar. Bateman, I., B. Day, I. Lake, and A. Lovett: 2001, ‘The Effects of Road Traffic on Residential Property Values: A Literature Review and Hedonic Pricing Study’. Technical report, University of East Anglia, Economic & Social Research Council, and Univeristy College London. Bickel, P., S. Schmid, and R. Friedrich: 2002, ‘Estimation of Environmental Costs of the Traffic Sector in Sweden’. Mimeo, Institut f¨ur Energiewirtschaft und Rationelle Energieanwendung (IER), Universit¨at Stuttgart. Bluhm, G. and E. Nordling: 2005, ‘Health Effects of Noise from Railway Traffic - The HEAT Study’. Inter-Noise. Boverket: 2003, ‘Buller: Delm˚al 3 - Underlagsrapport till f¨ordjupad utv¨ardering av milj¨om˚alsarbetet’. Boverket, Karlskrona. Carlsson, F., E. Lampi, and P. Martinsson: 2004, ‘Measuring Marginal Values of Noise Disturbance from Air Traffic: Does the Time of Day Matter?’. Working Papers in Economcis 125, Dept. of Economics, Gothenburg University, Gothenburg, Sweden. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

160

¨ Henrik Andersson and Mikael Ogren

Carson, R. T., N. E. Flores, and N. F. Meade: 2001, ‘Contingent Valuation: Controversies and Evidence’. Environmental and Resource Economcis 19(2), 173–210. Day, B.: 2001, ‘The Theory of Hedonic Markets: Obtaining Welfare Measures for Changes in Environmental Quality using Hedonic Market Data’. Economics for the Environment Consultancy, Mimeo. de Vos, P., M. Beuving, and E. Verheijen: 2005, ‘Harmonised Accurate and Reliable Methods for the EU Directive on the Assessment and Management of Environmental Noise – Final technical report’. Technical report. http://www.harmonoise.org. Dreyfus, M. K. and W. K. Viscusi: 1995, ‘Rates of Time Preference and Consumer Valuations of Automobile Safety and Fuel Efficiency’. Journal of Law and Economics 38(1), 79–105. European Commission: 1998, ‘White Paper on Fair Pricing for Transport Infrastructure Use’. European Commission: 2000a, ‘Proposal for a Directive of the European Parliament and of the Council relating to the Assessment and Management of Environmental Noise’. Technical Report COM(2000) 468 final, 2000/0194(COD), Commission of the European Communities. European Commission: 2000b, ‘Recommended Interim Values for the Value of Preventing a Fatality in DG Environment Cost Benefit Analysis’. mimeo, 08/06/04 available at: http://europa.eu.int/comm/environment/enveco/others/recommended interim values.pdf, 08/06/04.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

European Commission: 2002, ‘Environmental noise directive 2002/49/EG’. ExternE: 2005, ‘ExternE: Externalities of Energy. A Reserach Project of the European Union’. Internet, www.externe.info, 10/7/05. Friedrich, R. and P. Bickel: 2001, Environmental External Costs of Transport . Heidelberg, Germany: Springer Verlag. Hammitt, J. K.: 2000, ‘Valuing Mortality Risk: Theory and Practice’. Environmental Science & Technology 34(8), 1396–1400. Hartung, C. F.: 2000, ‘Vibrations and external noise from train and track - a literature survey’. Report F227, Chalmers Solid Mechanics. HEATCO: 2005, ‘Developing Harmonised European Approaches for Transport Costing and Project Assessment, Deliverable four - Economic values for key impacts valued in the Stated Preference surveys’. http://heatco.ier.uni-stuttgart.de, 8/29/07. Imagine: 2005. http://www.imagine-project.org/. Johansson, P.-O.: 2002, ‘On the Definition and Age-Dependency of the Value of a Statistical Life’. Journal of Risk and Uncertainty 25(3), 251–263. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Marginal Cost Pricing of Noise in Railway Infrastructure

161

Kahneman, D. and J. L. Knetsch: 1992, ‘Valuing Public goods: The Purchase of Moral Satisfaction’. Journal of Environmental Economics and Management 22(1), 57–70. Kalivoda, M., U. Danneskiold-Samso, F. Kr¨uger, and B. Bariskow: 2003, ‘EURailNoise: a study of European priorities and strategies for railway noise abatement’. Journal of Sound and Vibration 267, 387–396. Kihlman, T.: 2005a, ‘Developments in Environmental Noise Policies’. Budapest, pp. 35– 40, Proceedings of Forum Acusticum. Paper 964-0. Kihlman, T.: 2005b, ‘Developments in Environmental Noise Policies’. ForumAcusticum. Krupnick, A.: 2007, ‘Mortality Risk Valuation and Age: Stated Preference Evidence’. Review of Environmental Economics and Policy 1(2), 261–282. Lindberg, G.: 2003, ‘Marginella bullerkostnader: en genomg˚ang av SIKAs och V¨agverkets arbete samt analyser med effektkedjemetoden’. SIKA Rapport 2003:1, Bilaga 3, Trafikens externa effekter - Uppf¨oljning och utveckling 2002 - Bilagor, Stockholm, Sweden. Lundstr¨om, A., M. J¨ackers-C¨uppers, and P. H¨ubner: 2003, ‘The new policy of the European Comission for the abatement of railway noise’. Journal of Sound and Vibration 267, 397– 405. Metroeconomica: 2001, ‘Monetary Valuation of Noise Effects: Final Draft Report’. Mimeo, May, Preparde for The EC UNITE Project.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Nash, C.: 2005, ‘Rail Infrastructure Charges in Europe’. Journal of Transport Economics and Policy 39(3), 259–278. Naturv˚ardsverket: 1996, ‘Buller fr˚an sp˚arburen trafik – Nordisk Ber¨akningsmodell’. Rapport 4935, Naturv˚ardsverket (Swedish Environmental Protection Agency), Stockholm, Sweden. Navrud, S.: 2004, ‘The Economic Value of Noise Within the European Union - A Review and Analysis of Studies’. Mimeo. Nijland, H. A., E. E. M. M. Van Kempen, G. P. Van Wee, and J. Jabben: 2003, ‘Costs and Benefits of Noise Abatement Measures’. Transport Policy 10(2), 131–140. ¨ ¨ Ohrstr¨ om, E., A. Sk˚anberg, L. Barreg˚ard, H. Svensson, and P. Angerheim: 2005, ‘Effects of Simultaneous Exposure to Noise from Road and Railway Traffic’. Inter-Noise. Petersson, M.: 1999, ‘Noise-related roughness of railway wheels - testing of thermomechanical interaction between brake block and wheel tread’. Licentiate thesis, Chalmers University of Technology. Chalmers Solid Mechanics. Plovsing, B. and J. Kragh: 2000a, ‘Comprehensive Outdoor Sound Propagation Model. Part 1: Propagation in an Atmosphere without Significant Refraction’. Delta report AV 1849/00, Delta, Lyngy, Denmark. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

162

¨ Henrik Andersson and Mikael Ogren

Plovsing, B. and J. Kragh: 2000b, ‘Comprehensive Outdoor Sound Propagation Model. Part 2: Propagation in an Atmosphere with Refraction’. Delta report AV 1851/00, Delta, Lyngy, Denmark. Prop. 1997/98:56: 1998, ‘Government bill 1997/98:56: Transportpolitik f¨or en h˚allbar utveckling’. Government bill by the Swedish Government. Rosen, S.: 1974, ‘Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition’. Journal of Political Economy 82(1), 34–55. Rothengatter, W.: 2003, ‘How good is first best? Marginal cost and other pricing principles for user charging in transport’. Transport Policy 10(2), 121–130. Sansom, T., C. A. Nash, P. J. Mackie, J. Shires, and P. Watkiss: 2001, ‘Surface Transport Costs and Charges: Great Britain 1998’. Final report, Institute for Transport Studies, University of Leeds, UK. Schmid, S. A. and R. Friedrich: 2002, ‘External costs of transport noise. a bottom-up appraoch’. Inter-noise 2002. SIKA: 2002a, ‘Nya banavgifter? Analys och f¨orslag’. SIKA Report 2002:2, SIKA (Swedish Institute for Transport and Communications Analysis) and Banverket, Stockholm, Sweden.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

¨ SIKA: 2002b, ‘Oversyn av samh¨allsekonomiska kalkylprinciper och kalkylv¨arden p˚a transportomr˚adet’. SIKA Report 2002:4, SIKA (Swedish Institute for Transport and Communications Analysis), Stockholm, Sweden. SIKA: 2008, ‘Samh¨allsekonomiska kalkylprinciper och kalkylv¨arden f¨or transportsektorn’. SIKA PM 2008:3, SIKA (Swedish Institute for Transport and Communications Analysis), Stockholm, Sweden. ¨ S¨oderstr¨om, L.: 2002, Oresundsf¨ orbindelse med ett hinder mindre: Effekter p˚a integrationen vid avgiftsfrihet , Chapt. Priser och samh¨allsekonomi, pp. 35–47. Lund, Sweden: ¨ Oresundsuniversitetet. UIC: 2005. http://www.uic.asso.fr/environnement/Railways-Noise.html. Union of Railways.

International

Wall, R.: 2005, ‘Buller - En f¨orstudie om samh¨allsekonomisk v¨ardering av buller med direkt ljudm¨atning’. Mimeo, V¨agverket, Borl¨ange, Sweden. Wilhelmsson, M.: 1997, ‘Trafikbuller och fastighetsv¨arden - en hedonisk prisstudie’. Meddelande 5:45, Division of Building and Real Estate Economics, Royal Institute of Technology, Stockholm, Sweden.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

In: Railway Transportation Editor: Nicholas P. Scott

ISBN: 978-1-60692-863-9 © 2009 Nova Science Publishers, Inc.

Chapter 5

PLANNERS, COGNITION AND THE ROUTE TOWARDS IMPROVED PLANNING SUPPORT: AN EMPIRICAL STUDY INTO THE USE OF APS’S IN THE NETHERLANDS RAILWAYS René J. Jorna, Wout van Wezel and Joep Bos Faculty of Economics and Business, University of Groningen (Groningen, The Netherlands)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Abstract In many organizations, planning and scheduling are extremely important because they determine how and when the company’s resources will be used. This is especially the case in railway organizations, where many resources must be coordinated nationwide: train coaches, train drivers, ticket collectors, railroad tracks, shunting yards, shunting staff, etc. For such complex planning and scheduling situations, computer support is indispensable. In the past decades, mathematical algorithms have been incorporated in scheduling systems to improve schedules, but in many cases, this has not resulted in the expected improvements in performance. We performed experiments to investigate this problem for the decision support system that is used by staff planners of The Netherlands Railways. In the experiments, planners solve simple and complex problems with and without support of an algorithm. We report on the effect of the use of the algorithm on mental load and task performance.

1. Introduction The Netherlands Railways (in Dutch: de Nederlandse Spoorwegen; NS) daily transports one million passengers. Transportation takes place with the help of 2700 railroad carriages, which approximately run 5000 train services per day. The trains run between 384 stations in the Netherlands. The NS itself consists of several independent business units, like NS-stations, NS-real estate, etcetera, of which NS-Passengers (Dutch: NS-Reizigers) is the most important one. Among the departments in this business unit, the unit Production, taking care of logistics, is responsible for all planning and scheduling.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

164

René J. Jorna, Wout van Wezel and Joep Bos

Within the NS, four kinds of planning divisions exist. The first concerns timetables and other plans. Leaving aside the timetables, the second concerns the partitioning in planning rolling stock and planning rolling staff. The third concerns the partitioning in local planning and central planning (of stock and staff) and the last concerns the distinction in year plan (long term) and day plan (short term), again of stock and staff. Overall approximately 300 planners are continuously involved in making plans and schedules. The planning of train services is impossible without the help of advanced software support. Recently, the planners of the rolling staff departments of Year and Day Plan started to use an advanced planning system STAFF, including a newly developed algorithm to create schedules automatically. The system developers state that the algorithm in STAFF works very well and should lead to improvements in the planning process. The planners, however, indicated that the performance of the algorithm does not meet their expectations. As a consequence, the planners have resorted to manual planning again. It is not our intention in this research to shift the blame from software designer to planner or the other way around. Our basic interest is in why this misunderstanding or performance distrust takes place. If planning support systems do not function as expected, various reasons can be examined: software failure, hardware problems, incomplete modeling, or organizational culture. We focus our research on the knowledge and on the reasoning patterns of planners in interaction with STAFF. We studied eight planners of the department of rolling staff planning, especially the Day Planners, who plan personnel on trains, that is to say ticket collectors and engine drivers. In our study, we investigated the influence of the support by planning algorithms on the task performance of the planners. During our study, Day planners were studied solving two different planning assignments (with varying task complexity). The influence of task complexity of a planning exercise and the level of decision support (manual versus algorithmic) were analyzed in relation to task performance. In addition, the variations in task complexity and decision support level and its influence on mental workload were investigated. This could in turn influence task performance. In the remainder of this chapter, we will first describe the problem context. Subsequently, we provide an overview of the relevant literature. Then we give the experimental design, the results of the experiments, and a discussion of the results. We end with conclusions and directions for further research.

2. Problem Description The shift schedule of the NS contains a large quantity of types of shifts. Shifts specify the tasks of ticket collectors and train drivers. A shift may consist of the following order: Start on station A, be engine driver to station B, go from there back to station A. Have lunch for a period of 30 minutes and travel as a passenger from A to C. Go as engine driver from C to D and from there to B and then back to A. Every shift consists of a number of these tasks. Each shift needs to end in the same railway station as it started and train drivers and ticket collectors do not stay on the same trajectory the whole day. Therefore, a shift specifies a sequence of train rides, meeting the following constraints and goals: •

The maximum shift duration is 9.5 hours. Shifts that start very early or end very late may take at most 8 hours.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Planners, Cognition and the Route towards Improved Planning Support • • • • • •

165

Each shift must contain a lunch break of at least half an hour. A train driver must rest for at least 15 minutes before changing from train A to B. Shift changes must be scheduled within the original shift time period as much as possible. Shifts must have some variation (not driving between the same cities the whole day) Shift changes must be avoided as much as possible. On average shifts should be 8 hours.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The shift schedules need to be chanced frequently, for example after changes in the timetable or because of track maintenance. The eight planners on the Day Planning department are responsible for rescheduling shifts of train drivers and ticket collectors. These planners are using the planning system STAFF since two years (see Figure 1). This system has a GUI to create plans by dragging and dropping. The last half year the system includes an algorithm to automatically create plans. The planner can influence the algorithm by setting parameters for specifying bonuses or costs for e.g., shift length, taxi rides for engine drivers, overtime, etc. The use of the algorithm has not yet resulted in the expected increase in performance. The use of algorithms in advanced planning systems has been researched from different perspectives. In the next section, we elaborate on previous research in this area.

Figure 1. Screenshot of planning support program used within NS (in Dutch).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

166

René J. Jorna, Wout van Wezel and Joep Bos

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

3. Perspectives in (Advanced) Planning Systems: A Restricted Literature Review Many planners are supported by software to manually create and edit plans. Such planning software can be seen as a kind of “word processor” for planning. Plans can be copied, altered, printed, saved, and some basic calculations can be made. The software supports this kind of tasks, but the planner is the thinking actor. It is also possible that software can provide functionality to automatically generate schedules. Then the main problem solving actor is the software and not the planner. In general, there are three main approaches to schedule or plan generation (Van Wezel, 2006). We will outline the three categories shortly. First, there are approaches that focus mainly on the domain without analyzing the way in which the human planner solves the problems. In this way, the possibilities of the computer are not restricted by the knowledge or cognitive constraints of the human planner. In such approaches, characteristics of domain entities and their relations are analyzed and modeled (for example, capacity of machines, shift requirements, historical data of working hours, etc.), and an algorithm is formulated that can efficiently find a schedule that does not violate constraints. Examples are Operations Research techniques, Constraint Based Scheduling techniques, and Artificial Intelligence planning techniques. In the Operations Research community, it has long been acknowledged that the human decision maker has a strong influence on the use of techniques developed to solve planning and scheduling problems. For example, Jackson and Keys (1984, p. 476) state that “the nature of the decision makers will also greatly affect the type of solution needed to problems and the problem-solving methodology needed to reach that solution." More recently, Meredith (2001) looks back and reflects on the contribution of Operations Research on solving problems in practice. He concludes that the contribution has been limited to generic models and techniques, and that a large gap remains with the perception of problems by decision makers such as planners. Moreover, modeling the domain is too often considered a one-time-exercise, which makes the outcome rather sensitive for changes in the real world, and the method used to structure the domain is not always adequately selected (Bertrand & Fransoo 2002; Mingers & Rosenhead 2003; JORS 1986). So even within the Operations Research community, this approach is criticized for not recognizing different perspectives on the domain (Ackoff, 1979) and questioned on the real contribution of the solution approach (Corbett & Vanwassenhove, 1993). Second, in the 80’s and 90’s, knowledge based systems have been proposed to solve the above problems. These systems focus on imitating the human problem solving processes with rule bases or case bases. For this approach, the problem solving behavior of the human scheduler must be analyzed. A disadvantage of this approach is that although the system inherits the capacity of abstract reasoning that is so typical of humans, it also has the myopic fire fighting tactics that human schedulers practice (Smith 1992). An example of this approach is the ZKR-software we created for nurse scheduling (Mietus 1994; Jorna et al. 1996; Jorna, 2006). Although cognitive models are part of the analysis to develop algorithms, interaction between the user and the algorithm during problem solving gets little attention in this paradigm. As with the domain-oriented approach, the distribution of tasks between the computer and the user is mainly towards the computer.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Planners, Cognition and the Route towards Improved Planning Support

167

Third, in the 90’s, it became clear that the use of knowledge bases alone would not improve the use of scheduling algorithms in practice much. By then, both the Operations Research and Artificial Intelligence paradigms were extended to allow more interaction between the computer and the human user during the problem solving process. The focus of such mixed initiative or interactive systems is on improving the solution by establishing a coalition between the computer and the user. In this approach, not the domain or the problem solving process is the main focal point, but the roles that the computer and human can play in solving the problem. The computer and the human are both acknowledged as indispensible, and both should be allocated the tasks they are best at. Many scheduling systems that apply this approach have been described in scientific literature since the 90’s, e.g., in the proceedings of the ARPA-Rome Laboratory planning initiative workshop, and contemporary Advanced Planning Systems have implemented this approach as well (Van Wezel & Barten, 2002; Van Wezel & Jorna, 2007). Given that ample theories and methods exist to develop high quality scheduling algorithms, the main question in mixed initiative systems is the allocation of tasks to human and machine. However, theoretically founded human/computer allocation models that are specific to scheduling do not exist. In mixed initiative scheduling systems, the starting point is that algorithms can solve scheduling and planning problems better than humans can. The contribution of the mixed initiative paradigm is the explicit consideration that algorithms will never be able to solve the whole scheduling problem, that the human will always be necessary, and, consequentially, that human and computer should work together to solve scheduling problems. However, human/computer interaction in mixed initiative systems is formulated from the perspective of the functioning of the algorithm, for example: •

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

• • •

The planner can choose from a number of alternative solutions that are generated by algorithms (Lauer et al. 1994; Ulusoy & Özdamar 1996). The planner may specify weights on goal functions (Smed et al. 2000; Gabrel & Vanderpooten 2002), after which the algorithm generates a schedule. The planner can steer the backtracking process of the algorithm (Bernardo & Lin 1994). The planner can specify parameters for the algorithm (Ulusoy & Özdamar 1996; Dockx et al. 1997; Oddi & Cesta 2000; Meyers et al. 2002).

Essentially, the algorithm in the STAFF system can be categorized in the category of mixed-initiative systems; it is embedded in the GUI and the planners can specify relevant parameters to influence the outcome of the algorithm. In the field of cognitive ergonomics, this approach is called static allocation. Static allocation has been criticized for several reasons since the 80’s. There is general agreement that, given an adequate model of a scheduling problem, the computer is better and faster at finding the best conflict-free solution or a solution that violates as few constraints as possible. However, because algorithms can be continuously improved and extended, the role of the human is constantly reduced (also called the ‘leftover’ principle), ultimately resulting in a monitoring role. Such a role might lead to lower job satisfaction, complacency, and performance degradation, which can have severe consequences when the human must perform actions in case of changed circumstances, exceptions, or automation failures

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

168

René J. Jorna, Wout van Wezel and Joep Bos

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

(Bainbridge, 1983). Furthermore, allocation decisions might not be appropriate under all circumstances but situational factors are not taken into account in static allocation. Most notably, time pressure, cognitive workload, and the current status of related tasks interact with each other and influence the appropriate allocation. This problem is accounted for in many areas of human/computer interaction, for example, car driving support, aircraft piloting, airtraffic control, industrial process control, and anesthesiology (Michon, 1993; Hoc, 2000). In these examples, the design of human-machine cooperation considers criteria such as cognitive workload, situation awareness, complacency, skill degradation, risk of automation failure, trust, and cost of incorrect decisions to decide on the task allocation (Parasuraman et al., 1993). To come back to our initial research question, we wish to investigate whether and why the planners currently do not use the algorithm, which is supposed to support them, as expected. The literature on mixed initiative scheduling systems viewed from the perspective of criticism on static task allocation gives us cues on the causes. Specifically, we investigated the consequences on task performance and mental workload of using the algorithms under different conditions. Given the availability of STAFF, and given the long experiences of the planners, what causes can be detected in terms of knowledge availability and reasoning patterns that prevent planners from using STAFF? One of the causes may be related to mental load and we included this in the study. We designed, therefore, two real world assignments. The assignments are comparable to the usual NS-planning problems, but are restricted in time. Normally, solving a planning problem may take several days. We restricted this to a couple of hours. Three kinds of data were gathered in the study: quality of the solution, various aspects of reasoning patterns found in think aloud protocols during assignments, such as backtracking and number of reasoning steps, and experienced mental load. In the next section, we outline the experimental design.

4. Experimental Design 4.1. Research Design As mentioned in the previous section, the current approach in the development of algorithms is to focus on the problem that has to be solved. Adequate human/computer interaction is regarded as a consequence rather than as a prerequisite. Specifically looking at the consequences of the algorithm for the end user has not been done to date. From literature in other areas, however, we know that this kind of automation can have severe consequences for the task performance, for example because of an increased mental load. Furthermore, the advantages and disadvantages of using the algorithm might depend on the specific situation. Combining these factors, we get the following research design (Figure 2). Independent variables are task complexity and level of planning support, whereas the independent variable is task performance. The task performance is elaborated into various sub-variables, for example, the number of conflicts in the schedule, the number of shifts used, the number of taxi rides necessary, etc. Cognitive load is an intermediate variable. It is affected by the task complexity and the decision support level of the application, but has itself an effect on the task performance. The higher the cognitive load because of a high task complexity, the lower the task performance might be.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Planners, Cognition and the Route towards Improved Planning Support

169

Figure 2. Conceptual model of relationship between independent (DSL, TC) and dependent (TP) variables.

With the level of decision support, we vary whether the planner uses the algorithm or not. The task complexity is a situational factor; we will provide both a simple and a complex problem. The mental load will be measured using the NASA TLX questionnaire. The task performance will be measured in terms of reasoning steps, time to find a solution and quality of the solution.

4.2. Subjects

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The experiments were performed with 8 planners of the department Staff Day Planning. All planners have experience with the STAFF system, and three planners have more than 3 months of experience with the algorithm. The average age of the planners is 52 years, and they work on average 26 years within the Netherlands Railways. Most planners have worked on the planning department at least 5 years. Before becoming a planner, most planners used to work themselves as train driver or ticket collector. All 8 planners will solve the simple problem. As the complex problem takes approximately 8 hours to solve manually, only 2 planners solved the complex problem.

4.3. Procedure Because the planners had to solve the problem with and without the use of algorithms, two similar simple and two similar complex problems were created. By varying the sequence, learning effects were accounted for. Furthermore, there were at least two days between the two assignment sessions with a planner. The experiments were done on a standalone laptop with the same system as the planners use in their daily work. The following procedure was followed for each experiment: 1. 2. 3. 4.

De planner reads the assignment. De planner solves the problem using the system. The planner fills out the NASA TLX questionnaire. The researcher analyzes the results.

During the experiment, the researcher counts a number of activities of the planner: 1. Use of the searching functionality in the system.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

170

René J. Jorna, Wout van Wezel and Joep Bos 2. Number of times of backtracking (# of withdrawn reasoning steps). 3. Number of decisions that result in constraint violations (# of wrong reasoning steps).

4.4. Assignments: The Problems the Planners Solved Two assignments were designed. The simple one had a limited number of shifts, stations and misbalance (Table 1). The complex one was more elaborated. Because the planners had to work with STAFF, the program itself had to be adjusted. Normally, STAFF supports all shifts and stations, but in our assignments only a selection was relevant. The redesigned support of STAFF had to fulfill the following requirements: the stations and shifts have to be real, the route acquaintance of the engine drivers had to be supported, and a real problem should be used, not a toy problem. This means that tracks lengths, station names and the railway net should be realistic (see figure 3). Trains start on simplified points of time and work with an interval of 5 minutes. The new timetable had to be inserted in STAFF, because the rolling stock plan and timetable are basic for planning rolling staff. Table 1. Task complexity in assignments

Number of shifts Number of stations

Simple (1) Complex (2) 31 94 4 6

3000 serie

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

2000 serie Alkmaar

Hoorn

100 serie

Trein 4000 serie Haarlem

Amsterdam

Figure 3. Graphical and realistic display of 4 stations, the train numbers and the net/route between the stations.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Planners, Cognition and the Route towards Improved Planning Support

171

5. Results In this section, we discuss the following results. We start with the simple problem and we report for 8 planners the effects on their mental load for the support without and with algorithms. We then discuss for the same 8 planners the effects of the support without and with algorithm on their task performance. We then turn to the complex problem, and here we encountered a real life problem. It turned out that solving the complex problem took more than 6 hours for each individual planner in the without algorithm situation. We could have chosen to omit the complex problem completely. However, we wanted to continue our real world study of better forms of planning support and therefore, and in cooperation with the management and planners, we reduced the number of planners that made the complex assignment without and with algorithms to two. They made the simple and complex problem and they used support without and with algorithms (See Table 2) Table 2. Combination of kind of problem, kind of support and the participation of the planners Problem Simple problem Complex problem

Support Manually With Algorithm Manually With Algorithm

# of planners N=6+N=2 N=6+N=2 N=2 N=2

Mental Load X X X X

Performance X X X X

5.1. Simple Problem

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The simple problem was solved by all planners (N = 8). We will discuss the influence of support on mental load and on task performance. Table 3. Mental load for simple problem Planner 1 2 3 4 5 6 7 8

Manually 26 45 35 48 36 22 41 24

With algorithm 15 31 35 36 42 21 44 66

Average SD

34,6 9,8

36,3 15,5

Mental load. Table 3 shows the mental load. The average mental load for the simple problem is low (scale 1- 100), both without (manual) and with using the algorithm. A small

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

172

René J. Jorna, Wout van Wezel and Joep Bos

difference was found resulting in a little heavier mental load for the algorithm situation, but the difference was not significant (D = 1.6; df = 14; p = 0.40). Therefore, the level of support does not influence mental load when solving a simple planning problem. Task performance. Table 4 shows the task performance for the simple problem with and without using the algorithm. The first column shows the performance indicators. The second and fourth columns show the results for respectively solving manually and using the algorithm. The third column shows the difference. + indicates when there is a noteworthy difference, indicating that using the algorithm gave better results; 0 indicates that there was little or no difference. Because there were notable differences between planners, we present the standard deviation for both situations in columns 5 and 6. Table 4. Task performance for simple problem (N = 8)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Indicators (1)

Conflicts # Passenger tasks # Taxi rides >30 minutes exceeding shift length Extra shifts Empty shifts Solution time (minutes) # Reasoning steps Reasoning Steps withdrawn Wrong steps Search / Filter > Related to stock

Average Manual (2)

Algorithm vs Manual (3)

0,1 24,6 0

0 + 0

0 14,7 0

0,35 8,16 0

0 4,46 0

1 0 1,4 63,3 137,1 5 8,7 47,4 16,9

0 0 + + + + + + +

0,6 0,3 3,7 43,1 53 2,3 3,1 11,6 2,7

0,83 0 1,2 12,93 45,91 4,21 5,15 21,13 15,15

0,52 0,71 1,39 19,14 39,1 1,77 3,45 5,4 1,6

Average SD SD Algorithm Manual Algorithm (4) (5) (6)

From Table 4 it can be seen that if there is a difference between manual and algorithm, the algorithmic support has the advantage. Except for the indicators “extra shifts”, “ >30 minutes exceeding shift length”, “conflicts”, and “taxi rides”, task execution with algorithms is better than manually (t-tests were performed on differences (N = 8) and showed significant). Without much difference in mental load for the two kinds of support, STAFF with algorithm gave better results. 8 out of 12 indicators showed better performances for the support with algorithms. We will discuss the indicator differences in more detail. Most planners plan without conflicts, meaning that only one had conflicting constraints in its solution. This was with manual support (1 conflict for 8 planners = 0,1 as an average). Manual and algorithmic support therefore show no difference. This is remarkable because one should expect a worse performance for manual support. Concerning goal functions, that is to say goals or soft constraints that planners had to fulfill, the following were mentioned: amount of passenger tasks, taxi rides, excess of shift length, amount of empty shifts, and extra shifts. With algorithmic support the planners planned less passenger tasks (15 versus 25) and the amount of empty shifts were reduced. On

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Planners, Cognition and the Route towards Improved Planning Support

173

the other hand the differences with regard to the goal functions: shifts exceeding > 30 minutes, the amount of extra shifts, and the use of taxi rides - all three as little as possible were negligible. The differences between manual and algorithmic support are really visible in the time the planners used to give a solution and in the number of reasoning steps. On average, the planners execute the task much faster with algorithmic support. However, there are big individual differences. One planner planned much faster with manual support. Concerning the reasoning steps, planners used up to three times less steps with the algorithmic support. In addition, fewer steps are withdrawn, less wrong steps are made, and the search/filter function in STAFF is less used. The conclusion for the simple task is that the advantage of algorithmic support lies in the faster task execution and in the less complicate reasoning activities of the planners. The algorithmic support alleviates the thinking activity of the planners, as one might expect. On the other hand, the quality of the outcome is not positively affected by algorithmic support compared to manual support. One might have expected this, but the simplicity of the task might have prevented this advantage to take place. Another very important outcome is the individual differences of the planners on all indicators. Support for one planner really is something else compared to support for another planner.

5.2. Complex Problem

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

As we already explained, only two planners solved the complex problem. It took them between 3 and 4 hours (see table 6). We will show a qualitative analysis of mental load and task performance. Mental load. The result of effect on mental load is not conclusive (Table 5). One planner had a slightly higher mental load and one planner had a slightly lower mental load. The differences, however, are very small. Table 5. Mental load for complex problem (N = 2)

STAFF (manual) STAFF (algorithm)

Planner 1

Planner 2

Average

44 46

67 61

55,5 53,5

Task performance. Table 6 shows the results for the task performance. In column 1 the indicators are mentioned, in columns 2 and 3 the algorithmic and manual support for the two planners P1 and P2, in columns 4 and 5 the averages for algorithmic and manual support and in column 6 the comparisons, with + for algorithmic better, - for manual better and 0 for indifferent. As can be seen, the results are mixed. On most performance indicators, the use of the algorithm has a positive effect. There are two exceptions. First, the number of empty shifts is lower for planner 1 and higher for planner 2 using the algorithm. Second, the time needed increases slightly for planner 1 when using the algorithm, whereas it significantly decreases for planner 2. We will discuss the results in more detail.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

174

René J. Jorna, Wout van Wezel and Joep Bos Table 6. Task performance for complex problem ( N = 2)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Algorithmic

Manual

Algorithmic Manual

Algorithmic versus Manual

P 1*

P2

P1

P2

Average

Average

# Conflicts

0

0

0

0

0

0

0

# Passenger tasks # Taxi rides >30 min. exceeding shift length # Extra shifts # Empty shifts

62 0

34 1

25 0

87 0

48 0,5

56 0

+ -

21 11 4

1 13 12

18 14 18

14 13 5

11 12 8

16 13,5 11,5

+ + +

Time for solution (minutes)

279

91

245

335

185

290

+

# Reasoning steps # Steps withdrawn # Wrong steps Search/filter > concerning rolling stock

398 29 59 80 25

144 1 8 13 3

680 17 38 83 20

738 10 35 110 69

271 15 33,5 46,5 14

709 13,5 36,5 96,5 44,5

+ + + +

Concerning conflicts, we see no differences in the manual and algorithmic support. Both planners gave solutions in which no constraints were violated. Concerning goal functions passenger tasks, taxi rides, > 30 minutes excess, amount of extra and of empty shifts - we see that algorithmic support, except for the small difference in taxi rides, is better than manual support. However, there are big differences between the planners. Planner 1 plans more passenger tasks and planner 2 less passenger tasks in the algorithmic support and planner 2 has more empty shifts in the algorithmic support. In general, the support with the STAFF algorithmic function give better performances than manually. As with the simple problem, the time to give a solution in the complex problem is on average shorter in the algorithmic support condition. However, as can be seen in Table 6, planner 1 needs a little bit more time in the algorithmic support condition. Table 7. Task execution for planner 1 and 2 (N = 2)

Better with algortithm (+) Worse with algorithm (-) Indifferent (0)

Planner 1

Planner 2

3 7 2

9 1 2

The number of planning steps is less in the algorithmic condition compared to the manual solution. However, the planner that needed more time in the algorithmic condition used less reasoning steps. More steps are withdrawn in the algorithmic condition but that is due to planner 1. However, in general less wrong steps are made. The most important conclusion again is the big individual differences between planners, which we show in table 7. One

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Planners, Cognition and the Route towards Improved Planning Support

175

cannot in general conclude that performance is better with algorithmic support compared with manual support of STAFF.

5.3. Comparison of Simple and Complex Task (N = 2) Finally, we compare the results of the conditions and the problems with one another, alas only for two planners. We begin with the mental load results and continue with the indicators on the performances. Mental load. As one might expect, the mental load increases when the problem gets more complex. Both planners report that the algorithmic support reduces mental load, but more for the simple than for the complex problem (Table 8). Table 8. Mental load

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Simple problem Complex problem

STAFF (manual)

STAFF (algorithm)

35,1 55,3

23,1 53,7

Task performance. From Table 9 it can be seen for the two planners that task performance decreases with increasing task complexity. With the exception of conflicts and number of taxi rides, the planners perform worse for the complex task. This must of course be weighed against the longer time that is necessary to complete the complex problem. This does not imply that the solution for the complex problem is worse. In this sense the problems can not be compared. From the comparison of the simple and complex problem, and algorithmic and manual support for the two planners we conclude the following. First, planners avoid conflicts as much as possible. They do that irrespective of complexity and kind of support. Solutions with conflicts are no solutions at all. The lesson from this conclusion for support is that the trust of planners in support depends on the solution that algorithms give. Software support that yields conflicts is not trusted. Second, algorithmic support is more useful for complex problems. For simple problems, the surplus value of algorithmic support is questionable, also if we look at the results of the 6 planners, that we discussed earlier. Third, planners differ strongly, with regard to preference of goal functions, time to find solutions and in reasoning steps. The consequence of this result is that uniform support, neither in the possibilities of manual nor in that of algorithmic support, is realistic. Personalized, context dependent and individually adjustable support functions with the manual as well as in the algorithmic mode are essential.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest Ebook

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Table 9. Task execution (N=2) Planner 1 Simple* Complex Algo. Man. Algo. Man.

Planner 2 Simple Complex Algo. Man. Algo. Man.

Ave. Simple

Ave. Complex

Conflicts

0

0

0

0

0

0

0

0

0

0

Passenger task Taxi ride >30 min. excess shift length # Extra shifts # Empty shifts Algo. > extra > empty

11 0 1 0 4 1 5

16 0 0 0 2

62 0 21 11 4 16 6

25 0 18 14 18

15 0 1 0 4 0 4

19 0 1 0 3

34 1 1 13 12 13 3

87 0 14 13 5

15 0 1 0 3 1 5

52 0 14 13 10 15 5

Time for solution (minutes)

25

60

279

245

51

81

91

335

54

238

# Reasoning steps # Steps withdrawn # Wrong steps Search/filter > concerning rolling stock

26 3 4 *7 1

134 8 6 18 3

398 29 59 80 25

680 17 38 83 20

69 3 5 20 5

152 6 19 44 19

144 1 8 13 3

738 10 35 110 69

95 5 9 22 7

490 14 35 72 29

Planners, Cognition and the Route towards Improved Planning Support

177

5.4. Discussion Because of the sample size we can only draw tentative conclusions. First, the effects of the use of the algorithm depend on situational factors, i.e., problem complexity. Second, the overall results show no unequivocal positive or negative influence of the use of the algorithm on mental load. A possible explanation is that the use of cognitive resources shifts from creating a solution to interpreting the solution generated by the algorithm. The think-aloud protocols indicate support for this notion, for example in the following excerpt from an experiment using the algorithm: “I was looking why the shift is empty. Because he (the algorithm) has made it empty, because it is on top. I’ll look at Hoorn…. ….Alkmaar Hoorn, here Alkmaar Hoorn as well, this is strange. I now see the problem.” However, the experiments with planner 1 and planner 2 show that the mental load increases when the problem becomes more complex. Interestingly, planners 1 and 2 feel that the use of the algorithm decreases their mental load especially with the simple problem, whereas this was not the case for the complex problem. Thus, based on planner 1 and planner 2 who both solved the simple and complex problem, the use of the algorithm decreases the cognitive load only in the case of simple problems. Third, task performance for simple problems is positively influenced by use of the algorithm, but not unequivocally for complex problems. Although these conclusions must be interpreted with care, they nonetheless carry an important message. The simple problem was a toy problem that does not occur often in practice; actual problems are more like the complex problem, and it is there where the results are mixed.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

6. Conclusion Making algorithms to support human planners is a challenging task. From a mathematical perspective the basic stance is to look in detail at the problem structure and come up with algorithms that yield fast and high quality schedules. The user, however, is mainly concerned with aspects of usability, which include factors that are difficult to formalize in algorithms, for example, does the planner understand the outcome of the algorithm, can the schedule be explained to the staff that will execute it, can the planner change the schedule easily, etc. When a good algorithm is not properly aligned to the planner, the expected performance increase might not be reached. We have examined this problem from a cognitive perspective for a shift scheduling algorithm in the Netherlands Railways. More specifically, we analyzed mental workload and task performance of planners that solve respectively a simple and a complex problem, both with and without support of an algorithm. Tests done by the developers of the algorithm show that the algorithm, if properly used, will yield much better results than the human planners currently are able to get manually. The analysis of the experiments with the planners, however, shows that using the algorithm currently is not unequivocally better than manually solving the problem. The experiments show that on average results are slightly better, that the problems are solved faster with use of the algorithms, and that the average planners’ mental workload does not change much. However, looking in more detail at the results shows that there are significant individual differences. Varying the decision support level resulted in variable outcomes on task performance. Whether mental load decreases, the time to solve the problem decreases, or the

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

178

René J. Jorna, Wout van Wezel and Joep Bos

quality of the outcome increases, depends on the individual planner and the complexity of problem. Two planners solved a problem of which the complexity resembles reality, both manually and with use of the algorithm. Interestingly, their mental load did not decrease when they used the algorithm. We think this is a possible cue for further research. The source of mental load that planners face when manually solving a scheduling problem changes after the introduction of an algorithm; the think-aloud protocols indicate that they now spend more time on configuring parameters and interpreting the outcome. If this is confirmed in further studies, it shows that it is necessary to pay attention to the interface between the planner and the algorithm.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

References Ackoff, R. L. (1979). Future of Operational-Research Is Past. Journal of the Operational Research Society, 30(2), 93-104. Bainbridge, L. (1983). Ironies of automation. Automatica, 19, 775-779. Bertrand J.W.M. & Fransoo J.C. (2002) Operations Management Research Methodologies Using Quantitative Modeling, International Journal of Operations and Production Management, 22(2), pp. 241-264. Corbett, C. J., and Vanwassenhove, L. N. (1993). "The Natural Drift - What Happened to Operations-Research." Operations Research, 41(4), 625-640. Dockx K, De Boeck Y, Meert K (1997). Interactive scheduling in the chemical process industry. Computers and chemical engineering : an international journal. vol. 21, no. 9, pp. 925-946 Gabrel, V, Vanderpooten D (2002). Enumeration and interactive selection of efficient paths in a multiple criteria graph for scheduling an earth observing satellite. European Journal of Operational Research. vol. 139, no. 3, pp. 533-542 Hoc, J.M. (2000). From human-machine interaction to human-machine cooperation. Ergonomics, 43, 833-843. Jackson, M.C. & P. Keys, (1984). Towards a system of system methodologies, Journal of the Operational Research Society, 35, 6, pp. 473-486. Jorna, R.J. (2006). Cognition, planning and domains: An empirical study into the planning processes of planners. In: W.M.C. van Wezel, R.J. Jorna, A.M. Meystel (Eds.), Planning in Intelligent Systems: Aspects, motivations, and methods. New Jersey: John Wiley & Sons, pp. 101-136. Jorna, R.J., H.W.M. Gazendam, H.C. Heesen, W.M.C. Van Wezel. (1996). Plannen en roosteren: Taakgericht analyseren, ontwerpen en ondersteunen, Lansa Publishing B.V., Leiderdorp. JORS (1986). "Report of the commission on the future practice of Operational Research." Journal of the Operational Research Society, 37, 829-886. Lauer J, Jacobs LW, Brusco MJ & Bechtold SE (1994). An Interactive, Optimization-based Decision Support System for Scheduling Part-time, Computer Lab Attendants. Omega : the international journal of management science, vol. 22, no. 6, pp. 613-626 Meredith, J. R. (2001). "Reconsidering the philosophical basis of OR/MS." Operations Research, 49(3), 325-333.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Planners, Cognition and the Route towards Improved Planning Support

179

Michon, J. A. (Ed.). (1993). Generic Intelligent Driver Support. London: Taylor & Francis. Mietus, D.M., (1994). Understanding planning for effective decision support. PhD Thesis, University of Groningen, The Netherlands. Mingers, J., and Rosenhead, J. (2004). "Problem structuring methods in action." European Journal of Operational Research, 152(3), 530-554. Myers, K. L., Tyson, W. M., Wolverton, M. J., Jarvis, P. A., Lee, T. J., and desJardins, M. (2002). “PASSAT: A User-centric Planning Framework”. In: Proceedings of the 3rd International NASA Workshop on Planning and Scheduling for Space. Oddi A. & A. Cesta (2000). Toward interactive scheduling systems for managing medical resources. Artificial Intelligence in Medicine. vol. 20, no. 2, pp. 113-138 Parasuraman, R., R. Molloy, and I. L. Singh (1993). “Performance consequences of automation-induced ’complacency’,” Int. J. Aviation Psychology, vol. 3, pp. 1–23. Smed J, Johtela T, Johnsson M, Puranen M, Nevalainen O (2000). An Interactive System for Scheduling Jobs in Electronic Assembly. The international journal of advanced manufacturing technology. vol. 16, no. 6, pp. 450-459 Smith, S.F. (1992). Knowledge-based production management: approaches, results and prospects. Production Planning & Control, 3, 4, pp. 350-380. Ulusoy G, Özdamar L (1996). A framework for an interactive project scheduling system under limited resources. European Journal of Operations Research. vol. 90, no. 2, pp. 362-375 Van Wezel, W. & B. Barten (2002). Hierarchical Mixed-Initiative Planning Support. In: Grant, T. & C. Witteveen (Eds.), Plansig 2002. Proceedings of the 21th workshop of the UK Planning and Scheduling Special Interest Group. Delft: Delft University of Technology. Van Wezel, W. & R.J. Jorna (2008). Tracing cognition, tasks and support: The Dutch Railroad Case. Cognition, Technology, and Work, in press. Van Wezel, W., Jorna, R. J., Meystel, A. M. (2006). Planning in Intelligent Systems, Wiley, New Jersey. Van Wezel, W.M.C. (2006). Interactive scheduling systems. In: W.M.C. van Wezel, R.J. Jorna, A.M. Meystel (Eds.), Planning in Intelligent Systems: Aspects, motivations, and methods. New Jersey: John Wiley & Sons, pp. 101-136.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

In: Railway Transportation Editor: Nicholas P. Scott

ISBN: 978-1-60692-863-9 © 2009 Nova Science Publishers, Inc.

Chapter 6

REMAINING FATIGUE LIFE ESTIMATION OF EXISTING RAILWAY BRIDGES Siriwardane Sudath Chaminda1, Mitao Ohga1,a, Ranjith Dissanayake2,b and Kaita Tatsumasa1 1

Department of Civil and Environmental Engineering, Ehime University, Bunkyo-cho 3, Matsuyama 790-8577, Japan 2 Department of Civil Engineering, University of Peradeniya, Peradeniya 20400, Sri Lanka

Abstract

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The assessment of the remaining fatigue life of a railway bridge for continuing service has become more important. The present day accepted fatigue life assessment approach to railway bridges is generally based on a combination of measured stress histories, Miner’s rule and railway-code–provided fatigue curve. Even though the past strain measurements are available for major bridges, most of the old bridges do not have past strain measurements. Therefore, the application of available methods is limited to particular bridges where the information of previous loading histories is available. Furthermore, in the case of existing railway bridges where the detailed loading history is known, Miner’s rule might provide incorrect results because of its omission of load sequence effect. Therefore, it is inaccurate to use Miner’s rule for the remaining fatigue life estimation of railway bridges because most railway bridges are subjected to variable amplitude loadings. Meanwhile, a new damageindicator–based sequential law was developed to capture the loading sequence effect of variable amplitude loads more precisely. The authors have applied this theory to estimate the remaining fatigue life of railway bridges in their previous publications, and it was proven that sequential law is more applicable to railway bridges than the previous theories. However, there are more issues to be addressed in generalizing sequential law as it applies to railway bridges. This chapter therefore proposes a reasonably accurate method to predict the past stress histories from present day measured strains and a method to apply the new damage-indicator– based sequential law for remaining fatigue life estimation of railway bridges. Initially, this chapter describes the proposed method to predict the past stress histories. Then, the method of a b

E-mail address: [email protected] E-mail address: [email protected]

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

182

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al. application of new fatigue law is described comprehensively. It describes i) the basic concept of fatigue law, ii) a technique that utilizes transfer of the partially-known code-provided Wöhler curve to the fully-known curve, iii) the proposed extension of new fatigue law to evaluate secondary stress- (stress concentration- ) based fatigue life of bridge connections and iv) experimental verifications of proposed methods. Then the proposed method is applied to estimate the remaining fatigue life of an old existing railway bridge. Finally, comparisons of the results are made with Miner’s rule-based previous estimation. Hence validity and applicability of the proposed approach are discussed.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

1. Introduction In the past two decades, a significant amount of effort has been directed towards the development of structural health monitoring and non-destructive assessment methods to manage civil structures more efficiently (Sherif et al., 2006; Jung et al., 2004). Recent sudden failures of structures have also occurred. Some examples of recent unexpected failures are the collapse of the Interstate 35W bridge in Minneapolis, Minnesota in July 2007; the Hoan Bridge failure in Milwaukee, Wisconsin in 2000; and the roller coaster accident in Suita, Japan in 2007. Among these failures, fatigue is cited as one major mode that especially causes unexpected failures of structures. Fatigue is the progressive, localized, permanent structural change that occurs in materials subjected to fluctuating stresses and strains that may result in cracks or fractures after a sufficient number of fluctuations. Fatigue fractures are caused by the simultaneous action of cyclic stress, tensile stress and plastic strain. Structures that are subjected to fluctuation of stress are liable to suffer from fatigue, and this may be caused by loads, which are very much lower than those that would be necessary to cause failure under a single application. Out of various forms of structural failures applicable to bridges, the railway bridges are most likely to fail due to the effect of fatigue, because stresses of components are generally subjected to fluctuation with moving traffic. Today, many railway bridges are nearing the end of their design fatigue lives. Therefore, rail authorities all over the world are paying special attention to the remaining fatigue life of these bridges. Furthermore, the fatigue behaviour of wrought-iron and older steels, which were mainly used for the construction of these bridges, is not well known. These observations, coupled with the lack of information on loading history of these bridges, raise questions about their fatigue performance (Imam et al., 2005). As a result, the assessment of remaining fatigue life of a railway bridge for continuing service has become more important than ever, especially when making decisions regarding structure replacement and other major retrofits. The major objective of this chapter is to furnish a comprehensive understanding of how to predict the remaining fatigue life of railway bridges more precisely. Initially, the chapter describes the currently available fatigue life estimation method in detail. The deficiencies with shortcomings of the presently-accepted method have been described as a next step. Then, a new method is proposed to improve the accuracy of predicted fatigue life. The verification of the proposed method is conducted by comparing the predicted results with experimental results. Then all of these methods are applied to estimate the remaining fatigue life of an old existing railway bridge. Finally, comparisons of the results are made and, hence, validity and applicability of the proposed method is discussed.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

183

2. Available Method The present day accepted fatigue assessment approach to railway bridges is generally based on a combination of measured stress histories under actual traffic load (Köröndi et al., 1998), Miner’s rule (Miner, 1945; Suresh, 1998) and railway-code–provided fatigue curve (also referred to as S-N or Wöhler curve). The following subsection comprehensively describes this combination.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

2.1. Stress Evaluation In fatigue life estimations, it is essential to determine the stress ranges generated by the passage of trains over the bridge. Therefore, it is required to know the stress cycles (stress histories) distributions of all the critical components of the bridge for trains that are included in the timetables obtained from bridge owner for present and past rail traffics over the bridge. These stresses are determined from the continuous strain measurements of identified fatigue critical components of the bridge from as built stage. When weights of rail traffic are varying during a certain time interval (generally called duration of loading block) and a similar pattern is repeating during a particular period, it is enough to measure the strains throughout each considered time interval at each period. These stresses are generally called “primary stresses” and they might be uniaxial or multiaxial state, which generally depend on the geometric properties of the critical component. As an example, these critical stresses are mostly uniaxial for truss girder bridges. But for plate girder structures, it might be biaxial. In such situations, the multiaxial stresses are transformed into equivalent uniaxial stress (generally called as effective stress which characterizes the deformation of the material) cycles, which should generate the same fatigue life as that due to the multiaxial. Finally, these stress histories (uniaxial or equivalent uniaxial stress cycles) were converted into stress ranges by using the reservoir counting method (BS 5400 part 10,1980) and hence the stress range histogram can be obtained for all critical components.

2.2. Determination of Fatigue Curve The stress concentration effects around the discontinuities such as connections, notches or cracks against the primary members of bridges were found to be one of main reasons for fatigue damage (Fisher et al., 1984). Further it has been identified that the rotational fixity of connections and the variation in the clamping forces at the connections (Akesson, 1994) are the major causes leading to fatigue damage of structures. However, it cannot easily predict secondary stresses (stress concentration effects) around the discontinuities. In such situations, classification of details (detail class) at the connection (BS 5400 part 10, 1980), are considered to capture the fatigue damage due to the secondary stresses around the discontinuities. The classification of each component of a detail depends upon the direction of fluctuating stress relative to the detail, location of possible crack initiation, geometric arrangement, the method and standards of manufacture (quality of the workmanship) and current condition of the detail (Table 17 of BS 5400 Part 10, 1980). Different classes of fatigue curves are illustrated related to mentioned classification of details. Further, each class

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

184

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

of curve also is again classified in to different curves based on required probability of failure of the fatigue test data. Curve, which lies along the mean of the test results usually called as mean curve and it has a 50% probability of failure. The curve, which lies along the mean minus twice standard deviation of test data, exhibits 2.3% probability of failure and this is called “design curve”. However, currently there are two different ways to predict an appropriate fatigue curve. These are described below. •

•

By identifying the real detail of the component (type of the discontinuity such as connection, etc.), the suitable detail class is selected. Then relevant curve is directly chosen from the code by considering required probability of failure. In present, this way is widely used. If the correct stress concentration factor at the detail (connection, discontinuity, etc.) is able to obtain from numerical analysis or, etc., the code provided curve that describes the classification of no details (class for plain section, for example Class B curve in the BS 5400 Part 10, 1980) is modified using appropriate stress concentration factor. However, this might provide more accuracy than the previous method.

Hence the number cycles to failure (N) at each stress range, which was obtained in section 2.1 can be obtained from the selected S-N curve.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

2.3. Fatigue Life Estimation The cumulative damage law of Miner’s rule (Miner, 1945; Suresh, 1998) is employed to obtain the remaining fatigue life of each critical components of the bridge. In case of most railway bridges, the loadings vary during certain time interval (named as duration of loading block) and the similar pattern is repeating during particular period (number of years). The number of yearly repetition of various stress ranges (n) for each time period can be found using both stress cycle distributions and the frequency of rails during the loading block. For the particular ith period of age, Miner’s summation of fatigue damage, α i = ( n N ) i * (Number of years for period where similar loading block is repeating). Miner’s summation of the total cumulative fatigue damage for all periods α = α i where i = 1 to k which is equal to number of divided periods of age. Therefore remaining Miner’s summation of fatigue damage at present= (1 − α ) . Hence the remaining life of each member, assuming that the future sequence and the weight of rail traffic is the same as that for the present period = (1 − α ) / α present years, where α present = ( n N ) i .

∑

∑

∑

3. Deficiencies of the Current Method The above method (section 2) has been used continuously from several decades. However, experiences from engineering practices have indicated that fatigue analysis based on current method produces unrealistic prediction and also difficult to apply in some situations. Indeed consideration of the current method, recently researchers have highlighted that some of deficiencies of the current method causes to produce unrealistic output and also causes to

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

185

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

reduce common applications too. Following content describes these deficiencies in detail for complete understanding of the reader. 1. The fatigue evaluation based on field measured stress range histograms under actual traffic load (section 2.1) proves to be a more accurate and efficient method for existing bridges (Köröndi et al., 1998). Generally, past strain measurements are available for major bridges. Most of the old bridges do not have past strain measurements. Therefore, application of available methods is limited to particular bridges where the information of previous loading histories is available (Zhou, 2006; Alampalli et al., 2006; Roeder et al., 2006; Li et al., 2001, 2001, 2001, 2002, 2003 & 2006; Agerskov et al., 1999; Suresh, 1998). And from an extensive literature review the authors have found that there is no generalized method for evaluating the fatigue life of these old bridges. 2. Even though, the detail class (section 2.2) is determined by considering varies factors such as, direction of fluctuating stress, location of possible crack initiation, geometric arrangement, quality of the workmanship and current condition, it might give over or under prediction to real secondary stress state. As a result, fatigue lives estimated using the code given design S-N curves exhibits pessimistic results. Therefore it is important to investigate accurately the fatigue damage due to secondary stresses at the connections or discontinuities of existing railway bridges. 3. Miner’s rule, which considered as the fatigue criterion for previous method (section 2.3), has always been acknowledged as a simplification that is easy to use in design where detailed loading history is unknown. However under many variable amplitudeloading conditions, life predictions have been found to be unreliable since it does not properly take into account the loading sequence effect (Suresh, 1998; Dattoma et al., 2006; Mesmacque et al., 2005). Therefore, it is uncertain to use the Miner’s rule for remaining fatigue life estimation of railway bridges because most railway bridges are subjected to variable amplitude loadings. Some of fatigue theories such as fatigue model based on continuous damage mechanics (Li et al., 2001), non-linear fatigue damage model (Li et al., 2002) etc. are originated to over come this shortcoming of Miner’s rule. But the application of these theories to railway bridges is found to be very few, because the determination of model parameters of these theories depends on complex experiments. Meanwhile, a new damage indicator-based sequential law (Mesmacque et al., 2005) was developed to capture the loading sequence effect of variable amplitude loads more precisely. The authors have applied this theory to estimate the remaining fatigue life of railway bridges in their previous paper (Siriwardane et al., 2008) and it was proved that sequential law is more applicable to railway bridges than the previous theories. However, there are some more to be attended to generalize sequential law in application of railway bridges and these points are described below. •

When using the sequential law to estimate fatigue life, a fully known Wöhler curve (full range of number of cycles) is essential to use as the related fatigue curve. The corresponding code provided curves, only describes stress ranges, which correspond to more than ten thousands of failure cycles (usually called as

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

186

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

•

•

partially known Wöhler curve). Therefore, it is important to have a generalized method to transfer partially known curve to fully known S-N curve. Even though the sequential law has been compared with the experimental fatigue life of some materials (Mesmacque et al., 2005), fatigue damage evaluation has not been compared with real damage of materials. Therefore, it is important to verify the sequential law by comparing the theoretical damage with real damage of some materials because remaining life basically depends on current state of fatigue damage. In the authors’ previous work, sequential law has been applied to estimate secondary stress based remaining life of a connection. But, the used extension of Sequential law for multiaxial fatigue has not been verified with experimental results. Therefore, it is important to verify proposed extension of sequential law for multiaxial fatigue by comparing predicted lives with experimental results.

Mentioned deficiencies and shortcomings of the current method highlight the necessity of precious method to estimate remaining fatigue life of railway bridges. Therefore following sections describes the proposed methods to estimate reaming fatigue life of railway bridges in detail.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

4. Introduction to a New Method This section outlines the summary of proposed method to estimate the remaining fatigue life of existing railway bridges where the past strain measurements are not available. This method fundamentally based on combination of predicted primary or secondary stress histories, uniaxial or multiaxial sequential law and fully known fatigue curve which obtained from code or experimentally. Following key differences are highlighted between previous method (section 2) and the proposed method. 1. Evaluation of past stress histories are not based on past measured strains. A reasonably accurate procedure to predict the past and future stress histories from present day measured strains is described in proposed method. 2. The partially known Wöhler curve has been extended to a fully known curve. The technique, which utilizes transfer of the partially known code given or experimentally obtained Wöhler curve to a fully known curve, is clearly indicated. 3. The new damage indicator based-sequential law is utilized instead of Miner’s rule, which was used as the fatigue theory in the previous method. The following section of the chapter describes the proposed method for remaining fatigue life estimation in detail.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

187

5. Proposed Remaining Fatigue Life Estimation Method The following subsections describe the contents of the proposed remaining fatigue life estimation method. Initially it describes the newly proposed technique to predict stress histories, and then describes the concept of sequential law in detail. The technique, which utilized to obtain the fully known fatigue curve, is also discussed. Then details related to proposed extension of sequential law for multiaxial fatigue are mentioned under separate subsection. Prior applying this method to a railway bridge, the sequential law and fully known S-N curve determination procedure are verified by comparing theoretical damage behavior with real fatigue damage behavior of few materials. Accuracy of the extension of sequential law for multiaxial fatigue is also checked by comparing predicated life with experimental life. These two verifications are also mentioned in separate section onward. Finally, the proposed method is applied to estimate the remaining fatigue life of an old existing railway bridge.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

5.1. Stress Histories Prediction Method This section describes the different method to predict the past and future stress histories from present day measured strains of exiting bridges. The method is mostly applicable to the bridges where the information of previous stress histories is not available or the bridges which were never subjected to strain measurements. It is mainly based on appraisal of particular bridge by attending condition evaluation, FE analysis, material testing, experimental static and dynamic load testing. Initially a condition survey has to be carried out to assess the present geometric condition and damages. Generally, it consists of detailed visual examination, in-situ measurements of each component of the bridge and non-destructive field examinations. Then laboratory tests are carried out to determine the current mechanical properties and chemical composition of the bridge materials. Then static and dynamic load testing is recommended as the next major step to study the real behavior of the bridge under various load combinations. The obtained results are used to develop a proper analytical model and further assists in evaluating actual dynamic factors of each structural component. Finally the bridge is subjected to finite element (FE) analysis under test and actual loadings to determine stresses and deflections, as well as variations of stresses under moving loads. Material properties which are obtained through laboratory tests and current geometric properties obtained from condition assessment are applied to the FE model for more realistic outputs. The validation of the FE model has to be done by comparing the results from analysis with those from field-tests. The FE model which gives better comparison to load test results is named as “validated analytical model”. Hence, validated analytical model is used to obtain past and present static stress histories due to passage of trains specified by the owner. Due to the dynamic effect of moving trains, the actual working stresses should be higher than the analytical static stress. Therefore dynamic factors are used to multiply the static stress to get the service stresses of each member. Finally, the stress histories have to be converted into stress ranges using the reservoir counting method (BS 5400, part 10, 1980). The described stress evaluation method is briefly summarized as shown in Figure 1 to make the idea more clear.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

188

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

RAILWAY BRIDGE

Condition survey

Material Testing

Filed Load Testing

Present geometric details, damages and corrosion

Current mechanical properties and chemical composition of materials

Present stresses, deflections, accelerations and dynamic factors

FE Analysis

Validated Analytical Model

Past /Future Stress Cycles

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Reservoir Counting Method

STRESS RANGES Figure 1. Flow chart for the past /future stress evaluation method.

5.2. Fatigue Criterion: New Damage Indicator-Based Sequential Law A new damage indicator-based sequential law (Mesmacque et al., 2005) was recently originated to produce a more realistic fatigue life in case of variable amplitude loading situations. The hypothesis behind this fatigue law is that if the physical state of damage is the same, then fatigue life depends only on loading condition. In the case of bridges, it can be said that if the connection detail type and environment are the same, only the loading condition matters in fatigue. A detailed description of the damage stress model and the definition of damage indicator, Di is described in the corresponding paper (Mesmacque et al., 2005). Here only the concept is summarized with an algorithm for comprehension (flow chart described in Figure 2). Suppose a component is subjected to a certain stress amplitude or stress rangeσi for ni number of cycles at load level i and Ni is the fatigue life (failure number of cycles) corresponding toσi (Figure 3).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges A new damage indicator based sequential law ni number of cycles atσi stress level

If σ i > σ ∞

No

Ni failure number of cycles at σ i stress level (from Wöhler curve) NiR=Ni-ni: Residual life

σ (i) eq : Damage stress for NiR number of cycles (from Wöhler curve) Di

=

σ ( i ) eq − σ i σu −σi D=Di

No If D σ ∞

No

σ (′i +1) eq associated number of cycles N/(i+1)R (from Wöhler curve) N (i+1) R =N/(i+1) R – ni+1: Residual life

σ ( i +1) eq :Damage stress for N(i+1)R number of cycles (from Wöhler curve) Di +1 =

σ (i +1) eq − σ i +1 =D σ u − σ i +1 i ⇒ i+1

Figure 2. Flow chart for new damage indicator based sequential law.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

189

190

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

Hence, the residual life at load level i can be obtained as (Ni−ni). The stressσ(i)eq which corresponds to the failure life (Ni−ni) is named as ith level damage stress amplitude or stress range (otherwise can be introduced as stress amplitude or stress range relevant to the residual life). Hence, the new damage indicator, Di is stated as,

Di =

σ ( i ) eq − σ i σ u −σ i

(1)

whereσu is the intercept of the Wöhler curve with the ordinate at one-quarter of first fatigue cycle. Furthermore, it can be stated that, σ u is the ultimate tensile strength amplitude or range for rotating bending test-based S-N curves and it is the ultimate shear strength amplitude or range for torsional fatigue test-based S-N curves. Same damage is then transformed to load level i+1 and hence damage equivalent stress at level i+1 is calculated with the relation,

Di =

σ (i ) eq − σ i σ (′i +1) eq − σ i +1 = σu −σi σ u − σ i +1

(2)

Further simplification of Eq. (2),

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

σ (′i +1) eq = Di (σ u − σ i +1 ) + σ i +1

(3)

where σ (′i +1) eq is the damage equivalent stress amplitude or stress range at the level i+1. Thus the corresponding equivalent number of cycles to failure N (′i +1) R can be obtained from the Wöhler curve as shown in Figure 3. The σ i +1 is the amplitude or range of applied stress at the level i+ 1and suppose that it is subjected to n(i +1) number of cycles, then the corresponding residual life at the load level i+1, N (i +1) R is calculated as, N (i +1) R = N (′i +1) R − n(i +1)

(4)

Hence the damage stress amplitude or stress rangeσ(i+1)eq ,which corresponds to N (i+1) R at load level i+1, can be obtained from the Wöhler curve as shown in Figure 3. Then the cumulative damage at load level i+1 is defined as,

D( i +1) =

σ (i +1) eq − σ i +1 σ u − σ i +1

(5)

At the first cycle the damage stress amplitude or rangeσ(i)eq is equal to applied stressσ1 and corresponding damage indicator becomes Di=0. Similarly at last cycle, the damage indicator becomes Di=1 whenσ(i)eq is equal toσu. Therefore, the damage indicator is normalized to one (Di=1) at the fatigue failure of the material and same procedure is followed until Di=1. Here, the defined fatigue failure is the time taken for the occurrence of the first through-thickness crack at the location of maximum stress of the structural component. In the

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

191

case of railway bridge components, it can probably be taken as the time taken for initiation of crack near a connection (rivet or bolt).

log(σ )

σu

σ (i +1) eq σ (′i +1) eq σ (i) eq σi (0,0)

N ( i +1) R N (′i +1) R ( N i − ni ) N i

log( N )

Figure 3. Schematic representation of parameters in Wöhler curve.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

5.3. Fully Known Fatigue Curve Prediction Method Generally, the S-N curves, which are obtained from the assessment code or fatigue tests of present material, are more appropriate for existing railway bridges. However, the chosen fatigue curves only describe stress ranges, which are corresponding to more than ten thousands of failure cycles (usually called the partially known Wöhler curve). In the case of sequential law, it is essential to use the Wöhler curve for full range of the number of cycles. Therefore, the chosen partially known Wöhler curve, which is obtained from assessment code or experimentally, has to be transferred to fully known Wöhler curve. This section describes how this partially known curve can be transferred to full range with reasonable accuracy. This method mainly based on Kohout and Vechet Wöhler curve modeling technique (Kohout et al., 2001). Initially the available partially known curve has to be drawn in the log-log plot as shown in Figure 4. Then draw the three important straight lines, Line 1: Asymptote σ = σ u for the low cycle region (horizontal line across the ultimate strength), Line 2: Asymptote σ = σ ∞ for the high cycle region (horizontal line across the fatigue strength), Line 3: Tangent for the region of finite life described by the equation of partially known curve. The points of intersection of the tangent (Line 3) with the asymptotes (Line 1 & 2 respectively) occur at N=B and N=C. Hence, the equation of fully known curve can be written as the form of,

⎛N +B⎞ σ = σ∞⎜ ⎟ ⎝ N +C⎠

b

where b is the slope of tangent in the region of finite life.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

(6)

192

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

log(σ )

Line 3

σu

Line 1

b = tan α

σ∞ (0,0)

α

Line 2

B

C

log( N )

Figure 4. Graphical representation of fully known Wöhler curve modeling methodology.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

5.4. Extension of Sequential Law for Multiaxial Fatigue The described sequential law is extended to evaluate the fatigue damage due to multiaxial stresses. For this extension, the simple as well as the commonly used approach has been considered. It basically consists of transforming cyclic multiaxial stresses into equivalent uniaxial stress (generally called as effective stress which characterizes the deformation of the material) amplitude or range, which should generate same fatigue life as that due to the multiaxial stresses. The equivalent stress amplitude or range is then used to enter a uniaxial SN curve to determine the damage indicator Di of the sequential law. The usual methods for making such transformations are extensions of von Mises (octahedral shear stress) or Tresca (maximum shear stress) yield criteria for proportional loading conditions (Suresh, 1998; Mesmacque et al., 2005; Sigley et al., 1989; Chamat et al., 2007). Here, von Mises yield criteria is used for the transformation. Hence multiaxial stress fields at critical locations are considered in terms of equivalent von Mises stress amplitudes or ranges. The following steps are performed to predict the multiaxial fatigue life. •

Initially the von Mises stress (effective stress) histories due to passage of trains for critical location are obtained. Let σ x , σ y and σ z denote the three normal components of stress applied to fatigue specimen and let τ xy , τ yz and τ zx denote the three shear stress components. Hence, effective stress can be determined from the following general relation.

σe = •

1 2

(σ x − σ y ) 2 + (σ y−σ z ) 2 + (σ z − σ x ) 2 + 6(τ xy2 + τ yz2 + τ zx2 )

(7)

Then the von Mises stress histories are converted into effective stress ranges and effective mean stress by using the reservoir counting method (BS 5400, part 10, 1980). The effective stress range and mean stress are determined respectively by following equations which are derived only for the proportional loading conditions.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

193

σ r ,e = σ max,e − σ min,e

(8)

σ m ,e = (σ max,e + σ min,e ) / 2

(9)

where σ max,e and σ min,e are the maximum and minimum effective stress in considered effective stress range. •

The modified Goodman equation (Suresh, 1998) was used to do the mean stress correction of each effective stress range using following relation.

σ r ,e = σ r ,e |σ

1m = 0 σ 2 m =0 σ 3 m =0

⎧ σ m ,e ⎫ ⎨1 − ⎬ σu ⎭ ⎩

(10)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The σ u is the ultimate tensile strength of the material. Hence, the effective stress ranges are transferred to equivalent fully reversed uniaxial cyclic stress ranges. •

The code provided curve that describes the classification of no details (class for plain section, for an example Class B curve in the BS 5400 Part 10) or fully reversed tension-compression fatigue test based partially known S-N curves should be modified to full range using mentioned method in section 5.3.

•

Finally these equivalent fully reversed cyclic stress ranges are utilized instead of uniaxial stress at the sequential law (section 5.2). Hence damage indicator Di can be obtained and the remaining fatigue lives are estimated for critical locations by following the procedure mentioned in section 5.2.

Here it is important to note that the described procedure is applicable to critical points, which are subjected to proportional loading conditions. In other words, it can be stated that this procedure is applicable when the state of stress is at the elastic limit of material.

6. Experimental Varifications The Sequential law has been compared with the uniaxial experimental fatigue life of some materials in previous studies (Mesmacque et al., 2005). However, the fatigue damage evaluation has not been compared with real damage of materials. In similar vein, proposed extension of Sequential law for multiaxial fatigue has also not been verified with experimental results. Therefore, following the subsections verify i) the damage behavior of sequential law by comparing the theoretical damage with real damage of some materials and ii) the proposed extension of sequential law for multiaxial fatigue by comparing predicted life with experimental results.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

194

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

6.1. Verification of Damage Behavior Two kinds of experimental data (Shang et al., 1999) of constant fatigue amplitude for the normalized 45 steel and 16 Mn steel are used to verify the sequential law and associated fully known Wöhler curve modeling technique. For normalized 45 steel, the yield strength and the ultimate tensile strength are 371.7 MPa, 598.2 MPa respectively. For 16 Mn steel, the yield strength is 382.5 MPa, and the ultimate tensile strength is 570.7 MPa. The damage variable D, was experimentally determined by measuring the static relative ductility change of material.

1300 σ

Stress Range (MPa)

1000 900

⎛ N + 200 ⎞ = 560⎜ ⎟ ⎝ N + 600000 ⎠

−0.089

800 700 600 From the PSN curve (Zheng et. al, 1995) Predicted curve

500 400

100

102

104 Number of cycles N

106

108

1300 σ

1000 900

Stress Range (MPa)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5. Predicted S-N curve for 16 Mn steel.

⎛ N + 30000 ⎞ = 580⎜ ⎟ ⎝ N + 20000000 ⎠

−0.112

800 700 600 From the PSN curve (Zheng et. al, 1995) Predicted curve

500 400

100

102

104 106 Number of cycles N

108

Figure 6. Predicted S-N curve for 45 C steel.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

195

The partially known S-N curves of above materials were obtained from different research papers (Zheng et al., 1995 & 2005). Then these curves were transferred to fully known S-N curve using the described method in section 5.3. The obtained fully known curves for two materials are as shown in Figure 5 & 6. The sequential law is applied to obtain the theoretical damage evolution. The comparisons between the damage evolution curves and the experimental results are shown in Figure 7 & 8. This comparison shows that there is a good agreement between sequential law predicted damage evolution and the real fatigue damage of these materials than Miner’s rule predicted damage. This verification reveals that the sequential law and associated fully known Wöhler curve modeling technique can be applied to obtain reasonable accurate remaining fatigue life of existing railway bridges. 1.4

1.4 From sequential law at 337.1 MPa

1.2

From Minre's rule at 337.1 MPa

1 0.8 0.6

1.2

From the Miner's rule at 373.5 MPa

0.8 0.6 0.4

0.2

0.2

0

0 0.2

0.4

0.6

0.8

From Sequential Law at 373.5 MPa

1

0.4

0

1

0

0.2

0.4

N/Nf

(a)

0.6

0.8

1

N/Nf

(b)

Figure 7. Comparison of theoretically predicted fatigue damage evolution with experimental damage of 16 Mn steel. (a) σ a = 337.1 MPa; (b) σ a = 373.5 Mpa. 1.4

1.4 Experimental 405.8 MPa (D.G.Shang et al. 1999)

Experimental 330.9 MPa (D.G.Shang et al. 1999)

1.2

From Sequential Law at 330.9 MPa

Fatigue Damage D

Fatigue Damage D

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Experimental 373.5 MPa (D.G.Shang et al. 1999)

Fatigue Damage D

Fatigue Damage D

Experimental 337.1 MPa (D.G. Shang et al.1999)

From the Miner's rule at 330.9 MPa

1 0.8 0.6

From the Miner's rule at 405.8 MPa

1 0.8 0.6

0.4

0.4

0.2

0.2

0

0 0

0.2

0.4

0.6

(a)

0.8

N/Nf

1

From Sequential Law at 405.8 MPa

1.2

0

0.2

0.4

0.6

(b)

0.8

1

N/Nf

Figure 8. Comparison of theoretically predicted fatigue damage evolution with experimental damage of 45C steel. (a) σ a = 330.9 MPa; (b) σ a = 405.8 Mpa.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

196

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

6.2. Verification of Extension of Sequential Law for Multiaxial Fatigue The multiaxial fatigue test results of two materials are compared with theoretically predicted fatigue life in this section. The two materials are 18G2A and 10HNAP steels. The first one is a low alloy higher-strength steel for building and welded structures, such as bridges, high pressure pipelines of high diameters, cranes, overhead cranes, elements of cranes and ships etc. The other material is a low-alloy of higher resistance to atmospheric corrosion structural steel. The tests were performed, in the high cycle fatigue regime under variable amplitude combined bending and torsion loading, at Opole university of technology (Marciniak et al., 2008). The normal and shear stresses are in phase loading (proportional loading). The dominating loading frequency was 20 Hz. The time interval of loading repetition (duration of loading block) is equal to 2000 seconds. Fatigue tests were performed under two proportional loading combinations for 18G2A steel (Table 1). For 10HNAP steel, three proportional loading combinations were tested as shown in Table 1. For all these loading combinations, the nominal normal and shear stresses historieses for each loading blocks (duration of 2000 Sec) were received from the Opole university of Technology. The fragments of nominal normal and shear stresses histories, obtained across the critical cross section of one loading combination, are drawn in Figure 9 to make an idea about variable amplitude nature of loading histories. The partially known fully reversed tension-compression S-N curves of above materials were obtained from different research papers (Karolczuk et al., 2004; Kluger et al., 2004).

Table 1. Life times and nominal maximum stresses of variable amplitude multiaxial fatigue test

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Material 18G2A Steel

10HNAP Steel

Designation for test NWL 1

Maximum bending stress 219

Maximum shear stress 219

NWL 4

360

180

KWL 2

212

212

KWL 4

342

171

KWL 8

320

160

Experimental fatigue life(s) 99600 78240 134700 148720 155520 167020 48060 41400 45240 37200 100560 95520 106240 92940 99300 84600 99900 93900 84420 123660 122890

Average fatigue life(s) 130630

42795

98948

90705

123275

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Bending Stress (MPa)

Remaining Fatigue Life Estimation of Existing Railway Bridges

200 100 0 -1 0 0 -2 0 0

4

5

(a)

T im e ( S e c )

6

Torsional Shear Stress (MPa)

3 00 2 00 1 00 0 -1 0 0 -2 0 0 -3 0 0

4

5

(b)

T im e ( S e c )

6

1300 1000 900

Stress Range (MPa)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 9. Fragments of history of nominal stress for: (a) bending and (b) torsion for NWL4.

⎛ N + 855 ⎞ ⎟ ⎝ N + 2320000 ⎠

−0.122

σ = 408⎜

800 700 600 500 400 100

102

104 106 Number of cycles N

108

Figure 10. Predicted S-N curve for 18G2A steel.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

197

198

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al. 1300 Stress Range (MPa)

1000 900

⎛ N + 615 ⎞ ⎟ ⎝ N + 1281000 ⎠

−0.105

σ = 504⎜

800 700 600 500 400 100

102

104 106 Number of cycles N

108

Figure 11. Predicted S-N curve for 10 HNAP steel.

Calculated life (Sec)

106

Previous Method (Miner's Rule) Proposed Method (Sequential Law)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

105

104 4 10

105 Experimental life (Sec)

106

Figure 12. Comparison of calculated life with experimental life.

Then these curves were transferred to fully known S-N curve using the described method in section 5.3. The obtained fully known curves for two materials are as shown in Figures 10 & 11. The proposed extension of sequential law for multiaxial fatigue (section 5.4) is applied to obtain the theoretical fatigue life of each and every mentioned loading combination. The comparisons between the calculated life and the experimental results are shown in Figure 12. This comparison shows that there is a good agreement between sequential law predicted multiaxial fatigue life and the real fatigue life of these materials than Miner’s rule predicted life. This verification reveals that the proposed extension of sequential law of multiaxial fatigue and associated fully known Wöhler curve modeling technique can be applied to obtain reasonable accurate fatigue life of materials which are subjected to multiaxial state of stress.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

199

7. Bridge Description and Prediction of Stress histories The selected bridge is one of the major railway bridges in Sri Lanka spanning 160 m (Figure13). It is a six span-riveted bridge with double lane rail tracks having warren type semi through trusses, supported on cylindrical piers. The bridge deck is made of wrought iron and the piers are made of cast iron casings with infilled concrete. The bridge was constructed in 1885. Details of trains carried by the bridge and their frequencies illustrate that the bridge is suffered from variable amplitude loading. Following subsections are related to prediction of past and future stress histories of the bridge. Sections 12 and 13 comprehensively describe the method of application of proposed methods to estimate the remaining fatigue life of the bridge.

Figure 13. General views of the riveted bridge.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The proposed method in section 5.1 is applied to obtain the reasonable accurate past, present and future stress histories of critical components of the bridge by performing extensive condition survey, laboratory testing, field-testing, and analytical work.

7.1. Condition Survey The condition survey revealed that some places of the bridge have been subjected to mild corrosion due to the absence of anti corrosive coating (Figure 14). No visual cracks were observed in any component of the super structure. In situ measurements of member sizes, connections and support bearings verified the fact that the existing drawings were applicable and only few significant variations were observed.

Figure 14. Some identified corroded locations of the bridge.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

200

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

7.2. Material Testing The sampling of materials, specimen preparation and testing were carried out according to the ASTM standards. The chemical analyses as well as microscopic examinations lead to the conclusion that the bridge super structure material is wrought iron. The obtained values for elastic modulus, yield strength, ultimate strength in tension, fatigue strength and density are 195 GPa, 240 MPa, 383 MPa, 155 MPa and 7600 kg/m3, respectively.

7.3. Field load Testing Static and dynamic load testings were performed to study the real behavior of the bridge under various load combinations. The in situ measurements were performed using two M8 engines, each weighing 1120 kN, which is the heaviest rail traffic in current operation. The bridge was instrumented with strain gauges placed at selected locations to measure normal strains. In addition, the triaxial vibrations were recorded at several locations using accelerometers. Displacement transducers were used to measure vertical deflection at three places around the mid span area of the bridge. The measured locations are shown in Figure 15. C

C

S10 S1

S11

S5

S3

S12T, B

S21T, B S20T, B

S7 S16T, B

S2

S6

S4

S14

S19T, B

D2

S18T, B D1

S15T, B

(a) Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

S17T, B

S9

S8

S13

D3

Displacement gauge locations

(b)

Strain gauge locations

Sx Sx T, B Dx

Strain gauge at member “x” Strain gauges are attached both top and bottom of girder at member “x” Displacement gauge location number “x”

Figure 15. Locations of the strain gauges and displacement gauges: (a) main truss girder, (b) horizontal bridge deck.

To obtain the different type of load combinations, which are critical to the bridge, the two test-engines were placed together as well as moved under different speeds. The considered three static load combinations are defined as static load case (SLC) 1,2 and 3 by considering criteria of maximum shear effect, maximum bending effect (maximum deflection) and maximum torsion effect to the bridge deck respectively. The loading positions corresponding to the mentioned three load cases are shown in Figure 16. The criteria, which were considered for dynamic load combinations, basically illustrate that impact effect to the bridge with

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

201

Side view: 37.6 Tons per each arrowhead Plane view: 9.4 Tons per each marked point

(a)

Side view: 37.6 Tons per each arrowhead Plane view: 9.4 Tons per each marked point

(b)

Side view: 18.8 Tons per each arrowhead Plane view: 9.4 Tons per each marked point

(c)

40

20

S19,T,B

60

60

S7 S5 S3 S1

40

Primary Stress(MPa)

S15,T,B S14 S13

Primary Axial Stress (MPa)

Primary Axial Stress (MPa)

Figure 16. Loading positions corresponding to three static load cases (a) SLC 1 (b) SLC 2 (c) SLC 3.

20

40

20

0

0 0

1

2

3 4 Time (sec)

5

0

1

2 3 Time (sec)

(a)

4

0

5

2

4 Time (sec)

(b) S16,T,B Vertical Displacement (mm)

0

40

20

6

(c) Vertical Acceleration ( mm/sec2 )

60 Primary Stress (MPa)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

0

-10

200

100

0

-100

-20

-200

0 0

2

4 Time (sec)

(d)

6

0

1

2

3 Time (sec)

(e)

4

5

0

1

2

3 Time (sec)

4

5

(f)

Figure 17. Field measurements of the bridge due to heaviest load: (a) stresses at bottom chord of the main girder, (b) stresses at diagonal members which are usually subjected to tensile stress, (c) stresses at stringers, (d) stresses at secondary cross girders, (e) vertical displacement at midspan, (f) vertical acceleration at midspan.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

202

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

30

20

10

0

80

80 Miximum stress (MPa)

Maximum stress (MPa)

Mid span displacment (mm)

40

60

40

10

20

30 40 50 Train speed (km/h)

0

40

20

20

0

60

0

10

20

(a)

30 40 50 Train speed (km/h)

0

0

10

20

(b)

30 40 50 Train speed (km/h)

(c)

Figure 18. Dynamic factor determination curves (maximum response variation with speed): (a) main truss girder, (b) secondary cross girders, (c) stringers.

different levels of speed and traction force effect. Apart from the above mentioned formal field load testing, the bridge was subjected to a two days continuous field measurement program under present day actual traffic. When the bridge is affected by maximum load due to the present day heaviest train passage, the obtained sample measurements are shown in Figure 17. Finally the dynamic factors were obtained as 1.3, 1.4 and 1.4 for main truss girders, secondary cross girders and stringers respectively by using the curves illustrated in Figure 18.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

7.4. Development of Validated Analytical Model The Bridge deck was analyzed using the finite element (FE) method employing the generalpurpose package SAP 2000 (Ranaweera et al., 2000). A three-dimensional (3D) model (Figure 19) of one complete middle span of the bridge was analyzed under test loading and actual loading to determine stresses in members and deflections, as well as variations of stresses under moving loads. The bridge deck was modeled with 3D frame elements and the riveted connections are assumed to be fully-fixed (Imam et al., 2005). Even though the cross girders are ideally supported on the bottom chord of the main truss girder, the assumption of rotational stiffness behavior with magnitude 18200 kNm/rad (about the local second axis) for representative connection of cross girder to truss were found to be in better agreement with field measurements than the pinned connection assumption.

(a)

(b)

(c)

Figure 19. 3D frame element model for single span: (a) deflected shape for SLC 2, (b) axial force diagram at SLC 2, (c) bending moment diagram at SLC 2.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

203

Table 2. Comparison of FE analytical results with load test results Static load

Displacement (mm) Location of measurement

SLC 1

SLC 2

SLC 3

D1

D1

D1

Stress (MPa)

Load test FEM 19.4

21.3

-

21.0

22.5

19.1

Location of measurement

Load test FEM

S6

-40.2

-40.6

S5

51.4

57.3

S15,T,B

47.3

48.2

S6

-37.8

-37.7

S5

44.5

43.6

S15,T,B

53.5

53.9

S6

-39.5

-39.9

S5

35.2

41.5

S15,T,B

39.0

44.7

The validation of FE model was done by comparing the results from analysis with those from field-tests as shown in Table 2. From the results of static load cases it was seen that there is a good agreement among analytical results of the FE model and the measurement of the actual bridge. Therefore, the considered 3D frame element model was defined as “validated analytical model”.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

7.5. Past and Future Stress Histories Since used types of trains are changing with age of the bridge, the age had to be divided in to several periods. According to the time tables of the bridge, it could be decided that the traffic sequence is almost constant during a single week of each period of age. Then the validated analytical model was used to obtain the elastic stress histories of each critical member during a single week of each period. Therefore a week is considered as a loading block for each period. Due to the dynamic effect of moving trains, the actual working stresses should be higher than the analytical static stress. Therefore, the dynamic factor of each member, which was found experimentally in section 7.3, was used to multiply the static stress to get the service stresses.

8. Remaining Fatigue Life Estimation Based on Primary Stresses Remaining fatigue life evaluations of critical members in each member set are discussed in this section. Evaluations are specially based on primary stresses, which are determined by the global analysis of whole structure. The uniaxial sequential law (section 5.2) with associated curve predicting technique (section 5.3) is employed to obtain a more realistic fatigue life for the bridge.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

204

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

8.1. Determination of Stress Ranges

900

900

800

800

700

700

Number of cycles (per week)

Number of cycles (per week)

The stress histories which have been obtained from the previous section (section 7) are uniaxial. Therefore these are directly considered as primary stresses for the analysis. These primary stress histories are converted into stress ranges as described in section 5.1. The stress range histograms for one critical member have been drawn in Figure 20 for all periods of its age.

600 500 400 300 200 100 0 10

15

20

25

30

35

40

45

50

55

65

70

(a)

100

10

15

20

25

30

35

700 600 500 400 300 200

40

45

50

55

65

70

75

45

50

55

65

70

75

Stress range (MPa)

(b)

800

Number of cycles (per week)

Number of cycles (per week)

200

5

700 600 500 400 300 200 100

100

0

0 5

10

15

20

25

30

35

40

45

50

55

65

70

5

75

Stress range (MPa)

10

15

20

25

30

Number of cycles (per week)

200

150

100

50

0

35

40

Stress range (MPa)

300

(c)

250

Number of cycles (per week)

300

900

800

(d)

250

200

150 100

50 0

0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

0

3 00

(e)

Number of cycles (per week)

2 50

5

10

15

20

25

300

Stress range (MPa) Number of cycles (per week)

400

75

Stress range (MPa) 900

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

500

0

5

300

600

2 00

1 50

1 00

50

30

35

40

45

50

55

60

65

70

75

50

55

60

65

70

75

Stress range (MPa)

(f)

250

200

150 100

50

0

0 0

5

10

15

20

25

30

35

40

45

Stress range (MPa)

(g)

50

55

60

65

70

75

0

5

10

15

20

25

30

35

40

45

Stress range (MPa)

(h)

Figure 20. Sample stress range histograms for critical member of set DT3 for (a) period from 1995 to date, (b) period from 1985 to 1994, (c) period from 1975 to 1984, (d) period from 1970 to 1974, (e) period from 1946 to 1969, (f) period from 1928 to 1945, (g) period from 1910 to 1927, (h) period from 1885 to 1909.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

205

8.2. Determination of Wöhler Curve Field investigations reveal that the riveted connections of the bridge represent lapped or spliced connection behavior with normal clamping force. Therefore, riveted connections were classified as class Wrought-iron (WI), which is proposed by the UK railway assessment code (Network Rail, 2001). Hence the S-N curve, which is mentioned under the UK railway assessment code for WI detail class connection (Network Rail, 2001), was transferred to a fully known Wöhler curve using the method mentioned in section 5.3. The obtained function and the geometrical shape of the new fatigue curve, which correspond to class WI riveted connection, are illustrated in Figure 21. 400

Stress Range (MPa)

300 σ = 20

⎡ N + 1050 ⎤ ⎢⎣ N + 1000000000 ⎥⎦

−0.202

200

100

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

0 1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

1.E+10

Number of Cycles N

Figure 21. Predicted Wöhler curve for wrought iron material.

8.3. Application of Sequential Law The new damage indicator (present Di value) was calculated from the date of bridge construction to the present. Assuming that future sequence of passage is similar to that of the present day, the calculated remaining fatigue life for critical members of each member set is shown in Table 3. In the critical members where the stress range was entirely within the compression zone, the effect of fatigue damage was ignored (BS 5400, part10, 1980).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

206

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

Table 3. Summary of remaining fatigue lives for critical members Bridge component

Reference number

Remaining fatigue life from today (years)

of members

Miner’s rule

Sequential law

Main girder bottom chord

13

305

323

Main girder bottom chord

14

156

165

Main girder bottom chord

15

157

169

Cross girders

16,17

20

12

18,19,20,21

24

13

Truss diagonal (tension member)

1

179

191

Truss diagonal (tension member)

3

168

171

Truss diagonal (tension member)

5

131

138

Truss diagonal (tension member)

7

152

162

Stringers

9. Remaining Fatigue Life Estimation Based on Secondary Stresses Remaining fatigue life evaluation of a critical member of one set is discussed in this section and evaluations are especially based on secondary stresses, which are generated around the riveted connection due to stress concentration effect of primary stress. The considered secondary stresses are usually subjected to multiaxial state of stress.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

9.1. Considered Riveted Connection The considered riveted connection was selected corresponding to a highly stressed member, which was found from the detailed investigation of section 7. The selected connection is shown in Figure 22. Considered highly stressed member is corresponding to member number 3. The condition survey (section 7.1) reveals that there are no visual cracks in any component of the connection. In-situ measurements of member sizes and connections verified the fact that the existing drawings were applicable. Further it can be said that comparatively the maintenance work carried out on this connection thus far is satisfactory. Secondary (local) stressed area

Critical member (a)

(b)

S1

S3

S5

S2

S4

S6

Primary stressed area (c)

Figure 22. (a) One of the critical riveted connections of the main truss girder, (b) close view of the critical connection and the critical member, (c) schematic representation of the critical member and related areas for primary and local stresses.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

207

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

9.2. Secondary Stress Evaluation When a hot rivet is inserted into the hole of plates in order to connect them and when the second head is formed from the protruding shank, the rivet gets shortened in length due to cooling. However most of shrinkage of the free rivets is restricted by the connected plates, which consequently are compressed through the thickness. The residual tensile force in the rivet and the compressive force in the plates get balanced each other; i.e. called as “clamping force”. Therefore the clamping force from the rivet generates a complex tri-axial stress state in the connected plate in the vicinity of the rivet hole (Akesson, 1994). Finally the clamping force seems to be affected by the mechanism of distribution of stresses along the connection. The experience from the field practice reveals that resulting clamping force could vary substantially due to normal conditions. Therefore, it could consequently not be given a reliable value. Furthermore, one can assume a certain relaxation of the rivet clamping force due to creep, fretting of the interfacing plate surfaces, overloading (due to residual plastic deformation), etc. with the time (when bridge is getting older). However, the secondary stress analysis, which corresponds to normal clamping force, becomes more difficult. Because, the geometry of the problem consists of all the rivets with all members and the connected ply. The considered criterion of this section corresponds to the low or high clamping force at the rivets. Therefore, to obtain reasonable accurate results, only critical member without rivets can be considered as relevant geometry for secondary stress analysis in this section. The one of critical member of the truss girder (member 3) was subjected to further analysis of secondary stress evaluation. The nine-node isoperimetric shell elements were used for the FE mesh as shown in Figure 23. To represent the effect of no clamping forces in the rivets, the actual air gap restraint conditions were applied to represent the unilateral contact between rivet and plate. Similarly, to simulate the effect of friction forces due to clamping of the rivet (rotational fixity of riveted connection), the fully bonding unilateral contact behavior between rivet and plate was implemented. The individual deformations of rivets due to loading were not captured in this model. During this analysis, three main different features of rivets were considered such as, i). high clamping force is appears on all the rivets, ii). clamping force disappears from one rivet to the next, and iii). release of contact-ness of rivet while all the rivets have no clamping force. In the third feature, rivets, which are highly contacted with connected ply, transfers the total load properly and these rivets are named as “ active rivets” during this chapter. To make the continuity of stress field between global model (shown in Figure 19) and the sub-model (shown in Figure 23), it is required to use any interface between the two models at every iterative step. In this model, corresponding tensile stress history of the considered member, which has been obtained from section 7.5, is applied on the bottom face (ab of Figure 23 (a)) as a uniform pressure P (Imam et al., 2007; Kiss et al., 2000; Liu et al., 2006). The position of the ab boundary of the sub-model (shown in Figure 23) was determined based on far field primary stress of the member. In house FEM code was employed to perform a nonlinear kinematic hardening based elasto-plastic analysis. Corresponding increment of far field stress histories (P) of the member was imposed for the elasto-plastic incremental analysis. The obtained maximum stress contours are shown in Figure 23 only for third feature (when reducing the active number of rivets) of riveted connection. Since the proposed fatigue life estimation technique (section 5.4) describes the stress field at critical locations in terms of equivalent von Mises stress, the von Mises stress histories at critical location due to daily

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

208

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

passage of trains were obtained for fatigue life evaluation. Sample results are shown in Figure 24. The von Mises stress histories are converted in to stress ranges as described in section 5.1, which is similar to previous case.

a

Highly

Highly

Highly

Highly

Highly

stressed

stressed

stressed

stressed

stressed

location (144 Mpa)

location (172 Mpa)

location (174 Mpa)

location (203 Mpa)

location (228 Mpa)

b

(b)

(c)

(d)

(e)

(f)

P (a)

Figure 23. (a) Fine FE mesh, (b) maximum von Mises stress contour when all six rivets are active, (c) maximum von Mises stress contour when five rivets are active, (d) maximum von Mises stress contour when four rivets are active, (e) maximum von Mises stress contour when three rivets are active, (f) maximum von Mises stress contour when two rivets are active. 250

Operating effective stress (MPa)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

200 P rimary s tre s s hi s tory Stre s s his tory w he n all 6 rive ts are ac ti ve Stre s s his tory w he n 5 rive ts are ac tive

150

Stre s s his tory w he n 4 rive ts are ac tive Stre s s his tory w he n 3 rive ts are ac tive Stre s s his tory w he n 2 rive ts are ac tive

100

50

0 0

1

2

3

4

5

6

-50

T im e (s ec)

Figure 24. The von Mises stress histories at critical locations due to heaviest rail traffic for each of the considered features of riveted connection.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

209

9.3. Determination of Wöhler Curve The appropriate fatigue curve for this evaluation, the mean S-N curve of class B, which represents the classification of no detail (class of plain section), was obtained from the UK railway assessment code for Wrought iron material ((Network Rail, 2001)). Further, it describes only the fatigue behavior of wrought iron material and the obtained fatigue lives from this curve show 50 % probability of failure. The chosen fatigue curve only describes stress ranges, which correspond to more than ten thousands of failure cycles. Therefore the chosen partially known Wöhler curve was transferred to fully known Wöhler curve by using the method mentioned in section 5.3. The obtained function and the geometrical shape of the new fatigue curve are illustrated in Figure 25.

600

Stress range (MPa)

500

σ

400

⎡ N + 300 ⎤ = 60 ⎢⎣ N + 120000000 ⎥⎦

−0.1674

300

200

100

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

0 1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

1.E+10

Number of cycles N

Figure 25. The mean Wöhler curve for wrought iron material (transferred to full range using method described in section 5.3).

9.4. Fatigue Life Evaluation The proposed extension of sequential law in multiaxial fatigue (section 5.4) was utilized in this section to obtain a more realistic service life for the bridge. Although the hypothesis behind the fatigue model is the same as the previous case, only difference is that the active stress field is considered as the equivalent von Mises stress for multiaxial state of stress. Therefore, it is possible to use the previous fatigue concept with algorithms, which has been described under section 5.2, for fatigue life estimation by replacing active stress by equivalent von Mises stress. The considered member (member 3), which is the one of most critical member of main girder, was subjected to fatigue evaluation by considering three major steps. These steps describe the variation of remaining fatigue life with condition of riveted connection, such as level of clamping force effect, degree of surface contact-ness of rivets, etc.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

210

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

Step 1: In the first step, effect of clamping force on all the rivets was considered at once and remaining life was evaluated. When a rivet is subjected to plastic loading, the effect of clamping force begins to deviate from initial value. However, the considered connection is operating in an elastic state of stress either it has a high clamping force effect (The maximum von Mises stress value is 102 MPa) or low clamping forces (The maximum von Mises stress value is 144 MPa). Therefore it is possible to assume that the clamping force of the considered connection is not significantly changing with the time. As a result, fatigue life was obtained by considering that the deemed feature of clamping force remains the same from bridge construction to the date of failure. The calculated results are shown in Table 4.

Table 4. Comparison of both primary and secondary stress based fatigue lives Remaining fatigue life (years)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Primary stress approach Reference number of the member

Previous method

Proposed method

3

168

171

Secondary stress approach High clamping force (upper bound)

Low clamping force (lower bound)

Previous method

Proposed method

Previous method

Proposed method

Infinite

Infinite

Failed 17 years before

Failed 4 years before

Step 2: In this step, the fatigue life is obtained considering that the effect of clamping force disappears from one rivet to the next. The calculated results are shown in Table 5. Maximum stresses for first five features of the Table 5 (until five rivets have low clamping force, remaining one has high clamping force) are operating bellow the yield limits. (The maximum von Mises stress values are 109.3, 110.3, 111.4, 117.2, 122.1 & 144 MPa respectively from top feature of the Table 5). Therefore, here also it is possible to assume that there is no significant change of clamping force with the time and it is remained constant (at high clamping force rivets) or disappeared (at low clamping force rivets) in the elasto-plastic analysis of this step. The defined fatigue life in this section describes the time duration from the date when considered feature of riveted connection appeared, to the date of fatigue failure. Further, it was considered that the sequence and density of rail passage were similar to present period of operation.

Table 5. Fatigue life variation when the effect of clamping force is gradually decreasing Fatigue life (years) Considered features of the riveted connection One rivet has low clamping force, remaining five have high clamping force Two rivets have low clamping force, remaining four have high clamping force Three rivets have low clamping force, remaining three have high clamping force Four rivets have low clamping force, remaining two have high clamping force Five rivets have low clamping force, remaining one has high clamping force All six rivets have low clamping force

Previous Proposed method method Infinite Infinite Infinite Infinite Infinite 20

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Infinite Infinite Infinite Infinite Infinite 23

Remaining Fatigue Life Estimation of Existing Railway Bridges

211

Table 6. Fatigue life variation when active numbers of rivets are gradually decreasing while low clamping force Considered features of the riveted connection All six rivets are active Five rivets are active Four rivets are active Three rivets are active Two rivets are active

Fatigue life from today (months) Previous method

Proposed method

245 63 59 19.7 6.5

272 70 65 22.5 7.4

Step 3: In the third step the fatigue damage is evaluated based on a criterion called critical state of stress due to release of contact-ness of rivet while all the rivets have low clamping force. When the rivet is not properly in contact with the plate, particular rivets tend to transfer less or zero amount of total load, and this leads to unexpected stress redistribution around the riveted area. In this stage other rivets, which are carrying the load, are called as active rivets as described before. The fatigue lives were evaluated stepwise reducing of contribution of active number of rivets in the connection. The calculated lives are shown in Table 6 and the defined fatigue life in this section describes the time duration from date when considered feature of riveted connection appeared, to the date of fatigue failure. Further, it was considered that the sequence and density of rail passage was similar to present period of operation as similar to previous step.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

10. Comparisions and Discussion The obtained primary stresses based remaining fatigue lives (section 8) for the critical members were compared with the Miner’s rule-based previous method estimation (section 2) as shown in Table 3. In the case of main girder truss members, the sequential law-based remaining fatigue life illustrates higher values than Miner’s rule-based values. But it is the opposite for bridge deck members. Since the Miner’s rule estimation produces pessimistic results with increasing of loads and optimistic results with decreasing loads, it can be said that in case of truss members, the global increment of live load of trains with each period of age has greater effect on fatigue damage than local variation (increase and decrease of loading during a week). Similarly it can be seen that in case of bridge deck members (cross girders, stringers and bracings) the local variation of loading has a greater effect on fatigue damage than global increment of loading. However, it is revealed that the effect of members where the lives are very low (e.g., cross girders, stringers) becomes more significant in percentage terms. The secondary stresses based remaining fatigue lives (section 9) for one of critical member is also compared with the previous method prediction (section 2) as shown in Table 4, 5 & 6. The Table 4 shows that primary stresses based fatigue lives (in section 8) lie in between upper and lower bound of secondary stresses based estimation (in section 9). Therefore, it is possible to confirm that the UK railway code provided S-N curve, (Network

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

212

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

rail, 2001), represents the normal or intermediate effect of clamping force for wrought iron riveted connections of existing bridges. The results in Table 5 illustrated that the effect of clamping force in riveted connection tends to deviate the fatigue life considerably. Likewise, the lives, which are shown in Table 6, reveal that proposed method based-results deviate from previous method estimations. Although these types of deviations are particular to this bridge, it can be generally said that loading sequence effect critically influences the estimation of remaining fatigue life. Further, the obtained fatigue lives of all of the methods reveal that the active number of rivets, which are able to transfer the load, also changes the fatigue life significantly. Generally, the activeness of rivets may change with the time in service condition of a bridge and cause unexpected stress concentration near the riveted area.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

11. Conclusion This chapter comprehensively described the remaining fatigue life estimation methods for existing railway bridges. Initially, the available fatigue life estimation method was clearly indicated. The deficiencies of the presently accepted method have been described. Then a new method was proposed to improve the accuracy of predicted fatigue life. The verification of the proposed method was conducted by comparing the predicted results with experimental results. Theoretical fatigue damage behaviours were compared with real damage behaviour of two materials to verify the sequential law and associated S-N curve modeling technique (section 6.1). There, it shows that the sequential law method illustrates a closer relation to real damage behavior than previous fatigue theory, Miner’s rule. The verification of proposed extension of multiaxial fatigue (section 6.2) shows that there is a good agreement between sequential-law–predicted multiaxial fatigue life and the real fatigue life of some materials, more than Miner’s prediction. It tends to conclude that proposed extension of sequential law of multiaxial fatigue and associated fully-known Wöhler curve modeling technique produces a reasonably accurate prediction of multiaxial fatigue. Then, all of those methods were applied to estimate the remaining fatigue life of a riveted railway bridge in Sri Lanka. The comparisons of the results were also made. Finally, the chapter highlights major conclusions as follows. Condition evaluation of the bridge exhibits that the overall maintenance of the considered bridge is satisfactory, but there are localized mild corrosion at few places, and these need immediate attention. Due to fatigue, under current loadings, speeds and frequencies of operation, the lowest remaining life found for a member is 12 years. Thus it may be concluded that the bridge deck can be used for another 12 years provided that the speed, frequency, and weight of the trains are not increased. If proper maintenance work is carried out and the critical members are replaced with new members with longer life, the bridge will be able to provide further service. The 10 to 15 years’ variation of fatigue lives has been highlighted between previous and proposed methods. Obviously, the effect of parts for which the lives are very brief (e.g., cross girders, stringers) becomes more significant in percentage terms. These observations and the phenomenological validity of the new damage indicator-based sequential law tend to conclude that the application of sequential law is greatly advisable for the evaluation of remaining fatigue life of riveted railway bridges.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Remaining Fatigue Life Estimation of Existing Railway Bridges

213

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The remaining life is 160–170 years for member at location 3, when all six rivets have normal clamping force and failed before 10–20 years without clamping force at all six rivets. Further, it increases to infinity when it has high clamping force (Table 4). Therefore, it is possible to give an assurance to some extent that the UK railway- code–provided design S-N curve (Figure 21) captures the normal or intermediate effect of clamping force at riveted connections. As a result, the obtained function and geometrical shape of this fully known design S-N curve can be employed to assess the fatigue damage of other wrought iron bridges, which have riveted connections with normal clamping force. The fatigue life evaluation based on secondary stress analysis revealed that the clamping force and the activeness of rivets play a big role in fatigue damage. Hence, it can be concluded that it is greatly important to investigate accurately the condition of places where the stress concentration effect is severe such as notch, crack or connection area especially in old bridges for good judgment of fatigue life. The proposed past stress evaluation procedure is simple and applicable to a wide range of railway bridges. Further, it is realized that sequential law is dependent only upon the Wöhler curve, and current code-provided curves could be used in this method similarly to the previous method. The comparisons of remaining fatigue life highlight the necessity of a proper fatigue theory, which describes the loading sequence effect precisely. Finally, these observation tend to conclude that the application of the sequential law-based proposed method is advisable for the evaluation of remaining fatigue life of riveted railway bridges for which the detailed past stress histories are not known. Further, observations emphasize that the primary stress-based approach is more advisable for general use and secondary stressbased approach has been recommended for detailed studies. Since this investigation has not captured the effect of various types of microstructural changes and the effect of mesoscopic damage variables of particular materials at highly stressed locations, comparisons of the above approach with mesoscopic-level fatigue theories are currently on the way.

Acknowledgments The authors wish to express their sincere gratitude to Senior Professor M.P. Ranaweera and the team of experts who worked in the Sri Lankan Railway Bridge project for their great advice in carrying out this research. The kind support given by the Sri Lanka Railways (SLR) is also appreciated. The authors also convey their gratitude to Professor E. Macha and Professor Z. Marciniak for kindly providing us with the necessary loading histories of the multiaxial fatigue test free of charge.

References Agerskov, H.; Nielsen, J.A. Fatigue in steel highway bridges under random loading. Journal of Structural Engineering 1999, 125, 152-162. Akesson, B. Fatigue Life of Riveted Railway Bridges; PhD Thesis, Chalmers University of Technology, Göterborg, 1994.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

214

Siriwardane Sudath Chaminda, Mitao Ohga, Ranjith Dissanayake et al.

Alampalli, S.; Lund, R. Estimating fatigue life of bridge components using measured strains. Journal of Bridge Engineering.2006, 11, 725-736. BS 5400, Part 10. Code of practice for fatigue, Steel Concrete and Composite Bridges. British Standard Institutions: London, UK. Dattoma, V.; Giancane, S.; Nobile, R.; Panella F.W. Fatigue life prediction under variable loading based on a new non-linear continuum damage mechanics model. International Journal of Fatigue. 2006,28, 89-95. Fisher, J.W.; Yen, B.T.; Wang, D. Fatigue and Fracture Evaluation for Rating Riveted Bridges; NCHRP Report No 302;Transportation Research Board, National Research Council: Washington D.C, 1984. Imam, B.; Righiniotis, T.D.; Chryssanthopoulos, M.K. Fatigue assessment of riveted railway bridges. International Journal of Steel Structures. 2005, 5 , 485-494. Imam, B.M.; Righiniotis, T.D.; Chryssanthopoulos, M.K. Numerical modeling of riveted railway bridge connections for fatigue evaluation. Engineering Structures. 2007,29, 3071-3081. Jung, S. K.; Frangapol, D.M. Prediction of reliability and cost profiles of deteriorating bridges under time-and performance-controlled maintenance. Journal of Structural Engineering. 2004,130, 1865-1874. Karolczuk, A.; Macha, E.; Fatigue fracture planes and expected principle stress directions under variable amplitude loading. Fatigue & Fracture of Engineering Materials & Structures. 2005, 28, 99-106. Kiss, K.; Dunai, L. Stress history generation for truss bridges using multi-level models. Computers and Structures. 2000, 78, 329-339. Kluger, K.; Lagoda, T.; Application of the Dang-Van criterion for life determination under uniaxial random tension-compression with different mean values. Fatigue & Fracture of Engineering Materials & Structures. 2005, 27, 505-512. Köröndi, L.; Szittner, A.; Kálló, M.; Krisróf, L. Determination of fatigue safety and remaining fatigue life on a riveted railway bridge by measurement. Journal of Constructional Steel Research. Paper number 327, 1998, 46, 430. Kohout, J.; Vechet, S. A new Function for Fatigue Curves Characterization and Its Multiple Merits. International Journal of Fatigue. 2001, 23, 175-183. Li, Z.X.; Chan, T.H.T.; Ko, J.M. Fatigue analysis and life prediction of bridges with structural health monitoring data- Part 1: methodology and strategy. International Journal of Fatigue. 2003, 23, 45-53. Li, Z.X.; Chan, T.H.T.; Ko, J.M. Fatigue analysis and life prediction of bridges with structural health monitoring data- Part II: application. International Journal of Fatigue. 2001, 23, 55-64. Li, Z.X.; Chan, T.H.T.; Ko, J.M. Fatigue damage model for bridge under traffic loading: application made to Tsing Ma Bridge. Theoretical and Applied Fracture Mechanics. 2001, 35, 81-91. Li, Z.X.; Chan, T.H.T.; Ko, J.M. Determination of effective stress range and its application on fatigue stress assessment of existing bridges. International Journal of Solid and Structures. 2002, 39, 2401-2417. Li, Z.X.; Chan, T.H.T.; Zheng, R. Statistical analysis of online strain response and its application in fatigue assessment of a long span steel bridge. Engineering Structures. 2003, 25, 1731-1741.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Remaining Fatigue Life Estimation of Existing Railway Bridges

215

Li, Z.X.; Chan, T.H.T. Fatigue criteria for integrity assessment of long span steel bridge health monitoring. Theoretical and Applied Fracture Mechanics. 2006, 46, 114-127. Liu, Y.; Stratman, B.; Mahadevan, S. Fatigue crack initiation life prediction of railroad wheels. International Journal of Fatigue. 2006, 28, 747-756. Marciniak, Z.; Rozumek, D.; Macha, E. Fatigue lives of 18G2A and 10HNAP steels under variable amplitude and random non-proportional bending with torsion loading. International Journal of Fatigue. 2008, 30, 800-813. Mesmacque, G.; Garcia, S.; Amrouche, A.; Rubio-Gonzalez, C. Sequential law in multiaxial fatigue, a new damage indicator. International Journal of Fatigue. 2005, 27, 461-467. Miner, M.A.Cumulative Damage in Fatigue. Journal of Applied Mechanics. 1945, 12, 159164. Network Rail. Rail track Line Code of Practice, The Structural Assessment of Underbridges, RT/CE/C/025. Rail track, 2001. Ranaweera, M.P.R.; Aberuwan, H.; Mauroof, A.L.M.; Herath, K.R.B.; Dissanayake, P.B.R; Siriwardane, S.A.S.C.;, Adasooriya, A.M.N.D. Structural appraisal of railway bridge at colombo over kelani river; Engineering Design Center, University of Peradeniya, Sri Lanka, 2002. Roeder, C.W.; MacRae, G.; Leland, A.; Rospo, A. Extending the fatigue life of riveted coped stringer connections. Journal of Bridge Engineering. 2006, 10, 69-76. Shang, D.G.; Yao, W.X. A nonlinear cumulative model for uniaxial fatigue. International Journal of Fatigue. 1999, 21, 187-194. Sherif, B.; Shuichi, M.; Toshiyuki, O. Nondestructive damage detection scheme for steel bridges. Journal of Applied Mechanics. 2006, 19, 63-74. Siriwardane, S.; Ohga, M.; Dissanayake, R.; Taniwaki, K. Application of new damage indicator-based sequential law for remaining fatigue life estimation of railway bridges. Journal of Constructional Steel Research. 2008, 64, 228-237. Suresh, S. Fatigue of materials; Second edition, Cambridge University Press, UK. 1998. Zheng, X.L.; Lu, B.; Jiang, H. Determination of probability distribution of fatigue strength and expressions of P-S-N curves. Engineering Fracture Mechanics. 1995, 50, 483-491. Zheng, X.; Wei, J. On the prediction of P-S-N curves of 45 steel notched element and probability distribution of fatigue life under variable amplitude loading from tensile properties. International Journal of Fatigue. 2005, 27, 601-609. Zhou, Y.D.Assessment of bridge remaining fatigue life though field strain measurement. Journal of Bridge Engineering. 2006, 11, 737-744.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

In: Railway Transportation Editor: Nicholas P. Scott

ISBN: 978-1-60692-863-9 © 2009 Nova Science Publishers, Inc.

Chapter 7

MAINTENANCE MANAGEMENT STRATEGIES FOR STEEL RAILWAY BRIDGES: A CASE STUDY FROM VIETNAM Dinh Tuan Hai* Faculty of Urban Management, Hanoi Architectural University Km 10 Nguyen Trai Road, Hanoi City, Vietnam

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Abstract Bridge defects have become a major social concern in recent times. Defects on railway bridges in Vietnam have been identified as structural failure, corrosion, fatigue, functional obsolescence, aging and human intrusion. In addition, problems of low maintenance budget, rigidity and long lines of communication between relevant stakeholders, inadequate management attention and lack of prioritization in maintenance selection adversely impacts bridge structures. This study proposes a series of maintenance strategies for overcoming the current problems. Essential strategies are suggested for the short-term to delay the onset of deterioration by extending service life. Preventive strategies aim to eliminate potential causes early, thus the required resources and time expanded on remedies of future defects can be substantially reduced. It is recommended that preventive strategies should be given priority to eliminate causes of potential problems. Moreover, essential maintenance must be regularly carried out to eliminate serious defects. These goals can be achieved only if training of involved parties and their in-house personnel becomes imperative to identify failure modes and their root causes, and to decide on appropriate preventive and corrective solutions. Further research is needed to verify the proposed maintenance strategies on site and to modify them to fit the specific condition of Vietnam.

CE Database subject headings: Bridges; Defects; Maintenance; Railway; Steel; Vietnam

*

E-mail address: [email protected] Tel: (84) 4 – 3854 2073; Fax: (84) 4 – 3854 2073. Corresponding author: Dinh Tuan Hai, Faculty of Urban Management, Hanoi Architectural University, Km 10 Nguyen Trai Road, Hanoi City, Vietnam

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

218

Dinh Tuan Hai

Introduction The number of deteriorated bridges is on the increase in many countries, due to the adverse impacts of heavy traffic, climate, premature aging, and so on. Railway bridges have shown many evidences of structural defects and functional deficiencies. However, owing to budget limitations, maintaining them is now becoming significantly difficult. Therefore, the development of a broad array of comprehensive maintenance strategies for existing bridges is essential. Such strategies should be able to detect and evaluate current bridge problems. In addition, there is need for optimizing short and long-term maintenance plans within the limitation of available resources. Vietnam has approximately 2,600 route-kms of railway network for several single-track lines. The network is composed of a total 1,777 bridges with approximate length of 44 kms that is mostly built from steel structures (TMoVN, 2004). Currently, there are many bridges in poor physical condition and insufficient serviceability, causing several problems for trail operation and negatively impacting into development of the country. Moreover, not much effort is given to the analysis of current bridge defects and to propose appropriate maintenance actions, making it more difficult to overcome these problems (Hai, 2006). This paper therefore presents a brief overview of current physical condition of railway bridges in Vietnam. In addition, comprehensive maintenance strategies, essential and preventive, are proposed to eliminate the problems and their root causes. However, the scope of this study is limited to steel structures only, since the majority of railway bridges in Vietnam was constructed from either steel or steel-concrete combination.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

A. Brief History of Railway Bridges in Vietnam The railway system in Vietnam comprises several single-track lines as illustrated in Fig. 1 and Table 1. Construction of first bridges started in 1881 concurrently with the 1,726 km long Tran-Viet rail-track from Hanoi to Saigon. Subsequently, more bridges have been constructed to complete the Tran-Viet line in 1936 and other branches were added later (TMoVN, 2004). The system experienced long periods of construction and commissioning; through different ownerships. Railway bridges in Vietnam comply with different design standards, have a wide range of shapes and commissioning dates and have been subject to wars, adverse climate, and inadequate maintenance (TMoVN, 2002). Some historical landmarks of railway bridges in Vietnam are illustrated as follows. •

•

•

The period 1881-1930: One-span simple steel bridges were all manufactured in France. French standards were applied for bridge design and construction to allow bridges to accommodate 8-10 ton/axis trains at speeds of less than 25 km/h. The period 1930 – 1954: Large bridges used steel trusses from Germany and France to accommodate trains having imposed loads of 10 ton/axis and speeds up to 40 km/h. French standards were also applied for bridge maintenance and management. The period 1954 – 1975: The country was divided due to wars and bridges mostly served military purposes. U.S. and Soviet standards were applied separately in both sides of Vietnam that were actually two countries under a state of war.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Maintenance Management Strategies for Steel Railway Bridges

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

•

219

The period 1975 – present day: High demand of train transportation developed with the country’s unification in 1975 and economic boom since the 1990s has increased the number of bridges significantly. Bridge-related standards developed locally are now identically applied for the whole country in order to accommodate modern-style trains.

Figure 1. Existing railway network in Vietnam.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

220

Dinh Tuan Hai

Table 1. Existing railway lines in Vietnam. Existing railway lines of Vietnam

Distance

Track gauge (mm)

(km)

%

1,726

66

1,000

Cau Giat – Nghia Dan

30

1

1,000

Dieu Tri – Quy Nhon

10

Hanoi – Dong Dang

166

6

1,000 and 1,435

Yen Trach – Na Duong

31

1

1,000

Yen Vien – Lao Cai

285

11

1,000

Hanoi – Saigon

1,000

Tien Kien – Lam Thao

3

1,000

Pho Lu – Pom Han

24

1

1,000

Gia Lam – Hai Phong

91

3

1,000

Dong Anh – Quan Trieu

54

2

1,000 and 1,435

Kep – Ha Long

105

4

1,000

Chi Linh – Co Thanh

16

1

1,000

Luu Xa – Kep

56

2

1,000

Van Dien – Bac Hong

41

2

1,000

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Current Status of Railway Bridges in Vietnam General Physical and Serviceable Conditions TMoVN (2004) classifies railway bridges in Vietnam in terms of their length, materials used, construction time and load-carrying capacity (Fig. 2). In general, existing bridges have a wide range of shapes, commissioning dates and have been built using various design standards. Some have been subject to the impact of wars, adverse climate and inadequate maintenance (TMoVN, 2002). Many temporary steel bridges built to serve military purposes are still inservice now. They, of course, do not satisfy modern civilian trains and need to be replaced as soon as possible. There are differences in geographical locations ranging through mountains, delta and coasts. The climate in the northern area has four seasons (spring, summer, autumn and winter) whereas the southern area only has rainy and dry seasons. Moreover, a dramatic increase in train frequency, speed and imposed load has an adverse impact on bridge conditions (Tuong, 2002). It can be assumed that existing bridges categorized as “old” and “very old” do not satisfy the modern trains requiring at least 14 ton/axis at speed of 60-80 km/h in Vietnam. This is because the old bridges were either seriously damaged in the wartime or built in accordance with old French standards for loading capacity of 6-10 ton/axis at maximum allowable speed of 40km/h (Hai, 2006).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Maintenance Management Strategies for Steel Railway Bridges

221

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 2. Classification of existing railway bridges in Vietnam.

A previous research revealed that much damage has been appearing on railway bridges in Vietnam, though it varies widely according to location, climate, design standards, materials used, commissioning dates, etc (Hai, 2006). Several common types of deficiencies of steel structures, sorted in order of seriousness and frequency, are identified for structural failure, corrosion, fatigue damage, non-functioning, aging, and human intrusion (Table 2). Current failure modes are sorted into five groups classified in accordance with the frequency of damage as “very rare”, “rare”, “sometimes”, “frequent”, and “very frequent” respectively. On the other hand, the degree of seriousness of the detected defects is divided into “very light”, “light”, “moderate”, “serious”, and “very serious”. While “very serious” indicates a steel structure in critical condition, “serious” means even when a steel structure is temporarily stable it may become critical under extraordinary incidences.

Table 2. Current problems on steel structures of railway bridges in Vietnam and their main causes. Problems

Structural failure

Details Structural collapses.

Occurrences Very rare

Seriousness Very serious

Steel deformation.

Sometimes

Serious

Tearing of steel plates. Frequent

Moderate

Main root causes Attacks of missiles, bombs and explosions throughout elapsed wars. Impacts of high-force collisions. Other causes as overloads and overspeeds, construction mistakes and missing elements.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

222

Dinh Tuan Hai

Table 2. Continued Problems

Steel corrosion

Fatigue damage

Details

Occurrences

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Main root causes

Upper-structures Very frequent (Above bridge decks).

Light

Superstructures (Underneath of decks). Substructures.

Frequent

Serious

Not considered as they are almost concrete and stones.

Poor protective coats for metal surfaces.

Cracks of welded lines and weaken locations. Over-sags of bridge decks and girders.

Frequent

Serious

Frequent

Moderate

Traffic overloads and overuses (increased weights and speeds) as well as structural vibrations of trains impose on bridges. Exceeded collisions of large moving trains and vessels against bridges. Other causes of explosions missiles, and bombs due to the wars, external impacts, natural disasters and so on.

Over-displacement of Sometimes steel elements. Joint breaking and Sometimes damage of steel joints. Low load capacities of Frequent steel decks and girders. Limited speed Very frequent allowance of modern trains. Nonfunctioning Approach railroads do Rare not align with bridges.

Aging

Seriousness

Moderate Moderate

Serious

Moderate

Moderate

Low space clearances Sometimes of bridge decks and girders.

Serious

Aging bridges with Frequent ages 50 years upward. Signals of shipworms, Very frequent mosses and funguses.

Very serious

Material cancers of decays and erosion occur on protective coats.

Moderate

Frequent

Light

Adverse climate of high moisture, salt air, dramatically change in temperatures, etc. Low maintenance budget and lack of preventive maintenance.

Enormous changes of future conditions against initial design assumptions. Differences between bridges and railroads for speed and load capacities. Lacking of proper bridge management systems to store and update bridge data. Other causes of climate impacts, material degradations and aging, fatigues, etc. Adverse climates with high humidity, salt airs, heavy rains, etc. Lack of maintenance to remove aliens and to eliminate local defects off bridge surfaces. Use unsuitable and non-durable protective coats against weather attacks. There are many old bridges in services.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Maintenance Management Strategies for Steel Railway Bridges

223

Table 2. Continued Problems

Human intrusion

Details

Occurrences

Seriousness

Invading bridge decks Sometimes and pedestrian lanes.

Serious

Invading clearance spaces of bridges.

Moderate

Frequent

Invading substructures Rare of piers and abutments.

Moderate

Main root causes Bridges locate at ideal locations where there are many vehicles and people using them. Lack of strict protection and imposed actions against alien intruders. The history creates many bridges without clear boundaries against surrounding areas.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Bridge Administration The Vietnam Railway Corporation (VRC) is at present an independent state-owned enterprise under the control of the Ministry of Transport. In the year 1995, it entered into a business agreement with the Vietnamese government in which the government is responsible for train infrastructure. The VRC has all rights to use existing infrastructures, but should pay 10% of its operating revenue to the government. In term of organizational structure, the VRC is divided into three unions corresponding to geographic regions (Fig. 3). Bridge maintenance management has been administrated by 15 public-work government subsidiaries distributed across the unions. They carry out daily site maintenance and office management for all railway bridges in Vietnam. New bridge construction is under the management of three project management divisions with their staff appointed by the VRC.

Figure 3. Administrative structure for existing railway bridges in Vietnam.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

224

Dinh Tuan Hai

Bridge Maintenance Practice Rail tracks, including bridges are maintained mainly by 15 maintenance management enterprises, which are state-owned companies controlled by the VRC (TMoVN, 2004). Each enterprise is allocated a railway section of about 100-300 kms. Bridge maintenance is done manually using hand tools, beaters and hand tie tampers. Mechanized maintenance is not yet available, but it is planned for the future. Site works are carried out to ensure safe and smooth commissioning of railway bridges in Vietnam, to maintain the structures at reasonable quality and to reduce external impacts of overloads, climate, human intrusion, etc (Hai, 2006). The government wishes to preserve structural safety and traffic capacity and reduce the lifecycle cost of the existing bridges by extending their life. On the other hand, bridge preservation is not executed properly in the field due to lack of finance and attention to proper maintenance (NAoSRoVN, 2002). As per the current regulation in Vietnam, site maintenance works are classified into three categories, thus (Fig. 4). •

•

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

•

Regular maintenance: carried out every month by in-house personnel of the maintenance management enterprises under prefixed specification. It will be continued with predetermined schedule and budget. Periodic maintenance: varies from 2 to 5 years with time cycle, budget, scope, etc., depending on the actual physical and serviceable conditions of the bridges. Special maintenance: uniquely depends on specific condition of the bridges. They become compulsory when bridges are either physically unsafe or not serving traffic normally. This maintenance is outsourced to external contractors and supervised by the project management units (Fig. 5).

Figure 4. Categories of site maintenance for steel railway bridges in Vietnam.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Maintenance Management Strategies for Steel Railway Bridges

225

Figure 5. Procedure of special maintenance for steel railway bridges in Vietnam.

Bridge Management Practice The 15 maintenance management enterprises responsible for regular maintenance of the existing bridges also have full responsibility for their daily management. Their duties are (a) management of bridge inventory data, (b) inspection and evaluation of bridge health condition, and (c) security of traffic safety and operation. Therefore, these enterprises manage bridges by focusing on the following three aspects. •

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

•

•

Inventory management: Untrained staff have handled all bridge related data. The data is not properly compiled and classified since there is no computerized database system in Vietnam. Bridge site management: Existing bridges are frequently damaged by external agents such as human beings, animals and vegetations. However, some important bridges are fully guarded by the staff of the maintenance management companies. Inspection: Bridges are inspected regularly and periodically. Furthermore, they are under special evaluation every 3 to 5 years.

Major Problems on Steel Railway Bridges in Vietnam The previous overviews unveil five major problems currently experienced on railway bridges in Vietnam. These problems are further confirmed when the author conducted a series of comprehensive personal interviews with bridge stakeholders (e.g. field engineers, in-house staffs, maintenance workers, and local researchers). The problems have been summarized into both physical and maintenance management aspects as below. •

•

Problem 1: Railway bridges have poor physical and serviceable conditions. The common failure modes of steel structures are identified as structural failure, steel corrosion, fatigue damage, functional obsolescence, aging, and human invasion. Problem 2: Site maintenance currently gives equal-consideration to all aspects and work through random-selection. Priority maintenance criterion that enables the

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

226

Dinh Tuan Hai

•

•

•

assessment of maintenance demands and optimizing the available budget has yet to be applied. Problem 3: The allocated budget is usually small and covers only 30-50% of actual maintenance needs. Thus, site maintenance is limited to several bridges, while many failed bridges received little or no remedial treatment throughout their life. Problem 4: There is much distance between the bridge owner (the government) and the maintenance agencies (maintenance management enterprises). This creates bureaucracy, communication slowdown, cost escalation, and so on. Problem 5: Good management system encompassing computerized database has yet to be applied even through it is considered to be the most important factor for effectively managing and maintaining the existing bridges

Maintenance Strategies for Steel Railway Bridges in Vietnam Major problems of the existing bridges identified above should be eliminated early in order to prevent adverse impacts on safe and smooth operation of the railway network in Vietnam. The diversity of existing problems demands not just a single solution, but a series of maintenance strategies to remove the symptoms and to eliminate the root causes. Essential and preventive maintenance need to be differentiated in definition and time implementation. If done correctly, this will enhance the serviceability of the existing bridges and can significantly decrease severity of the current problems on steel structures of railway bridges in Vietnam.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Essential Maintenance Strategies Essential maintenance is defined as performance-based activities in that it is applied when a performance indicator of an existing bridge reaches a pre-defined target value. It is suggested for a short-term period in order to delay the onset of deterioration and so extend service life. The justification for essential maintenance is that, without it a whole bridge or its specific elements would be unsafe or unable to function normally (Das, 1999). The time of application is implicit and immediate once any performance limit is violated. Several maintenance strategies are, therefore, proposed to eliminate the current major problems mentioned above.

Site Maintenance Site maintenance seeks to delay the progressive development of existing defects. In the end, deterioration of steel structures still occurs but overall lifespan of whole bridges and their elements are significantly increased. Examples of site maintenance are removing rust, repainting on corroded structures, cleaning bridge surfaces against the presence of aggressive agents, and replacing failed members. In accordance with country-specific conditions for steel structures and their current defects, several types of site maintenance are tentatively recommended for railway bridges in Vietnam (Table 3).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Maintenance Management Strategies for Steel Railway Bridges

227

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Table 3. Proposal of specific site maintenance for steel railway bridges in Vietnam. Scope of site maintenance Remove rusts and repaint for corroded steel elements.

Locations Surfaces of steel structures, especially hidden and access-difficult locations.

Cleaning on steel surfaces, removing unexpected aliens. Periodic and special visual inspections.

All bridge structures, especially for hidden and access-difficult locations. All bridge structures, especially for and access-difficult locations.

Replace for weakened and damaged steel elements.

All key components such as steel decks, beams and girders.

Repair for steel defects of cracks, joint breaking, erosions, etc. Select maintenance scope to base on maintenance priority orders.

All steel elements, especially for key components of bridges. Bridge elements in specific and bridges of same networked routes in overall.

Purposes Delay progressive develop of chloride problems on steel structures. Prevent the occurrence of shipworms, mosses and funguses on steel surfaces. Detect new defects or record the progress of existing defects as well as their root causes. Eliminate catastrophic failures and serious damages of steel structures. Prevent existing defects further developing or reaching to ultimate levels. Orient limited maintenance budget into necessary site maintenance.

Priority Maintenance Method The proposed method is the algebraic combination of multiple criteria such as deficiency, specific location, health quality, and obsolescence. The criteria have different weights, and the ensemble of weights reflects the policies of the bridge owner. A priority ranking function is an evaluation of present-day needs for improvement of bridge physical and serviceable conditions. Finances and other resources allocated for the existing bridges are typically inadequate to maintain the safety and functionality of all bridges properly at the same time. The application of the priority ranking enables the conflicting objectives to be solved, ensuring good condition of bridges at lowest possible costs. It channels limited budget into the most necessary site maintenance to generate highest outcome and to promote the improvement of bridge quality and functionality. Currently, railway bridges in Vietnam are in poor conditions and so require large maintenance costs. However, the allocated annual budget is limited to about 30-50% of actual requirement, forcing the owner to abandon certain works (Hai, 2006). The selection of maintenance scope does not depend on bridge quality only, but also on other factors such as budget availability, social preference, human personality and intuition. The priority maintenance index (PMI) calculated in Equation (1) is proposed as a technique to determine socially preferential maintenance for the existing bridges on their networks (Hai et al., 2006). Certain bridges, especially those with high PMIs, can be selected for site maintenance while the total required cost remains within approved budgets. Three criteria are proposed into the PMI in accordance with country-specific condition of Vietnam as followings. •

The bridge location (BL): Evaluates the importance of an examined bridge for its specific location. Factors contributing into the value of the BL are its weighting (WBC), defined value (FBC) and attribution (αBC).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

228

Dinh Tuan Hai •

•

The health condition (HC): Reflects the physical and serviceable status of an examined bridge. Its value is combined from weighting (WHC), defined value (FHC) and attribution (αHC). The maintenance benefit (MB): The positive outcome or accrued benefit of an examined bridge if site maintenance is applied. The MB is calculated by its weighting (WMB), defined value (FMB) and attribution (αMB).

PMI = α LC W LC FLC + α HC W HC FHC + α MB W MB FMB

(1)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Establishing of Appropriate Inspection System The literature indicates that damage arises within 10 years or more after completion of construction or remedial works to adversely affect bridge safety and smooth commissioning (Nakatani et al., 2004). However, it has been reported that some types of damage require essential maintenance within four to seven years after previous repairs. In addition, the period in which bridge inspection results can be fully relied on is considered to be limited. Under adverse impacts of environment, heavy rail traffic, premature aging, etc., structural defects are likely to appear and progress quicker than initial expectations. Thus, damage if any, should be recognized at early stage to help the authorities appropriately decide on necessary works to minimize maintenance costs throughout a bridge‘s life.

Figure 6. Inspection procedure for existing railway bridges in Vietnam.

It is necessary to collect quantitative and objective information via site inspections to grasp the soundness of bridge structures, to forecast subsequent deterioration, and to decide on the type and execution time of the remedial measures. It is here proposed that an inspection procedure for railway bridges in Vietnam to consist of the six consecutive steps shown in Fig. 6. The definition of inspection is not simply limited to site surveys, but extended to the structural analysis and theoretical assessment of load-carrying capacities, reliabilities and behaviors through several uncertain options of safe-loads and ultimate-loads. The site inspection that forms the core of the proposed system is essential to detect defects and their potential causes. Three types are proposed for superficial (patrol), principal (periodic) and special inspections (Fig. 7). Superficial inspection of bridge structures is performed daily by patrol teams in accordance with formal instruction. In the periodic inspection, the condition of bridges is determined through visual observation, non-destructive

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Maintenance Management Strategies for Steel Railway Bridges

229

test and site measurement. Special inspection supplements the patrol and periodic inspections to be performed as required following any unusual incident or follow-up surveys after repairs.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 7. Categories of visual inspections for existing railway bridges in Vietnam.

Health Monitoring System The current defects occurring on railway bridges in Vietnam cannot be eliminated instantly due to budget limitation and adverse impacts of external condition. Change to the current situation requires substantial time and resources to implement a series of diversified maintenance strategies which will gradually remove the problems and their causes. In addition, health monitoring system is proposed to detect new defects and their progress, and to warn authorities on the seriousness of detected problems. Corrective remedies can therefore be carried out early to eliminate these defects and potential causes, and to minimize their adverse impacts on steel structures. An advanced health monitoring system (VR-HMS) with the computerized database is currently developed by the author to monitor the health condition of railway bridges in Vietnam (Hai, 2006). Two spheres are assigned into the VR-HMS for (a) monitoring of steel defects: cracks, fatigue damage, joint breaking, etc., and (b) monitoring of external conditions such as train weight, speed and volume, overall and local transpositions, and traffic vibration. Several local-appropriate advanced techniques (e.g. sensors, thermocouples and gauges) have been introduced into the VR-HMS to properly accomplish targeted objectives. Furthermore, a computerized inventory database is constructed to collect defect-monitoring data, to analyze the results and to report the status of monitored bridges for the authorities to take necessary corrective actions.

Preventive Maintenance Strategies Preventive maintenance is usually a time-based activity, that is, it is applied at pre-specified time periods over the bridge life cycle. It aims to eliminate potential causes of problems early, thus substantially reducing the cost, time and other resources expended on remedying future defects. If preventive maintenance is not done early, it will require more effort at a later stage

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

230

Dinh Tuan Hai

to keep the existing bridges from being critical or keep them functioning normally (Das, 1999). The mean times of initial and subsequent application of this maintenance are used as planning variables in the optimization process for determining optimized maintenance solutions. Several strategies are therefore proposed in an effort to preventively minimize the occurrence of defects on steel structures of railway bridges in Vietnam.

Early Maintenance Early site maintenance imposed on steel structures (e.g. periodical replacements of depreciated protective paints, installation of anti-collision devices, and regular cleaning) enables eliminating potential causes of future problems. Thus, deterioration of potential defects on structures can be substantially eliminated and the overall lives of bridges are definitely increased. In accordance with country-specific condition as well as their possible defects and potential causes, several types of early maintenance are tentatively recommended for railway bridges in Vietnam (Table 4). Table 4. Proposal of specific early maintenance for steel railway bridges in Vietnam.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Scope of early maintenance Periodical replacements of protective painting coats. Regular cleaning for steel surfaces and removing unexpected aliens. Routine visual inspections.

Replace for weakened and damaged steel elements. Installations of anticollision devices. Control of ongoing trains (loads and frequencies).

Locations Steel surfaces, especially hidden and access-difficult locations. All bridge structures, especially for hidden and access-difficult locations. All bridge structures, especially for and access-difficult locations. All key components such as steel decks, beams and girders. Substructures and upper-structures such as piers, abutments, and decks. Whole railway tracks from origins to destination points.

Purposes Early preventing of corrosion problems occurring on steel structures. Early preventing of premature aging problem caused by shipworms, mosses and funguses. Detect future defects and their potential causes that may occur on steel structures. Eliminate catastrophic failures and serious damages of steel structures. Prevent high-forced impacts caused by vessels, ships, trains, etc., colliding against steel structures. Prevent steel railway bridges to be overloading and overusing.

Appropriation of Standards and Specifications The bridge-related standards and specifications currently applied in Vietnam were written by local authorities. However, there were considerably compiled from many foreign standards (e.g. France, Soviet Union, US, Japan, and UK). This is due to the ownership changes of railway bridges throughout the history of Vietnam. The poor research facilities locally and limited resources coupled with high demands is a source of dilemma for bridge authorities. They have to apply these oversee standards into local practices of bridge design, construction and commissioning without test and modify them. Thus, several clauses copied from foreign sources prove to be inappropriate for the specific condition of Vietnam. Besides, several purely and locally developed standards that reflect either economic or social viewpoints seem not to be suitable for general bridge practice. There are a lot of inappropriate clauses written into the current standards, but they are scattered in different archives and issues. Their modification and correction deserve special studies by professional committees to broadly review the standards, propose alternatives, and authenticate these proposals. The author

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Maintenance Management Strategies for Steel Railway Bridges

231

carried out an intensive review of current bridge standards and legal frameworks of Vietnam and found several inappropriate clauses as listed in Table 5.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Table 5. Examples of inappropriate clauses of current bridge standards in Vietnam. Inappropriate clauses Protective painting coats are only required 0.3mm for steel surfaces.

Originated by Compiled from Soviet Union standards.

Causes of inappropriate In high-humid and salt-air condition of Vietnam, corrosion quickly occurs on steel surfaces those their protective paints are less than 0.5mm.

Safety factors are using in design for 1.5 times of actual imposed loads.

Compiled from Soviet Union standards.

Vehicle overloads and overuses are frequently two times greater than indicated maximum loads.

Used all imported materials from France in early time until 1954.

France imposed regulation for colonial country of Vietnam.

Paintings are local-made while steel have mostly imported from developed countries.

Local-made clause for cost and time deduction.

Many current steel structures were decayed as they are not suitable for Vietnamese climate that is very much defers with France climate. Paintings may not be appropriate for steel, therefore they are easily came off from their protected structures.

Maintenance agencies are not required to involve in bridge design and construction stages.

Local-made clause for separating between creation stage and operation stage.

Several problems are repeatedly happened on steel structures as previous lessons of maintenance have not been learned by designers and contractors.

Practice of using low quality materials of paintings, steels, etc.

Locally-made bridge specification for low construction cost.

Low-quality materials are not durable enough to withstand adverse climate, high moisture, etc., and prematurely degrade their quality.

Involvement of Maintenance Agencies in the Creation Stages of Bridges Currently, many defects originating during the design and construction stages have been found on steel structures of railway bridges in Vietnam that adversely decrease their quality and require high remedial costs. The creation stage needs proper controlling in order to ensure defect-free completed bridges with maintainability characteristics. The practice whereby maintenance management agencies and their staff are not required to be involved in bridge design and construction should be changed immediately. If so, the accumulated experience of maintenance management institutions can be efficiently mobilized at the early stage of bridges to eliminate repeated mistakes of design and construction. On the other hand, a good quality assurance/control system is essential to assure quality of design, construction and maintenance for bridges in Vietnam (Hai et al., 2006). It helps to create a favorable working environment where intuitive decisions are minimized and formal procedures are strictly respected throughout all stakeholders of bridges. Consequently, the occurrence of design and construction defects is unlikely and bridge-related data will be properly managed and utilized.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

232

Dinh Tuan Hai

Bridge Maintenance Management System Effective maintenance management systems for the existing bridges have been implemented in many countries. Chase and Gaspar (2000) mentioned the uses of the Pontis and the BRIDGIT at U.S. Federal Highway Administration to provide comprehensive support, to determine the optimum expenditures and to maintain a specified level of service for bridge population. Meanwhile the J-BMS was proposed for the Yamaguchi-prefectural government in Japan to evaluate bridge performance and to estimate deterioration and remaining service life. It is to generate maintenance strategies in considering actual cost, budget availability and effect of maintenance (Kawamura, 2001). In Vietnam, paper-based maintenance management is currently used; several computerized systems are however suggested for the existing bridges. One of them is the BridgeMan by British Parkman consultant under consideration (Dac, 2004). It is however limited to managing bridge inventory data only. The function for evaluation and prediction of physical condition and serviceability as well as actual expenditure and site maintenance are, so far, not available. A maintenance management model named VietRailway-BMS is hereby proposed to be applied individually by the 15 enterprises currently managing and maintaining all railway bridges in Vietnam. It consists of three modules for management, assessment and maintenance as shown in Fig. 8 (Hai et al., 2006). The management module administrates physical conditions and serviceability as well as original and updated bridge-related data. The assessment module is an intermediate stage to change from office ideas (management) to actual site-work (maintenance). It overviews bridge inventory data together with experts’ opinion, knowledge, experience, etc., in order to suggest suitable maintenance management activities for bridges. The maintenance module involves direct site-work with the objective to preserve and enhance bridge quality against the impacts of traffic, environment, elapsed time, etc. A computerized database is simultaneously constructed to store inventory data, to assess bridge conditions, and to output necessary information upon practical requirements of site engineers. Fig. 9 shows the organizational structure of the proposed database and the required activities in order from ① to ⑥.

Figure 8. Proposed maintenance management model for railway bridges in Vietnam.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Maintenance Management Strategies for Steel Railway Bridges

233

Figure 9. Organizational structure of the proposed database system.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Conclusions This paper introduces the current status of railway bridges in Vietnam and proposes maintenance strategies for eliminating structural and functional defects of steel structures and preventing the occurrence of potential causes of future problems. The results of this study are summarized as follows. 1. The historical development of railway network of Vietnam has been reviewed. In general, the existing railway bridges that were first constructed in 1881 have a wide range of shapes, commissioning dates and have been using various design standards. 2. Generally, steel railway bridges are in poor physical condition and offer poor serviceability with common failure modes of structural failure, steel corrosion, fatigue damage, functional obsolescence, aging and human intrusion. 3. The analysis of bridge maintenance management practice in Vietnam identified several major problems such as low maintenance budget, rigidity and long lines of hierarchies between bridge stakeholders, inadequate maintenance management system, and lack of equalization in maintenance selection. 4. To overcome current problems of steel railway bridges in Vietnam, a series of maintenance management strategies have been proposed. Essential strategies are suggested for short-term period in order to delay the onset of deterioration and to extend service life. Meanwhile, preventive strategies should aim to early eliminate potential causes, thus the cost, time and other resources expanded for remedies of

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

234

Dinh Tuan Hai future defects can be substantially reduced. Both essential and preventive maintenance, if done correctly, will enhance bridge quality and serviceability and significantly decrease the severity of current problems. 5. It is recommended that preventive maintenance strategies should be given priority to eliminate the causes of potential problems. In addition, essential maintenance strategies must be regularly carried out to remove the most serious defects. In order to accomplish these goals, regular training to teach all involved parties to identify failure modes, their root causes, and preventive and corrective solutions is imperative. New investment resources should be allocated to replace old and obsolesced bridges by new ones. Further researches need also to verify these proposed strategies on actual condition of railway bridges and to modify them according the facts in Vietnam.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

References Chase, SB. and Gaspar, L (2000). Modeling the reduction in load capacity of highway bridges with age. J. Bridge Engineering, 5(4), 331-336. Dac, LH. He, PV. and Hoc, BX (2004). Current conditions of concrete bridges and the need of inspection, maintenance and improvement. National Research Program DTDL 2003/2004, Vietnam. Das, PC (1999). Prioritization of bridge maintenance needs. Case studies in Optimal Design and Maintenance Planning of Civil Infrastructure Systems. Virginia, USA. Hai, DT (2006). Current status of existing railway bridges in Vietnam: an overview of steel deficiencies. Journal of Constructional Steel Research Vol. 62, 987-994. Hai, DT (2006). Maintenance management system for existing bridges – an explorative study. PhD Dissertation, Yokohama National University, Japan. Hai, DT. Yamada, H. and Katsuchi, H (2007). Present condition of highway bridges in Vietnam: an analysis of current failure modes and their main causes. J. Structure and Infrastructure Engineering. Hai, DT. Yamada, H. Katsuchi, H. and Sasaki, E (2006). Proposal of bridge maintenance management system. 3rd International Conference on Bridge Maintenance, Safety and Management. Porto, IABMAS, 113-114. Hai, DT. Yamada, H. Katsuchi, H. and Sasaki, E (2006). Proposal of priority maintenance method for network highway bridges. The Tenth East Asian-Pacific Conference on Structural Engineering and Construction, Bangkok, EASEC, Vol. 5, 393-398. Kawamura, K. Miyamoto, A. and Nakamura, H (2001). Proposal of a practical management system for bridges. 8th East Asian-Pacific Conference on Structural Engineering & Construction, Singapore. Nakatani, S. Teramoto, H. Tamakoshi, T. and Yoshida, Y (2004). Current status of Japanese road bridges and future perspectives. Ministry of Land, Infrastructure and Transport, Government of Japan. TMoVN: Transport Ministry of Vietnam (2004). Transportation infrastructure of Vietnam in 2003. Transportation Press: Hanoi. TMoVN: Transport Ministry of Vietnam (2002). The works of management, maintenance and repair of roads and bridges in Vietnam. Transportation Press: Hanoi.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Maintenance Management Strategies for Steel Railway Bridges

235

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Tuong TVD (2002). Solutions of improving and increasing capacities of road bridges at Ho Chi Minh City. Master Thesis, Transport University of Vietnam: Hanoi. NAoSRoVN: National Assembly of the Social Republic of Vietnam (2002). Law of land transportation of Vietnam. Hanoi.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

In: Railway Transportation Editor: Nicholas P. Scott

ISBN: 978-1-60692-863-9 © 2009 Nova Science Publishers, Inc.

Chapter 8

NOISE AND VIBRATIONS OF RAILWAY WHEELS: GENERATION MECHANISMS AND ATTENUATION Alfredo Cigada, Stefano Manzoni, Matteo Redaelli and Marcello Vanali Politecnico di Milano, Department of Mechanics, Via La Masa, 1 20156 Milan, Italy

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Abstract Railway noise emission and its reduction are considered among the most important topics in the future development of transportation systems, as they have a strong impact, which comes with a new line, on the people living nearby. Many elements contribute to the overall noise emission, but it has been evidenced that the most critical factor is the interaction between wheels and rails. This contribution is intended to give an overview of the noise emitted by wheels and rails from the basic emission mechanisms (rolling noise, squeal noise and impact noise) up to noise attenuation by means of passive/active control. To this purpose, the vibro-acoustic behaviour of different wheels is investigated both by means of experimental data and numerical simulations. Solid wheels, resilient wheels and wheels provided with constrained layers are deeply analysed to underline how different design strategies affect wheel emission under operating conditions. It is shown that a strict link exists between wheel mode shapes and the emitted noise. Particular attention is devoted to resilient wheels, as they are often adopted on city-tramcars, and thus their vibro-acoustic behaviour has a strong impact as an annoyance to people. Many aspects are discussed, especially those that have not yet been deeply considered in the state of the art. An example is the effect of the tread thickness value of resilient wheels on the noise emission. This comparison among different wheels also shows that a particular design approach can reduce the noise emission due to a certain generation mechanism but, on the other hand, it has no effects on other kinds of noise. The role of rails is also considered, focusing on their influence on the wheel dynamic behaviour and so on their effect on the global noise emission. Data gathered during laboratory tests are integrated and compared to data measured in different situations on real railway lines throughout Italy. City-tram lines and high-speed trains have been considered.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

238

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al. A deep comprehension of the vibro-acoustic behaviour of different wheels is the starting point towards any attempt to reduce wheel noise emission. The absence of a design approach that is effective for all of the main emission mechanisms makes the wheel-rail interaction a topic still open and where research is under continuous development. The final part of the contribution presents different possible approaches for noise mitigation and points to the new challenge: active noise reduction. The main difficulties associated with this strategy are discussed and some experimental results are presented too in order to show that, in some situations, technology is closing in on the possibility of finding new effective solutions for noise reduction.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

1. Introduction The importance of railway transportation is becoming greater and greater, as it offers many advantages when compared to other means of transportation. Many new lines are currently under construction and the building of many others is foreseen in the next years. Most attention is devoted to high-speed lines, which allow transporting people and goods over long distances in a very short time. Moreover, the adoption of a railway is considered a valuable way to transport people also within cities, contributing to fewer traffic jams and less air pollution. The increasing number of railways yields not only benefits but also drawbacks, mainly linked to people’s discomfort and sometimes to environmental pollution. One of the main problems is the noise emitted by trains and city-trams, which is often responsible for people’s annoyance. This is a critical aspect when considering surface trains, as the emitted noise can easily reach people. Another drawback is constituted by ground vibrations. Often such a problem is related to underground trains as the vibrations, generated by the train/substructure interaction ([1],[2]), can propagate through the tunnels up to houses and offices. Therefore, the problems related to the annoyance of people generated by railways are very complicated and constituted by many factors. This chapter is intended to face the problem of noise emission, giving a wide and exhaustive view of its main aspects.

1.1. Railway Noise When noise emission in the railway field is considered, it is immediately evident that many sound sources have to be accounted for, entailing a deep knowledge of many different engineering and physical aspects. Engines, inverters, wheels, rails and aerodynamic sources are only some examples of possible emitters. All of these elements contribute to the overall emission but their contribution is not in the same frequency ranges. Furthermore, the weight of each source changes with the train speed. Noise from propulsion and auxiliary systems is usually dominant at very low speed, while aerodynamic sources become very annoying for speeds higher than about 250 km/h. When the speeds are within these thresholds, the main noise contribution is due to wheels, rails and pads. Usually, sleepers contribute mainly in the frequency range around 500 Hz, rails are dominant between 750 and 1500 Hz and wheels are the main source between about 2 and 4 kHz. Nevertheless, wheel noise becomes significant over about 1 kHz ([3], Chapter 5 by M. T. Kalivoda).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

239

Furthermore, there are particular conditions under which wheel noise can become very annoying also at lower frequencies. This aspect is faced in the following discussion under the discussion of the different types of wheel noise emissions. It is evident that the problem of railway noise is very complex, and its exhaustive description would require a very wide knowledge of mechanical, thermal and fluid-dynamic fields. As wheel/rail emission is dominant in most cases, much attention is devoted to such an aspect in the state of the art. This chapter is particularly dedicated to the wheel emission. Nevertheless, a strict link exists between wheels and rails as they mechanically interact; therefore, the role of the rail is also analysed.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

2. Wheel Dynamics The sound radiation due to a wheel is generated by the wheel vibration. The mechanical excitation causing the vibration is due to the wheel-rail interaction. This excitation can introduce energy on a wide frequency range or can be characterised by few harmonics, depending on which mechanical phenomenon takes place when the wheel and the rail are in contact. This aspect is treated in the following discussion. Whatever causes the excitation, the wheel emits sound rightly due to its induced vibration. When the wheel-rail contact introduces energy in correspondence with the wheel eigenfrequencies, the wheel vibration becomes high and thus the sound emission can become loud. Not all of the wheel eigenmodes are effective in terms of sound radiation, as the mode shape and the related eigenfrequency value have an important influence on the emission level. The wheel can be approximated as an axisymmetric structure, and its main eigenmodes can be characterised by nodal lines. These nodal lines usually are nodal diameters and nodal circumferences. This means that these modes are distinguished in terms of their number of nodal diameters and nodal circumferences. Another way to name the mode shape is the following one: the axial modes are those where the main deflection is in the axial direction and nodal circumferences can exist, the flexural ones are those characterised by nodal diameters where the main tread deflection is in the axial direction and finally the radial modes are characterised by nodal diameters where the main tread deflection is in the radial direction. Figure 1 gives the sketch of the cross-section of a traditional wheel; here the two main components (tread and web) are evidenced. Figure 2 helps in the understandign of these three mode shape types. This figure shows a complete wheelset (two wheels on their axle with three brake discs) modelled by means of finite elements. Although other mode shapes can exist, these three categories are the most involved in the radiation of sound ([4],[5],[6]). Therefore, attention will be mainly focused on them. Wheel dynamics are also influenced by the contact with the rail and by the rolling movement. These aspects will be addressed further in the chapter.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

240

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

Figure 1. Cross-section of a traditional wheel (solid wheel)1.

Figure 2. Examples of mode shapes coming from a finite element model (the model represents a wheelset with solid wheels); (left) axial mode shape with a nodal circumference; (centre) flexural mode shape with two nodal diameters (the tread main deflection is out of the wheel plane); (right) radial mode shape with three nodal diameters (the tread main deflection is in the wheel plane)1.

1

This figure was published in Applied Acoustics, vol. 69, by A. Cigada, S. Manzoni and M. Vanali, Vibro-acoustic characterization of railway wheels, pp. 530-545, copyright Elsevier (2008).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

241

2.1. Wheel Sound Emission The sound power Wrad from a vibrating structure is given by ([3], Chapter 1 by D.J. Thompson):

Wrad = ρcSσ v 2

(1)

where ρ is air density, c is the speed of sound in air, σ is the radiation ratio (or radiation

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

efficiency), S is the surface area of the structure and v

2

is the spatially averaged mean

square normal velocity of the vibrating surface. The radiation ratio σ is defined [7] as the ratio of the average acoustic power radiated per unit of area of a vibrating surface to the average acoustic power radiated per unit of area of a piston, that is vibrating with the same average mean square velocity at a frequency for which the piston circumference greatly exceeds the acoustic wavelength. At high frequency σ tends to 1 while at low frequencies it is much lower than 1. It is stated in literature ([3], Chapter 1 by D.J. Thompson) that the wheel flexural modes with n (with n>0) nodal diameters usually are characterized by a σ which depends on the mode frequency value f (at low frequency values) and it is proportional to f 2n+4, while for radial modes the dependence is f 2n+2. Finally the dependence for axial modes (n=0) is f 4. As already mentioned, when the frequency gets higher σ is nearly 1. This means that the same eigenmode at two different frequency values does not provide the same sound emission, once the force excitement is kept constant. Concerning the directivity of wheel sound emission, it is stated in the state of the art that it depends on many parameters such as wheel shape, mode shape and frequency ([3], Chapter 1 by D.J. Thompson). The next section discusses the main mechanisms able to give raise to a significant wheel noise emission. Then, after having explained wheel dynamics and noise phenomena, it will be possible to discuss in details the vibro-acoustic behaviour of different wheels, which are built nowadays.

3. Noise Generation Mechanisms As already underlined, the wheel cannot be considered as a standalone body but the coupled components have to be considered too, when its dynamic behaviour is analysed. As the rail is in direct contact with the wheel it is easy to understand that there is a deep mutual influence between them. Different mechanisms can give rise to loud, or significant at least, noise emissions. The main ones are “rolling noise”, “squeal noise” and “impact noise”. They are discussed one by one in the next sections.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

242

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

3.1. Rolling Noise Both wheel and rail surfaces, in contact during the rolling movement of the wheel, are characterized by the presence of roughness. This is generated by many factors; one is the braking action which forces the wheel to slide on the rail. Something similar is related to accelerating trains. This roughness is the main factor in rolling noise generation: when the wheel is travelling along the rail, the roughness present on both the bodies generates relative displacements between them. This leads the bodies to exchange force actions. Then these loads cause wheel and rail vibrations and thus sound emission. The wheel-rail contact does not occur at a single point but rather over an area, named contact patch, which is usually about 10–15 mm long and with a lateral width of about the same dimension. When roughness wavelengths are short, if compared to the contact patch radius, their effect on the excitation on the wheel-rail system is reduced because of averaging across the contact patch (this effect is called contact filter) ([8]; [3], Chapter 1 by D.J. Thompson). Moreover irregularities with a wavelength λ induce an excitation at frequency f described by the following law:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

f = V/λ

(2)

where V is the train speed ([8], [9]). This means that the same roughness profile would cause vibrations and noise at different frequencies depending on the train speed. Usually the frequency range at which rolling noise is significant is limited up to about 5 kHz [10, 11, 12]. It has already been mentioned that wheel contribution is dominant about between 2 and 4 kHz, rails mostly contribute between 750 and 1500 Hz and sleepers are dominant around 500 Hz. Considering an interesting frequency range between 300 Hz and 5 kHz, the more the train speed increases the more the range of wavelengths, important for rolling noise generation, shifts towards higher values. At normal speeds the wavelengths between 1 mm and 100 mm are of interest. It is now evident that the rolling noise is characterized by a broadband sound emission since it is due to broadband actions which excite many eigenfrequencies. Other parameters, such as pad and fastener behaviours [13], are able to affect rolling noise but with a lower influence than the wheel and rail roughness. In the past years many models attempting to deeply describe the rolling noise birth have been developed. One of the most important has been worked out by Remington and Thompson ([4],[14]-[19]) and describes the coupling between the rail and the wheel. The interaction between the two bodies is treated introducing stiffness elements simulating the contact and the roughness. The equation system describing the dynamic behaviour of the global system presents coupled equations confirming the high mutual influence of the wheel and the rail. As roughness is one of the major elements involved in rolling noise generation, many studies have been carried out on it ([20], as an example). One of the most interesting is [11] where a sort of linear dependence (in logarithmic scale) has been experimentally found between the wheel roughness and the relative emitted noise (for not too high roughness values; with high roughness values some non-linear effects are present) relying on other previous studies. The presence of roughness on the rails does not affect the linear relationship

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

243

till when the levels of the wheel and rail roughness are similar. A global model, considering all the mentioned topics, is named TWINS [9, 10] and it is based on the previously mentioned model from Thompson. In such a model the noises coming from wheels, rails and sleepers are kept distinguished, as if they are different sound emitters. Another remarkable aspect concerning roughness is that it is in some way linked to the kind of adopted brake system [21–23]. The adoption of tread brakes is much more critical than the adoption of disc brakes from the point of view of roughness generation. When tread brakes are used, the wheel slides on the block and there is a certain critical temperature above which there could be thermoelastic instabilities. The consequence is that hot spots can appear and these heated areas expand radially in a way depending on the material nature. This causes an increased contact pressure and so, once again, thermal expansion closing the loop. The areas interested by the presence of hot spots are thus mechanically and thermally very stressed and this gives raise to wear in a faster way than on the other wheel areas. Sometimes there could even be material phase transformations, due to the high temperatures. This phenomenon is not important in terms of rolling noise when disc braked wheels are considered since the breaking action is performed on a surface which is not directly involved in the wheel-rail interaction.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

3.2. Squeal Noise While rolling noise is characterized by a broadband emission, the main feature of the squeal noise is that the emission is concentrated on one or at least few harmonics, generating a sound which is felt by the human hear just like a squeal, as suggested by the name. Such a feature is considered very annoying for the human beings [24]. Although the emission comes again from the interaction between the wheel and the rail, the generation mechanism is totally different with respect to the rolling noise. Squeal noise usually arises when a train negotiates a curve, especially a short radius one. Two main mechanisms can give raise to such a phenomenon [25]: •

•

squeal originates as the wheel velocity is made up of a rail tangential component and another one normal to it (usually called creep velocity); the lateral creepage of the wheel causes a stick/slip phenomenon. The consequence is that forces, orthogonal to the wheel, are induced and they excite the wheel flexural eigenmodes, thus generating the noise emission [25–28]. squeal can also be originated due to contact on the wheel flange [25].

The excited eigenfrequencies are self-excited and the phenomenon, which governs the vibrations, is a stick/slip one, so non-linear (as already mentioned). The slip behaviour is responsible for the instability of some wheel modes and drives them to grow exponentially. The stick behaviour is responsible for the limitation of this growth and imposes a limit cycle [26, 27]. This phenomenon is in some ways similar to what takes place with disc brakes, which suffer of squeal noise too. It is stated in literature [29, 30] that disc brake squeal appears when there is a coupling between an eigenmode of the disc and one of the pad (which is directly in

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

244

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

contact with the disc). Something similar has been experimentally observed between the wheel and the rail in [31]. Squeal noise comprehension is probably much more complicated than that of rolling noise as the non-linear effects play a crucial role. Furthermore there are many geometrical and mechanical variables able to affect the phenomenon [32], such as the lateral contact position of the wheel on the rail [33]. Before concluding the discussion on squeal noise, it is important to underline that such a sound emission can be loud also at low speeds (lower than 15 km/h), as confirmed also by reference [34].

3.3. Impact Noise Another wheel noise mechanism which is worthy of attention is the impact noise [35], which appears when the wheel meets important irregularities during its rolling. Examples of these irregularities could be rail junctions or important local defects on the wheel rolling surface itself. The most important parameters determining the emission level are the train speed, the irregularity geometry, the static load on the wheel and the rail kind. For example a rail sunk in the ground makes emission lower than a rail which is not sunk. Moreover, considering a not sunk rail, a step shaped obstacle generates more noise if the wheel faces it in descending direction than in ascending one [35]. On the other hand there are not great differences between the two situations for an immersed rail.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

4. Investigation of Different Railway Wheels Nowadays different kinds of railway wheels are used on trains and city-trams. The most analysed one is the solid wheel (often used on high-speed trains), whose structure is quite simple. It is constituted by a single steel part and a cross-section example is given in Figure 1. Sometimes a thin layer of highly damped viscoelastic material is placed between the web and a stiff constraining plate. This solution is able to effectively increase the overall damping and thus to reduce the wheel noise emission. A further solution is the resilient wheel (often adopted on city-trams). In this case the wheel is constituted by two steel parts and a certain number of rubber blocks is placed between them. A sketch of a resilient wheel cross-section is provided in Figure 3. The structure and the dynamic features of these three wheel kinds can be very different and therefore it is expected that their vibro-acoustic behaviours can be different as well. The next sections discuss the dynamic and acoustic characteristics of these wheels in order to highlight the reasons why they are responsible for different noise emissions. The wheel heavily interacts with the rail under operating conditions. This would require to analyse the wheel vibro-acoustic behaviour when the wheel is laid down on the rail. Nevertheless this complicates the analysis as wheel eigenmode identification becomes harder as wheel eigemodes can be not easily distinguishable from those related to wheel-rail interaction. This has pulled towards testing the wheels in free suspended condition. The tests are constituted by wheel modal analysis coupled to sound pressure measurements. It has

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

245

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

already been found out that the experimental tests under free suspended conditions are able to give reliable information on wheel sound emission under operating conditions [6]. Nevertheless the effect of the contact between the wheel and the rail is analysed too in Section 5.

Figure 3. Cross-section of a resilient wheel1.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

246

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

All the wheels have been tested mounted on their own bearings or on axles to stay close to real operating conditions as it is demonstrated [4] that this adds damping to certain eigenmodes. The accelerometer mesh for the modal analyses is constituted by a number of sensors which depends on the wheel size. Particularly the accelerometer layout has always been chosen in order to identify modes with a number of nodal diameters up to 5 and a number of nodal circumferences up to 2. Nevertheless it has been possible to recognize radial or flexural eigenmodes with a higher number of nodal diameters, although this number could not be found out due to spatial aliasing. The microphones used for the sound pressure measurements were aligned with the wheel main axis and they were at about 1.4 m from the wheel. This distance allows to account for the global wheel emission in the interesting frequency range (400–5000 Hz) and to have not any significant acoustic local effect.

4.1. Solid Wheel

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

It has already been mentioned that the structure of a solid wheel is constituted by a single steel part. Figure 1 shows the cross-section of a solid wheel currently adopted on high-speed trains of a European country. This wheel has been analysed and the results are discussed in the following as representative of the solid wheel type. The wheel was mounted on its own axle with the other wheel and three brake discs (which is the operating condition) for the abovementioned reason.

Figure 4. Tread point accelerances—axial direction (solid line) and radial direction (dotted line)1.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

247

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 4 (where FRF means Frequency Response Function) shows the tread point accelerances for the radial and axial directions. It is evident that a high number of modes exist up to 5 kHz. This is critical under the rolling noise point of view as many eigenmodes are excited during wheel rolling. Another critical aspect is related to the very low values of damping ratios. Flexural and radial eigenmodes are usually characterised by a damping ratio of about 0.01 % or a little higher [6]. Axial modes have low damping ratios too but when the axle is involved in the mode shape the damping ratio can increase [4, 6] up to about 0.3 %. Looking at the microphone results, it comes that the radial modes are the most important in terms of sound emission (Figure 5). This could seem odd as it would be expected that modes dominated by axial deflections are those characterised by the highest sound emission. This fact can be explained investigating the behaviour of the web. Figure 6 shows that a significant axial web deflection exists in correspondence of the radial eigenmodes; this strong coupling between tread radial deflections and one or two-circles web axial ones is the reason why radial eigenmodes are responsible for the loudest sound emission [6]. The same coupling does not exist when flexural eigenmodes are considered. It has to be underlined that the mentioned coupling is significant at least when curved webs are considered [4, 5] which is the present case. A finite element/boundary element model of the experimentally tested wheelset has been developed in order to investigate in details its vibro-acoustic behaviour [36]. Both the dynamic (finite elements) and the acoustic (boundary elements) models have shown reliable results up to 2 kHz when compared to experimental data; limited frequency shifts (less than 6 %) take place in correspondence of the main wheel eigenfrequencies. The acoustic model has allowed to find out the role played by the tread and the web in the overall sound emission in correspondence of different eigenmodes.

Figure 5. Amplitude of the frequency response function averaged on different impacts (radial and axial on the tread and axial on the web). The arrows indicate radial eigenmodes1.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

248

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 6. Amplitude of the frequency response functions for a radial impact excitement on the tread. Solid line for the mean of the axial responses of many accelerometers placed on the wheel web and dotted line for the mean of the radial responses of many accelerometers placed on the tread. Squares and circles are in correspondence of flexural modes (squares for the solid line and circles for the dotted line) while triangles indicate radial modes (tip up for the solid line and tip down for the dotted line)1.

Figure 7. Acoustic pressure emission of the web and the tread for an axial impact on the tread (boundary element simulation). The circles indicate the peaks of the dashed line (web) while the squares indicate the peaks of the solid line (tread). F indicates flexural modes while R indicates radial modes. The dB reference value is 2*10-5 Pa.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

249

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 8. Acoustic pressure emission of the web and the tread for a radial impact on the tread (boundary element simulation). The circles indicate the peaks of the dashed line (web) while the squares indicate the peaks of the solid line (tread). F indicates flexural modes while R indicates radial modes. The dB reference value is 2*10-5 Pa.

Figure 9. Radiated acoustic power of the web and the tread for an axial impact on the tread (boundary element simulation). The circles indicate the peaks of the dashed line (web) while the squares indicate the peaks of the solid line (tread). F indicates flexural modes while R indicates radial modes. The dB reference value is 10-12 W.

Figure 7 and Figure 8 (where R and F indicate the radial and flexural eigenmodes) provide the simulated acoustic pressure due to the tread and the web at a field point far away

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

250

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

from the wheel 1400 mm and aligned with its axis for an axial and a radial impact on the tread, while Figure 9 and Figure 10 display the radiated acoustic power for the same elements and excitations. The results of the radial and the axial impact tests cannot be directly compared as the force power spectra are different in order to stay close to the real performed tests. It is possible to note that the web global contribution is higher than that of the tread. This is mainly due to the fact that the web surface is much wider that that of the tread.

Figure 10. Radiated acoustic power of the web and the tread for a radial impact on the tread (boundary element simulation). The circles indicate the peaks of the dashed line (web) while the squares indicate the peaks of the solid line (tread). F indicates flexural modes while R indicates radial modes. The dB reference value is 10-12 W.

Considering the flexural eigenmodes at 907 and 1650 Hz, it is evident that the contributions of web and tread are pretty similar in terms of acoustic pressure at the field point. Relying on the fact that the web surface is much higher than that of the tread, this means that there is a low coupling between the tread and the web deflections for flexural eigenmodes, as already evidenced by the above experimental activity. On the other hand, a different situation takes place in correspondence of the flexural eigenmode at 329 Hz. In this latter case the web contribution (in terms of acoustic pressure at the field point) becomes more important if compared to the former modes. This could mean that the coupling between web and tread deflections gets higher at 329 Hz. Nevertheless a careful analysis of Figure 9 and Figure 10 evidences that the tread and web contribution in terms of acoustic power are similar. Relying once again on the fact that the web surface is much wider than that of the tread (now also the rear surface of the web has to be considered as acoustic power is accounted for), it comes that also in this case the mentioned coupling is low and the vibration activity of the tread is higher if compared to that of the web. Considering radial modes, the only one present within the 0–2 kHz range is that at 1551 Hz. It is evident that the contribution of the web is much higher than the tread one, in terms of both acoustic pressure and acoustic power. This means that a high coupling exists

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

251

between tread radial deflections and axial web ones, just in agreement with the experimental outcomes.

4.2. Solid Wheel with Viscoelastic Layers These wheels have a structure similar to that of the traditional solid wheels. Furthermore a thin layer of highly damped viscoelastic material is placed between the web and a stiff constraining plate (Figure 11). This is intended to reduce the noise emission due to the web.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 11. Sketch of the cross-section of a solid wheel with viscoelastic layers.

Figure 12. Mean axial response of the web (the mean has been calculated on many accelerometers) for a radial impact on the tread. The solid line represents the solid wheel with viscoelastic layers and the dotted line represents the traditional solid wheel.

Figure 12 provides the mean web axial response (averaged on many accelerometers placed on the web) for a radial impact on the tread. The same graph is provided for the above analysed traditional solid wheel in the same figure for a direct comparison. It is evident that a

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

252

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

significant improvement is obtained in terms of vibration reduction. The direct consequence is that the sound emission is reduced too. This is highlighted by Figure 13 where the measured acoustic pressure is displayed for the same wheels and the same excitement.

Figure 13. Acoustic pressure response for a radial impact on the tread. Solid line for the solid wheel with viscoelastic layers and dotted line for the traditional solid wheel.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Significant benefits in terms of rolling noise attenuation can be obtained as the action of the viscoelastic layers gives vibration reduction on a wide frequency band, as shown by the above figures.

4.3. Resilient Wheel It has already been mentioned that this kind of wheel is built up by two steel parts: an inner part and an outer one. A variable number of rubber blocks are placed between them in order to isolate the two components and to add damping to the global structure. Figure 3 shows, as an example, the cross-section of a resilient wheel mounted on the city-trams of an important European city. These wheels are nowadays widely used on city-trams and thus their properties, in terms of sound emission, are very important. This is due to the fact that city-trams run very close to houses and therefore their eventual noise radiation could be very annoying for people living there. Resilient wheel vibro-acoustic behaviour has not been widely investigated in the state of the art as in the case of the traditional solid wheel. Nevertheless some key points have been highlighted in different works. First of all it has been experimentally demonstrated by means of field measurements [37] that resilient wheels are effectively able to reduce rolling noise, if compared to solid wheels ([3], Chapter 6 by C.J.C. Jones and D.J. Thompson; [38]). Therefore solid wheel noise is dominated by rolling noise, although squeal can be important

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

253

too, while resilient wheel is mainly interested by squeal noise and rolling noise is effectively reduced thanks to its mentioned structure. Most literature has focused on the role played by the rubber blocks ([3], Chapter 6 by C.J.C. Jones and D.J. Thompson; [39]; [40]) as they are the main components influencing the wheel dynamics. Particularly it has been demonstrated that the stiffer are the blocks, the higher is the damping and the lower is the uncoupling between the inner and the outer part. A modal analysis of the resilient wheel of Figure 3 evidences that the damping ratios associated to flexural and radial eigenmodes are often between 0.15 % and 0.5 % and sometimes they can grow up to about 2 % [6].

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 14. Tread radial point accelerance for the resilient wheel1.

Figure 15. Tread radial point accelerance for the solid wheel of Figure 11.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

254

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 14 and Figure 15 present the tread radial point accelerances for the resilient wheel and the above discussed solid wheel. It is evident that the maxima peaks are much lower in the case of the resilient wheel and there are less eigenmodes up to 5 kHz. Furthermore Figure 16 shows the web mean axial responses for a radial impact on the tread for the same wheels. The benefit provided by the resilient wheel structure is evident. While solid wheel sound radiation is dominated by radial modes (due to the coupling between web axial deflections and tread radial ones; see Section 4.1), that of resilient wheels is dominated by flexural modes. This is almost due to the fact that the coupling between web axial deflections and tread radial ones is absent (or very low) in the case of resilient wheels.

Figure 16. Mean axial response of wheel web for a radial impact on the tread. A solid line represents the solid wheel and dashed line represents the resilient one1.

The mentioned resilient wheel has been tested twice. The figures shown up to now concern a set-up with 22 already used rubber blocks. The same wheel has been then tested with 26 brand new rubber blocks in order to experimentally investigate the overall effect given by an increase of the rubber block number. Figure 17 shows a comparison between the two wheel layouts in terms of sound emission. Although a general increase of the damping is expected when the number of blocks grows up, it is evident that some eigenmodes (for example that at 770 Hz) are interested by an increase of the sound emission level. This is particularly dangerous if such a mode is a flexural one (which is the present case) as they are the most involved in squeal noise emission. This means that a louder emission is obtained with 26 blocks rather than with 22 blocks, in the case squeal takes place in correspondence of the mentioned mode.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

255

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 17. Amplitude of the frequency response function between an axial impact on the tread and the microphone response. The solid line represents the wheel with 22 blocks and the dotted line represents the wheel with 26 blocks. Squares are in correspondence with the flexural modes for the wheel with 22 blocks while circles indicate the flexural modes for the wheel with 26 blocks1.

Attention has been often focused on rubber block properties in the state of the art, as they are the most effective features able to affect wheel behaviour. Nevertheless other construction features have to be accounted for, if a wider comprehension of wheel behaviour is required. Another resilient wheel has been tested to this purpose. This further wheel is mounted on citytrams of another important European city. These additional tests have allowed to look for possible vibro-acoustic behaviour differences between the first wheel (Figure 3, named wheel A in the following) and the latter one (named wheel B in the following). The external diameters are similar (600 mm for wheel A and 660 mm for wheel B) and the ratio between the tread external diameter and the tread internal diameter is similar too. In fact the ratio of wheel B is just 2 % higher than that of wheel A. The main differences between the two wheels are the thickness value (about 115 mm for wheel A and less than 85 mm for wheel B) and the structure of the wheel inner part. The number of rubber blocks is 26 for wheel A and 25 for wheel B and the vertical stiffness is very similar and differs less than 10 %. As mentioned, wheel B has a lower tread thickness so that its eigenfrequency is at a lower frequency value with respect to wheel A [41], once a certain mode shape is considered. Moreover the external radius of wheel B is higher than that of wheel A, which is a further fact causing eigenmodes at lower frequency values for wheel B [41]. Although the tread axial and radial point accelerances (Figure 18 and Figure 19) show similar vibration behaviours (similar number of eigenfrequencies up to 5 kHz, similar levels of the maxima peaks, similar identified damping ratios and so on), some differences can be highlighted in the behaviour of the webs at low frequency (below about 1 kHz). Looking at Figure 20 and Figure 21, it can be noted that the flexural mode with two nodal diameters (770 Hz for wheel A and 532 Hz for wheel B) has a much higher response for wheel A than

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

256

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

for wheel B when a radial excitation is considered. One of the main reasons is linked to the behaviours of the two treads. That of wheel A has a significant response for both the excitation directions while that of wheel B shows a high response only when it is excited in axial direction. This means that there is a high coupling between tread axial and radial deflections for wheel A but not for wheel B. It comes that the vibration response of web A is much higher than that of web B, once the radial force on the tread is kept constant and the mentioned mode (flexural with two nodal diameters) is considered. Therefore the sound emission of web A is louder than that of web B under the mentioned conditions (this is confirmed by experimental measurements).

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 18. Tread point accelerances for wheel A. The solid line indicates axial direction and the dotted line represents radial direction.

Figure 19. Tread point accelerances for wheel B. The solid line indicates axial direction and the dotted line indicates radial direction.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

257

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 20. Axial web response (mean of the response of many accelerometers placed on the wheel web) for wheel A. The solid line represents an axial impact on the tread and the dotted line represents a radial impact on the tread.

Figure 21. Axial web response (mean of the response of many accelerometers placed on the wheel web) for wheel B. The solid line indicates an axial impact on the tread and the dotted line indicates a radial one.

It is interesting to find out which are the reasons of the different degree of coupling between axial and radial tread deflections for the two wheels. One of these reasons is linked to the fact that wheel A tread thickness is higher than that of wheel B. A finite element model of an asymmetric steel ring, whose size is similar to those of the treads of the two considered resilient wheels, has been developed. Some dynamical analyses have been performed changing the thickness value step by step and investigating its effect. The absolute value of the ratio between the ring axial and radial deflections for the first

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

258

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

flexural and radial eigenmodes with two nodal diameters (the deflections have been computed in correspondence of an antinode of the considered modes) has been computed for the different adopted thickness values (Figure 22). This last figure shows that the lower is the thickness, the lower is the coupling between ring axial and radial deflections.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 22. Absolute value of the ratio between the axial and the radial deflections for the first flexural (up) and the first radial (down) modes with two nodal diameters as a function of the tread thickness. The ratio is calculated in correspondence of a tread antinode on the edge between tyre and tread outer lateral surface. The solid line indicates the flexural mode and the dashed line indicates the radial one.

A direct consequence is that when the mentioned eigenmode is involved in squeal emission, the web of wheel A will emit more than that of wheel B. It has been evidenced in the above discussion that wheel thickness value has significant fallouts on the vibro-acoustic behaviour of resilient wheels. It is possible to summarise them in the following way. An increase of the tread thickness value means that: •

•

•

the coupling between ring axial and radial deflections (at least for modes with two nodal diameters) increases and thus more vibration energy passes from the tread towards the web; once a certain mode shape is considered, the corresponding eigenfrequency is at a higher frequency value. This means that the eigenfrequency is in a frequency field where the radiation ratio (Section 2.1) gets higher (or at least not lower). A similar effect is due to external wheel radius: the lower is the radius, the higher are the eigenfrequency values. Furthermore, it is stated in the state of the art [42, 43] that the critical frequency for circular plates usually decreases when the plate thickness increases (where the critical frequency fc is an important parameter in terms of radiation efficiency: sound at frequencies higher than fc is radiated efficiently while sound at frequencies lower than fc is radiated less efficiently [44]). the tread radiating surface is higher.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

259

All the mentioned points cause an increase of the emitted noise. The influence of tread thickness is important, especially when squeal noise is considered. Let make the hypothesis that the flexural eigenmode with two nodal diameters (it is recalled that flexural eigenmodes are highly involved in the squeal phenomenon) is responsible for squeal noise emission. A thicker tread means that the web is characterised by a higher vibration level (due to the coupling between tread radial and axial deflections), that the tread emitting surface is wider and that the considered eigenfrequency is in a frequency range where the noise is efficiently emitted. It comes that, although the wheel performances in terms of rolling noise mitigation are mainly influenced by rubber block properties as evidenced in literature, the wheel behaviour under the squeal noise point of view is linked also to other features, just like tread thickness. This is due to the fact that squeal is associated to one or few eigenfrequencies; it has been experimentally and numerically found that variables different from rubber block properties can have significant influences on limited frequency ranges and on few eigenmodes, although the global behaviour is dominated by the rubber block features. The discussed topics underline that there are construction parameters, different from the rubber block properties, which are effectively able to influence the behaviour of resilient wheels under the squeal noise point of view. Thus they have to be deeply considered when a new wheel is designed.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

5. Wheel-Rail Contact The discussion on the vibro-acoustic behaviour of railway wheels has been performed up to now considering wheels in free suspended condition. This section is intended to show which are the main effects due to the contact with the rail. To this purpose the same wheelset (provided with solid wheels) discussed in Section 4.1 has been laid down on pieces of rails and a modal analysis has been carried out. The rail pieces have been fixed to the ground with supports really used on rail tracks and the accelerometer mesh has been the same of that adopted during the first modal analysis. The experimental results have then been compared to those of the modal analysis performed in the free suspended configuration. A general increase of damping is expected from the state of the art for those modes (characterised by nodal diameters) presenting an antinode at the contact point, while those with a node at the contact point should not be affected by a significant change of the damping ratio ([3], Chapter 1 by D.J. Thompson). The results from the modal analysis confirm this point but a further fact has to be highlighted. There are some axial modes (thus not having nodal diameters) which are not affected by a change of the damping ratio, although axial deflections are expected to exist at the contact point. One reason can be the following: often axial modes are coupled to axle deflections [4], which adds damping. Probably the damping ratios of these modes are already so high in the free suspended condition, that the contact with the rail is not able to give a further increase of the damping. When the laid wheel is analysed, it can be noted that many new eigenfrequencies can be identified. Often they are characterised by the presence of nodal diameters. Figure 23 shows the tread axial point accelerances for the suspended and the laid conditions (in addition some transfer functions between the axial impact on the tread and the responses of some accelerometers measuring the axial vibration in different tread points are given). It comes that

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

260

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

the spectrum is much denser in the case of the laid condition, due to the mentioned newly borne resonances. Furthermore the flexural eigenmodes are characterised by lower peaks for the laid configuration, if compared to the free suspended condition. The same conclusion can be stated for the radial modes.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 23. Tread axial point accelerances for the suspended and the laid conditions (in addition some transfer functions between the axial impact on the tread and the responses of some accelerometers measuring the axial vibration in different tread points are given for the laid condition). Solid lines indicate the laid configuration and dashed line indicates the suspended free one.

Figure 24 shows a zoom of Figure 23 where a newly borne resonance is evidenced by an arrow. The reason explaining most of these new eigenfrequencies can be obtained from the state of the art. Considering the suspended free wheel, there are some eigenmodes which are characterised by the fact that they are at two slightly different frequency values and that the nodes of one mode are more or less at the same position as the antinodes of the partner eigenfrequency and vice versa [4]. This is confirmed also by the finite element model of the wheelset (Section 4.1) and sometimes these “twin modes” are hardly distinguishable with experimental modal analyses. Situations like these have been found also in [4] and the phenomenon has been explained with slight imperfections in the wheelset axial-symmetry. This could give a suggestion about the observed difference between the suspended free and the laid conditions: when the wheel is laid down on rails the contact between the wheel and the rail would bring the wheel far away from a perfect axial-symmetry. As a consequence, the two eigenfrequencies, previously very close each other, would become more separated, highlighting the presence of the second eigenfrequency. This is critical in terms of rolling noise because the wheel response spectrum becomes denser. This means that the roughness wavelengths which can excite the wheel, at a certain fixed train speed, grow in number.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

261

Figure 24. Zoom of Figure 23. A new resonance is evidenced by the arrow. Solid lines indicate the laid configuration and dashed line indicates the suspended free one.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

It has been discussed in Section 2.1 that solid wheels with curved webs show an important coupling between axial web deflections and tread radial deflections and that this is critical in terms of noise emission. The same coupling has been experimentally evidenced also with the laid configuration, as also confirmed in [18]. This means that those modes presenting such a coupling (radial modes) are really able to give a significant noise emission under operating conditions. This will be discussed again in Section 7. In addition it has been found that some axial modes exist, which are affected by an increase of their transfer function peaks when the wheel is laid down on rails (Figure 25).

Figure 25. Zoom of the transfer function amplitudes between a radial impact on the tread and some web accelerometer responses evidencing the growth of the transfer function peaks for an axial mode in the laid configuration (dashed lines) with respect to the suspended free condition (solid lines).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

262

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

It can be concluded that the behaviours of the same wheel in the two different considered configurations are not heavily different as the same dynamic features can be found in both the layouts. Nevertheless it has been shown that some differences exist and that they can be of some importance when noise emission under operating conditions is considered.

6. Effect of Wheel Rotation The effect of rotation is particularly meaningful when modes with nodal diameters are considered. Each of these modes can be represented ([3], Chapter 1 by D.J. Thompson; [19]) by two rotating waves with angular dependence e ± i ( nθ + ωt ) where θ is the angle around the wheel, ω is the rotational speed (in rad/s), n is the number of nodal diameters of the considered mode and t is the time. These two waves have an amplitude which is the half of the mode amplitude and have opposite directions, as evidenced by the term e ± iωt. The composition of these two waves gives a standing-wave pattern, which is the mode shape. When the wheel is rotating with a rotation speed Ω, the two waves rotates at different rotational speeds (ω ± nΩ in module). Thus the wave rotating with the wheel becomes faster, while the other becomes slower. This means that the resonance peak is split into two peaks, corresponding to the two waves rotating at different speeds. Therefore, once wheel speed is kept constant, the roughness wavelengths able to excite the wheel at resonances increase. No fixed interference can occur between two partner waves both in rotating and nonrotating frames ([3], Chapter 1 by D.J. Thompson; [19]).

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

7. Rolling Noise under Operating Conditions It has already been mentioned that rolling noise (Section 3.1) is caused by the interaction between wheels and rails, and particularly by the roughness on wheel and rail surfaces. Some field measurements have been performed on Italian lines where high-speed trains usually travel. These high-speed trains mount solid wheels just like those presented in Section 4.1. The considered line is characterised by the presence of two different substructures in close proximity. This has allowed to test the acoustic emission due to the two different substructures at the same conditions (kind of trains, kind of mounted wheels and so on). These two substructures are the so called “Mass Spring” and “Slab Track”. In the case of the Slab Track, the rail is attached by resilient fasteners to a concrete base; in the case of Mass spring, the slabs to which the rails are connected are floating designed to act as seismic isolators. Firstly, the results of far field microphone measurements [45] are provided in terms of third-octave band, which is the usual layout in literature. This allows for a direct comparison with other measurements carried out within other research works. Then, microphone array measurements [46] of the noise emitted by wheelsets are discussed in relation to the outcomes of the modal analyses carried out on the considered solid wheel (Sections 3.1 and 5). Figure 26 presents the far field measurement for Mass Spring at different train speeds and Figure 27 gives the same kind of data for Slab Track (noise due to wheel/rail interaction is dominant at the considered speeds, as already mentioned in Section 1.1). It is evident that an

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

263

increase of train speed causes an increase of noise emission, especially at high frequency, where the wheel noise is dominant. Such a phenomenon is partly due to the higher energy exchanged by wheels and rails during their impacts (caused by the roughness) due to the higher speed. Nevertheless it has to be recalled (Section 3.1) that a change of train speed causes a change of the relationship between the excited vibration frequencies and the exciting roughness wavelengths (Equation 2). The dependence of the overall noise emission on the train speed is complex and is strongly affected by the wavelengths characterising the wheel/rail roughness.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 26. Third-octave band noise emission for different train passages on Mass Spring.

Figure 27. Third-octave band noise emission for different train passages on Slab Track.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

264

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

Figure 28 shows a comparison between the noise emission during passages on the two considered substructures at almost the same train speed. It comes that, under the present conditions, the noise due to the track is higher in the case of Mass Spring but wheel noise increases when slab Track is considered. This evidences that the strong coupling of many elements makes this topic very complicated. For example, a solution able to decrease wheel noise could increase rail noise and finally the overall effect could be null or worse. Microphone array processing [46] has been carried out on the same train passages. The aim was to recognise the noise emission only coming from the wheel (actually a complete separation between wheel and rail noise is not possible as these two interacting elements are spatially close).

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 28. Third-octave band noise emission for different train passages on Mass Spring and Slab Track at nearly the same speed.

Figure 29. Noise emission from train passage on Slab Track (array processed measurements averaged on four wheelsets).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

265

Figure 29 shows the sound emission due to wheels during a train passage. A careful comparison between this figure and Figure 4 (this latter figure shows the radial and axial point accelerances of the solid wheel in the free suspended condition) evidences that wheel noise is high in those frequency ranges characterised by many radial and flexural eigenmodes (for example the range between 1600 and 1800 Hz). On the other hand, the noise emission is definitely lower within the frequency ranges where there are none or few of these modes (for example between 1200 and 1600 Hz). This is a confirmation that the experimental modal analysis (coupled to acoustic pressure measurement) of the free suspended wheel is able to give reliable information on wheel sound emission under operating conditions (Section 4).

8. Railway Wheel Noise Attenuation

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Many works have focused on the possibility to mitigate railway wheel noise. It has already been mentioned that the wheel interacts with rails and, consequently, with the substructure. This means that wheel noise mitigation cannot be treated as a standalone topic. Other elements have to be accounted for (i.e., rails, pads and so on). The strong coupling of many elements makes this topic very complicated. For example, a solution able to decrease wheel noise could increase rail noise and finally the overall effect could be null or a worsening could be even got. Very different approaches can be adopted. These can be split into two main categories: passive and active solutions. Other subtypes can be individuated within these two categories; a possible further subdivision can be attempted relying on the elements on which the mitigation approach is applied to. A brief description of the most important attempts is given in the following (a part of the discussion on the passive remedies is based on references ([3], Chapter 6 by C.J.C. Jones and D.J. Thompson; [38]).

8.1. Passive Systems for Wheels 8.1.1. Damping Treatment Although solid wheels show very low damping levels, the contact between wheels and rails causes an increase of wheel damping (Section 5). Thus the addition of further damping cannot represent the only and ideal solution to get high rolling noise reduction. On the other hand, wheel-damping treatments could be very effective in the reduction of squeal noise. Many kinds of dampers have been used through all these years. Most of them could offer good performances in noise reduction. Nevertheless, it has to be remembered that a key point for an effective reduction of sound emission is to reach good damping performances in the frequency range dominated by wheel noise. Otherwise if the dampers offer a high reduction of the vibration levels in the frequency range dominated by rail noise, the overall effect on noise reduction is quite low. Wheels with “tuned absorbers” (together with other measures) have been tested too. Basically these absorbers are mass-spring systems added to a structure with the target to remove energy from some eigenfrequencies. Using a system with high damping loss factor,

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

266

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

the absorbers can even work as sources of added damping at frequencies above the massspring resonance frequency. Another alternative to add damping to the wheel is the use of thin layers of highly damped viscoelastic material sandwiches between the wheel and a stiff constraining plate. This solution has already been discussed in Section 4.2. Such devices, when used with treadbraked wheels, can be placed only in the web region because of the high temperatures reached in the wheel parts near the tread. The use of a thicker constraining plate could allow for a higher noise reduction but this kind of modification can be adopted only with disc-braked wheels with a straight web.

8.1.2. Wheel Shape Optimisation The cross-sectional shape of the wheel can highly influence its noise emission. The two most important factors are the thickness of the wheel and its radius. Furthermore, it has already been shown in Section 4.3 that tread thickness and radius values can significantly affect squeal noise emission of resilient wheels.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

8.1.3. Resilient Wheel It has already been shown in Section 4.3 that these wheels are very effective in mitigating rolling noise, if compared to the traditional solid wheels. It is reminded that the rubber blocks have a double effect: they provide the isolation of the web from the tread and also introduce an additional damping to the wheel. The rubber block stiffness governs the uncoupling effect and the damping addition with opposite effects. Thus a trade-off must be accepted. It has been found that the choice about the rubber block stiffness is very important since it has consequences not only on the noise emitted by the wheel but on that due to the rail too ([3], Chapter 6 by C.J.C. Jones and D.J. Thompson). 8.1.4. Reduced Wheel Radiation A further technique, which could be effectively used, is the reduction of the radiated sound due to a particular vibration level. The idea is to reduce the emission decreasing the radiation efficiency of the wheel through holes in the wheel. These holes can create an acoustic shortcircuiting between the front and the back of the wheel web. The effect depends on the size and the spacing of the holes. Another possibility is to mount a shield on the wheel so that the web cannot radiate sound.

8.2. Passive Systems for Tracks 8.2.1. Rail Pad Stiffness The noise radiated by the track is highly linked to the stiffness of the rail pads between the rail and the sleepers. Soft pads allow the rail to be uncoupled from the sleeper. This causes a decrease of the noise from the sleepers but, on the other hand, the rail can vibrate more freely and can generate higher noise levels. Stiff pads decrease the contribution from the rail but increase that from the sleepers. In the end a trade-off should be found case by case.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

267

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

8.2.2. Damping Treatments The use of rail dedicated dampers has the advantage, with respect of changing rail pad stiffness, to damp rail vibrations without affecting sleeper vibration. The only final effect is to increase the decay rates in the rail. The difficulty linked to this kind of methodology is that the rail is already highly damped; this feature makes it difficult to increase the damping factor. Among the most used kinds of damper there are constrained layers and tuned-absorber systems. The use of dampers with soft rail pads could represent the best solution allowing, at the same time, an increase of both rail and sleeper damping. Figure 30 helps in understanding how one of these devices can be set up. The idea is to cover the rail web with rubber layers and to load them using two iron plates. The rubber layers and the plates could be considered, in some way, as a sort of dynamic absorber (a sort of mass-spring system where the plate is the mass and the rubber layer is the spring), whose parameters have to be tuned in order to make the mass-spring resonance frequency (actually the system has several eigenfrequencies) as close as possible to a critical rail eigenfrequency. Therefore when the rail is excited in correspondence of this resonance, the added system starts to vibrate dissipating energy and the rail vibration results much lower than without the absorber.

Figure 30. Rail provided with a dynamic absorber.

One of the parameters which could be tuned to change the mass-spring resonance frequency is the torque with which the bolts are tightened. In fact the rubber layer compression and static stiffness change when the torque is changed. Figure 31 shows the point accelerances (measured by the authors) for different rail (actually a model of a rail with a length of about 50 cm has been used) configurations in order to give an idea of the device effect.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

268

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 31. Performances provided by a dynamic absorber (in two different configurations) in terms of point accelerances.

There are other possible solutions to decrease the sound emission from the rails: optimization of rail shape (the goal is to decrease the radiating area and the radiation ratio through the reduction of rail size), reduction of track mobility and modification of rail support. The discussion about the rail support is worth being paid special attention. Today the use of ballastless tracks (rail is attached by resilient fasteners to a concrete base) is quite popular for high-speed train lines. This kind of rail support is not good from the point of view of noise emission. It is possible to overcome this penalty through continuously supporting the rail by embedding it in a viscoelastic material. In this way it is possible to use a rail with a smaller section decreasing the radiation area. The authors have tested the dynamic behaviour of different rails. Here a piece of rail adopted for city-tram lines in an important Italian city is considered. The same kind of rail has been tested in its usual configuration (not embedded in a viscoelastic material) and embedded in a viscoelastic material (Figure 32). Both the rail pieces were fixed to the ground by means of fasteners usually adopted on lines and their length was higher than 1 m. Figure 33 shows a comparison (in terms of point accelerances) between the usual rail (not embedded in a viscoelastic material) and the embedded one. It is evident that the dynamic rail behaviour completely changes between the two configurations. The damping ratio values dramatically increase thanks to the added layers and the eigenfrequency peaks become much lower. As previously mentioned, this solution could be effective in limiting the noise contribution of the rail.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

269

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 32. Rail of a city-tram line embedded in a viscoelastic material.

Figure 33. Point accelerances for the rail of Figure 32.

Nevertheless it is important to remind a fact. It is stated in literature on disc brake squeal [47] that the addition of damping only to one of the interacting elements involved in squeal phenomenon cannot be a winning solution.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

270

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

Moreover there are many other factors affecting the overall value of noise emission. One of the most important is the roughness of wheels and rails. It has already been quoted that the use of disc brakes helps to have a lower level of roughness than the use of cast-iron block brakes. Of course the most common remedy for roughness is to grind the rail. A change of wheel/rail contact zone could affect noise emission too but this is a very difficult way to work. Another solution could be the use of some sorts of noise barrier placed very close to the rail or mounted on the trains. It is very important to underline that none of the abovementioned passive solutions (or their combination) is able to provide noise reduction higher than 10 dB. Nevertheless the best way is to use some of them in combination to get the best results.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

8.3. Active Systems In the last years more and more attempts have been performed in order to actively reduce wheel and rail noise, especially in terms of squeal noise. Some of them are briefly introduced here and then new original results are given by the authors. Two examples of active control are given here to show that very different possible solutions can be adopted to solve such a problem. In reference [48] a system based on piezo-patches has been adopted on a test rig to control squeal noise. From the experimental evidence and numerical simulations it has been found that the squeal noise can effectively be actively controlled. Nevertheless, since some actuator limits, no high performances have been experimentally got. It has also been found that this active system can be tuned only on one wheel bending mode in order not to destabilize a previously stable mode. A modal approach could be the right solution for this problem since it would allow not to have unwanted interferences between the modes. A further approach (which cannot be considered a real active control, although it involves an action on the rails) is given in reference [49], where squeal noise is mitigated by applying a friction modifier liquid on the rails before the city-tram passage. The change of friction coefficient between the wheels and the rails allows to achieve a significant reduction of squeal noise; furthermore the traction and braking seem to be lowly affected by the liquid application. This solution has been tested on real lines. The authors of this contribution have experienced that also water, applied on rails before tram passage, is able to highly mitigate squeal noise [31]. Concerning the first described approach, it has already been proved that active squeal noise attenuation is possible in terms of control algorithm. The problem is related to the high power needed to mitigate wheel vibration. In the case of reference [48] the problem was related to the piezo-patches, which were not able to provide enough power. This is mainly due to the fact that railway wheels usually have high mass and stiffness. The recent developments in the piezo-actuator field have allowed to improve their performances. Some tests have been recently carried out by the authors of this contribution to test the current performances of piezo-actuators and to find out if it would be possible to exploit them for squeal noise active reduction. Both piezo-patches and piezo-stacks have been tested on the resilient wheel of Figure 3 (with 22 rubber blocks). They have been chosen

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

271

Figure 34. Piezo-bender on the wheel tread.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 35. Scheme of the actions provided by the bender on the structure. E is the applied voltage, F is the obtained force and M is the obtained torque.

Figure 36. Piezo-stack on the wheel tread.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

272

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

Figure 37. Scheme of the actions provided by the stack-actuator on the structure. E is the applied voltage, F is the obtained force and M is the obtained torque.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

considering power requirements, electro-mechanical parameters and cost (thus the best possible actuators have not been chosen but a trade-off has been accepted; the performances of the tested actuators can be considered as an average of the currently available actuators). Figure 34 and Figure 35 show a piezo-bender attached to the wheel tread and Figure 36 and Figure 37 give the same for a piezo-stack. In the case of the stacks, they have been pre-loaded and two bearings have been adopted to prevent shear stresses. The first tests have been carried out on the non-rotating wheel (the wheel was laid down on three rubber supports). Therefore, it has been possible to place the actuators in correspondence of the antinodes of the mode to be controlled. In this case this mode is the flexural one with two nodal diameters (at about 750 Hz, the small mode frequency shift with respect to Section 4.3 is due to the different conditions: suspended free and laid down on supports).

Figure 38. Maximum axial acceleration at resonant frequency (at about 750 Hz), in correspondence of an antinode of the considered mode, induced to the wheel by a pair of piezoelectric benders (up) and by a pair of piezo-stack actuators (down).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

273

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 38 shows the maximum acceleration amplitude (in correspondence of an antinode of the considered mode) obtained by forcing the wheel with two benders and then with two stacks at the resonance frequency. The piezo-amplifier had a full scale of ±500 V for the benders and of ±50 V for the stacks. Nevertheless the best performances are not obtained for the full scale voltage. This is due to current limits on the amplifiers. It is evident that the performances for the stacks are much better, also considering that a less powerful amplifier has been adopted. The need to control high vibrations imposes severe restrictions on the actuator choice and the used actuators would not be able to control squeal vibrations which are characterised by very high levels (of course much higher than 45 m/s2). However, it is known from literature [26, 50, 51] that one of the squeal noise peculiarities is that it is due to a self-excited phenomenon and that it can be reduced, or even suppressed, controlling the first part of the unstable wheel oscillation, preventing its growth to the limit cycle [48, 51]. Considering that previous analyses on disk brake squeal [52] show the need to control vibrations about ten times lower than the limit cycle amplitude, it has been estimated by field measurements that a vibration with an amplitude of about 15 m/s2 has to be controlled in the considered cases if the aim is wheel squeal reduction. Therefore the mentioned test seems to confirm the possibility to effectively exploit stack actuators to attenuate squeal noise. Other tests have then been carried out on the non-rotating wheel before testing the rotating wheel. One of these tests is described here. The disturbance given by two shakers (Figure 39) consists for each shaker in a band-limited white noise (between 600 and 900 Hz) modulated by a low frequency sine wave (4 Hz). The two sine waves used as shaker inputs have a phase shift of 90° between them. This allows to have a non-stationary wave pattern (and thus no fixed mode shapes), which is close to real operating conditions [19] (Section 6).

Figure 39. Test configuration.

The performances of the control system shows that a satisfactory result of -9 dB is obtained for vibrations measured by the accelerometer 1 (Figure 40).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

274

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

After the mentioned laboratory activity, some experimental tests have been carried out on the rotating resilient wheel. The adopted test stand allows the wheel to roll over another wheel. This further wheel (which works as a rail) has a radius which is much greater than that of the railway wheel, in order to limit the effects due to its curvature. Some hydraulic actuators are able to force the railway wheel in lateral and vertical direction and this allows to simulate curve conditions.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 40. Control performances in terms of tread axial acceleration in correspondence of the accelerometer 1 of Figure 39. Solid line for controller switched off and dotted line for controller switched on.

The first performed tests have been carried out simulating curves where squeal emission is limited. Figure 41 gives the effect of the control system (constituted by an adaptive feedback algorithm, four stack actuators and four piezo-accelerometers). Consider, for example, the squealing frequency at about 750 Hz. The noise attenuation given by the control system is about 10 dB. In the present case the wheel vibration in correspondence with the squealing frequency is always lower than 10 m/s2. This situation is quite far from the usual in-line squealing conditions where the wheel vibration is much higher. Furthermore, 10 m/s2 is an acceleration value lower than the limit of 15 m/s2 previously given. Nevertheless, it has been shown that it is possible to significantly excite the wheel under operating conditions and thus to attenuate squeal emission. The next tests will be devoted to verify whether squeal can be attenuated also with higher wheel acceleration levels.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Noise and Vibrations of Railway Wheels

275

Figure 41. Control performance in terms of sound emission. The dashed line indicates controller switched off and the solid line indicates controller switched on2.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

9. Conclusion This contribution has dealt with the main aspects related to railway wheel noise emission. A description of the main emission mechanisms and an overview on the vibro-acoustic behaviour of different kinds of wheels have been given. Particular attention has been devoted to the problem of noise attenuation. Finally, it has been shown that it will soon be possible to attenuate squeal noise by means of piezoactuators.

Acknowledgments The authors are grateful to the European Projects InMAR and HIPERTRACK for having funded this research and to the Research & Development Department of Lucchini Sidermeccanica and to ATM–Azienda Trasporti Milanesi S.p.A. for enabling us to perform a part of the experimental tests presented in this contribution. The authors are also grateful to ERAS GmbH for their collaboration in the tests on squeal noise active reduction.

2

Courtesy of ERAS GmbH (Göttingen, Germany), which has been a partner of Politecnico di Milano within the European Project InMAR.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

276

Alfredo Cigada, Stefano Manzoni, Matteo Redaelli et al.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

References [1] S. Wolf, Potential low frequency ground vibration ( 160 km/h

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Table 1. Optimal track parameters (after Meissonnier, 2000) Attribute Geometry of wheel Rail section Rail pad stiffness Sleeper spacing Ballast – depth Sleeper stiffness

Value 1 in 10,000 wheels giving 250 kN impact force or better UIC 60 80 kN/mm 0.6 m 0.3 m 200 kN/mm

Subgrade stiffness variation

10% or better than mean value

Comments

0.5 mm rail deflection at 20 tonne axle load

The Rational Design Process Introduction The railway track structure should be designed in an appropriate manner to maintain, for a predetermined period, a suitable and uniform track stiffness by withstanding the combined

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

308

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

effects of traffic and climate to the extent that the subgrade is adequately protected and that railway vehicle operating costs, safety and comfort of passenger are kept within acceptable limits. There are a number of design procedures which have been developed by different agencies for such systems. As many of these are based, in part at least, on an empirically founded methodology extrapolating them for the design of high speed lines may not be appropriate (Burrow et al., 2007a). The trend towards using faster trains and heavier axle loads than those originally considered in many existing design procedures, in conjunction with the need to minimise whole life costs, requires a rational approach to design which is based on a thorough understanding of the track system and its influence on subgrade behaviour. A rational approach combines two main processes. In the first the stresses, strains and deflections induced by train loading in the component layers of the substructure are determined (Ullidtz, 2002). The second process consists of using experimental methods to determine allowable stresses, strains and deflections in the various materials which constitute the track substructure. To formulate the design, the induced stresses, strains and deflections in each of the component layers are compared with the allowable determined from experimentation. To determine the induced stresses, strains and deflections the applied traffic loading should be accurately represented and the component layers of the substructure should be characterized to meet the requirements of a theoretical model of the track system. Such a traffic / substructure model allows the stresses, strains and deflections throughout the track system to be predicted. In the second process, it is necessary to identify the most important parameters in the component layers which cause the track support system to deteriorate over time. By setting limits on these parameters, through serviceability requirements for example, the allowable stresses, strains and deflections may be established.

Traffic Characterization The effects of repeated traffic loads and climate reduce the performance of the track over time with a consequent lessening of its ability to carry traffic at design loads and speeds. The performance of the track at any particular time can be related to its condition at that time. The design period is the number of years from the time the track is opened for first use until a terminal condition, however defined, is reached.

Track Condition The condition of the track may be described by its functional and structural condition. The functional condition relates to the ability of track to serve the rail user, whilst the structural condition concerns the track’s ability to carry load and protect the subgrade. Under repeated loading, the track moves laterally and vertically causing deviations in line and level from the desired geometry. As these deviations are generally irregular, ride quality decreases and consequently the loads to which the track is subjected increase, causing increased geometry deterioration. Changes in the ride quality, or functional condition, can be measured by a number of parameters, the most common however concern changes in the vertical and

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

309

horizontal profile of the railway track. Track geometry may change over time due to a number of factors which are related to both the functional and structural property of the track system. For example, line and level deviations over time can be caused solely by poor ballast alignment, or they may be associated with structural problem related to the subgrade. Consequently, in order to identify whether the problem is functional and / or structural in nature and make appropriate decisions regarding maintenance it is necessary also to measure the track deflection or stiffness (Huille and Hunt, 2000).

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Design Period The choice of the optimum design period should be evaluated using a life cycle cost analysis to investigate various design options over a fixed analysis period. For example, one strategy could be to design the track substructure to last for the whole of the analysis period with maintenance carried out periodically to adjust the line and level. An alternative strategy could be adopted whereby several design periods are included within the analysis period with various components being replaced at the end of each design period. The most efficient solution should be determined by evaluating total life cycle costs including user, maintenance and renewal costs. Design Loading The track structure is subject to repeated vertical, lateral and longitudinal forces induced by traffic and the climate. These forces, transferred through the track superstructure, determine the dynamic loading environment that must be supported by the substructure (Selig and Waters, 1994). The forces applied to the track by moving vehicles are larger than the nominal static weights of the trains in question due to dynamic forces induced by variable track, vehicle characteristics and operating conditions. Research suggests that dynamic forces can be considered to result from vehicle speed effects and irregularities in the track or vehicle wheels. The latter are likely to increase over time as the track deteriorates. The dynamic forces may be determined from; field measurements; appropriate models of the vehicle track system, or; empirically founded formulae. These typically are of the following form:

Pd = K d Ps

(7)

where Pd = dynamic force, Kd = dynamic impact factor and Ps = static train load. A useful review of many commonly used impact factors is given by Stewart and O’Rourke (1988). Eisenman (1977) proposes a probabilistic approach, based on an empirical study of actual stresses, which takes into account both vehicle speed and track condition. From the study, Eisenman suggests that stresses induced in the rail are normally distributed (Figure 3) and the mean value is independent of the operating speed, V, but is a function of the track condition, ϕ. For speeds greater than 60 km / h, dynamic forces however were found to be a function of both the vehicle speed and track condition. These two cases can be represented by the following equations:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

310

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

k d = 1 + tϕ

when V < 60 km/h

(8)

⎛ V − 60 ⎞ k d = 1 + tϕ ⎜ 1 + ⎟ 140 ⎠ ⎝

when 60 ≤ V ≤ 200 km/h

(9)

where ϕ , the track condition, has a value of 0.1 for track in very good condition, 0.2 for track in good condition and 0.3 for track in poor condition. t is an integer which takes a value between 1 and 3 depending on the risk associated with the design (see below). The probability of occurrence of a particular stress or load, P is given by

⎛ 1 ⎞ f ( p) = ⎜ ⎟e ⎝ σ 2π ⎠

( P − μ )2

2σ 2

(10)

Where μ is the mean value of stress occurrence and σ is the standard deviation given by

σ = μϕ σ = μϕ ⎛⎜1 +

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

⎝

V − 60 ⎞ ⎟ 140 ⎠

when V < 60 km/h

(11)

when 60 ≤ V ≤ 200 km/h

(12)

Figure 3. Train induced stresses as a function of train speed (after Esveld, 2001).

Eisenman suggests that the design dynamic stress selected should be based on the application to which the design process will be applied. For example, Eisenman recommends

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

311

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

that for a design with safety critical considerations, such as that associated with calculating rail stresses and fastenings, the design dynamic stress should be equal to the mean value, μ, plus three standard deviations (i.e. t = 3). As the dynamic loads in Eisenman’s model are normally distributed about the mean, the dynamic load calculated with a value of t = 3 represents the maximum of all possible loads occurring within 3 standard deviations of the mean i.e. 99.7 % of all possible loads are likely to be less than this value (Figure 3). For calculations of the lateral load and those in the ballast bed, a value of t = 2 is suggested by Eisenman. Where the design concerns less safety critical work, such as the foundations then t = 1 is appropriate. As an example, consider a foundation design problem (t = 1) for a track with a design speed of 200 km / h and assuming the track condition will never fall below what is considered to be a good condition (ϕ = 0.2), then using equation 3 the dynamic amplification factor, K, is 1.40 or an increment of 40 % should be added to the mean (static) load. Equations 3 has been modified for use by the German railway authorities (Brandl, 2001) as follows:

⎛ 0.5(V − 60 ) ⎞ k = 1 + tϕ ⎜ 1 + ⎟ 190 ⎝ ⎠

for passenger trains when 60 ≤ V ≤ 300 km/h

(13)

⎛ 0.5(V − 60 ) ⎞ k = 1 + tϕ ⎜ 1 + ⎟ 80 ⎝ ⎠

for freight trains when 60 ≤ V ≤ 140 km/h

(14)

and ϕ = 0.15 for high speed lines and other main lines, ϕ = 0.20 for secondary lines and ϕ = 0.25 for other tracks. It is usual to consider the two axles of a leading bogie and the two of a trailing bogie as a single load repetition in order to calculate the stresses and deformations in the subgrade (Grabe and Clayton, 2003; Li and Selig, 1998a; Stewart and O’Rourke, 1988). This four axle loading regime has been shown, from field measurements, to be equivalent to a single load pulse at depths below approximately 0.6 m (Stewart and O’Rourke, 1988). For design purposes it is convenient to use a single loading configuration. However, as the loads applied to the track over its design period are likely to vary in magnitude, it is necessary to convert the number of applications of all loads to an equivalent number of repetitions of the design load. The equivalent number is that number of loads which will cause the same amount of track damage and can be calculated using equations for track damage which are typically of the form (Esveld, 2001 and Li and Selig, 1998): c

D = CPd N b

(15)

where D is damage, C is a constant for vehicles travelling at a particular speed, Pd is the dynamic load, N is the number of load repetitions and c and b are constant for particular track constructions. Then the equivalent number of repetitions, Ne, of the design load, Pdd, to achieve the same amount of damage as caused by Ni repetitions of any load Pdi, can be calculated as follows:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

312

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al. c

b

c

D = KPdd N e = KPdi N i

b

(16)

whence c b

⎛P ⎞ N e = ⎜⎜ di ⎟⎟ N i ⎝ Pdd ⎠

(17)

Where a four axle loading configuration is used, the above calculations involve the superimposition of loads, or the use of modelling to determine the stresses induced by the loading configuration at the depth of interest. This is usually the sub-ballast / subgrade interface. To quantify the spectra of likely dynamic loads a probabilistic approach may be used. Stewart and O’Rourke (1988) describe a method which considers an unequal load distribution on the four axles. In their method they assume that three of the axles carry equal loads whilst the fourth one carries a variable load. The magnitude and probability of occurrence of the latter are determined from field data describing the distribution of all dynamic loads.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Modelling the Track Support System Analytical models of the track superstructure and substructure are used to determine the effect of traffic loads on the stresses, strains and deformations in the system. These can then be compared with allowable stresses, strains and deformations of the various components in the track support system to formulate a design. With the advent of the personal computer design procedures have incorporated models based on layered elastic theory, the finite element (FE), the finite difference, boundary element and distinct element methods. These model individual components of the superstructure and substructure and are able to consider non-linear characteristics including plastic, viscous and viscoelastic deformations and strain rates which are non-linear functions of the stress level. Whilst each technique may have advantages for specific applications, arguably the FE method offers the widest and most robust range of computational capabilities (Schwartz, 2004). Further information on the types of models available and their relative merits may be found elsewhere (see for example Blair and Chan, 2006; Selig and Waters, 1994).

Track Foundation Properties Required for Structural Analysis Analytical models require each layer of the substructure to be characterised in terms of elastic parameters. Usually two parameters, the resilient modulus and Poisson’s ratio, are used. The resilient modulus, defined as the quotient of the deviator stress by the resilient strain in the direction of the major principal stress, can be determined directly from laboratory tests, or from an analysis of the response of measured in situ parameters. Poisson’s ratio is usually estimated. Granular Materials Under repeated loading conditions, the behaviour of granular materials is nonlinear and stress-dependent (Gomes Correia, 2004; Selig and Waters, 1994). Initially, for each cycle of loading some plastic strain occurs whilst the magnitude of the plastic strain decreases as the

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

313

number of loading cycles increases. Eventually, and if the stresses levels are moderate, after a number of loading cycles the resilient strain becomes constant and the material behaves elastically (see below). The stress level is the primary factor affecting the resilient modulus which has been shown to increase appreciably with increasing confining stress and slightly with increasing repeated deviator stress, provided that shear failure is not approached. Four main types of non-linear models are widely used to describe non-linear behaviour. These are the K- θ, the modified K- θ, the Boyce, and the orthotropic Boyce models (Gomes Correia, 2004; Lekarp et al., 2000). As modelling of resilient behaviour is complex, the simple and widely used, K- θ model was proposed in the 1960’s to describe the results of cyclic load triaxial tests carried out with a constant confining pressure (Gomes Correia, 2004; Brown, 1996; Brown and Pell, 1967). In this model the resilient modulus, Er, is given by

Er = k1θ1

k2

(18)

where k1 and k2 are material constants determined from experimentation and θ1 = 3 p′ where

p′ is the mean normal effective stress. Poisson’s ratio, ν, is constant and usually taken as 0.3. The inaccuracy of this model in computing stress conditions in a granular layer or sand subgrade is widely recognised (Gomes Correia, 2004; Brown, 1996). However, it can usefully be used to model a granular layer when effects in the layers above or below are required. Gomes Correia et al. (1999) also note that the accuracy of the model could be improved if a stress dependent Poisson’s ratio is used. Time (secs) 0.01

0.02

0.03

0.04

0.05

0.06 0.1

-0.1 5 -0.3

15

-0.5

-0.7

-0.9 20

25

-1.1

D e fle c tio n (m m )

10

F W D lo a d (k N )

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

0.00 0

FWD Load (kN) d0 d300 d1000 d1500

-1.3

-1.5 30 -1.7

35

-1.9

Figure 4. FWD load and sensor deflection time histories.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

314

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Burrow et al. (2007a) used the K- θ model to represent ballast and sub-ballast material behaviour in a finite element model of an existing section of railway track. In their study they determined the parameters k1 and k2 using a procedure based on matching closely actual deflections in the field obtained with the falling weight deflectometer (FWD) device with deflections obtained from the model (Figure 4). The study found that for the ballast material k1 = 175 and k2 = 0.5 (with ν = 0.2) and for highly contaminated ballast slurry k1 = 35 and k2 = 0.5 (with ν = 0.49). A modified version of the K- θ model was proposed by Uzan et al. (1992) in which the resilient modulus is given as a function of both the mean stress p and the deviator stress q.

Er = k1θ1 2 θ 2 k

Where θ2 =

k3

(19)

q and the deviator stress, q = σ1-σ3 and pa is a reference stress (such as pa

atmospheric pressure). k1, k2 and k3 are material properties obtained from experimentation. Although the modified model takes into account the effect of the shear stress (q), Poisson’s ratio is also assumed to be constant and consequently its use may lead to the inaccuracies as described above. The Boyce (1980) non-linear elastic model was developed using cyclic triaxial test apparatus that was able to vary the confining pressure. Consequently, it is able to take into account the effect of stress paths. The model is expressed in terms of the bulk modulus, K, and the shear modulus, G, as follows:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

1− n

⎛ p ⎞ ⎜⎜ ⎟⎟ p K= ⎝ a⎠ 2 1 β ⎛q⎞ − ⎜ ⎟ k a k a ⎜⎝ p ⎟⎠

(20)

1− n

⎛ p ⎞ G = Ga ⎜⎜ ⎟⎟ ⎝ pa ⎠ where

β=

(21)

(1 − n)ka 6Ga

ka, Ga and n are constants. With this model using elasticity theory, Young’s modulus, E, and Poisson’s ratio, ν, may be expressed as follows:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

315

1− n

⎛ p ⎞ 9Ga ⎜⎜ ⎟⎟ ⎝ pa ⎠ E= 2 ⎛ Ga ⎞ ⎡ ⎛q⎞ ⎤ ⎜ ⎟ 3+⎜ ⎟ ⎢1 − β ⎜⎜ p ⎟⎟ ⎥ ⎝ ⎠ ⎥⎦ ⎝ K a ⎠ ⎢⎣ 2 ⎛q⎞ ⎤ 3 ⎛ Ga ⎞ ⎡ − ⎜ ⎟ ⎢1 − β ⎜⎜ ⎟⎟ ⎥ 2 ⎜⎝ k a ⎟⎠ ⎢ ⎝ p ⎠ ⎥⎦ ⎣ ν= 2 ⎛ Ga ⎞ ⎡ ⎛q⎞ ⎤ ⎟⎟ ⎢1 − β ⎜⎜ ⎟⎟ ⎥ 3 + ⎜⎜ ⎝ p ⎠ ⎥⎦ ⎝ K a ⎠ ⎢⎣

(22)

(23)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The Boyce model represents the behaviour of materials with reasonable accuracy as it can describe the effects of the mean normal stress p and also of the stress ratio q/p (Gomes Correia, 2004). However, it has been found that the Boyce model does not accurately predict values of the shear strain, εq for low values of q/p, leading to unrealistic values of ka, Ga and n. Consequently, Hornych et al. (1998) proposed an orthotropic version of the model by introducing a parameter γ, to scale the principal stress σ1. Using their model the resilient volumetric (εv) and shear (εq) strains can be represented by the following (Gomes Correia, 2004):

p *n ε v = n−1 pa

2 p *n εq = n −1 3 pa

2 ⎡γ + 2 n − 1 ⎛ q * ⎞ γ − 1 ⎛ q * ⎞⎤ ⎜ * ⎟⎥ ⎢ + (γ + 2)⎜⎜ * ⎟⎟ + ⎜ p ⎟⎥ p G 3 ⎢⎣ 3K a 18Ga a ⎝ ⎠ ⎝ ⎠⎦

2 * ⎡γ − 1 n − 1 ⎛ q *⎞ 2γ + 1 ⎛ q ⎞⎤ ⎜ ⎟ ⎜ ⎟⎥ ⎢ (γ − 1)⎜ * ⎟ + + ⎜ p * ⎟⎥ p G 6 ⎢ 3K a 18Ga a ⎝ ⎝ ⎠ ⎠⎦ ⎣

and

p* =

(24)

(25)

γσ 1 + 2σ 3 3

q * = γσ 1 − σ 3 Pappin et al. (1992) demonstrated that the resilient response modelled for dry granular material is applicable to saturated and partially saturated conditions, provided that the principle of effective stress is observed (Brown, 1996). Gomes Correia (2004) describes a number of models developed using such an approach. However, in practice the determination of the effective stress state in a granular layer may not be straightforward.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

316

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

As mentioned above, the accuracy of the constitutive relationships may be improved if stress dependent Poisson’s ratios are used and for repeated load triaxial compression tests indicate that Poisson’s ratio is strongly correlated to the ratio of principal stresses (σ1/σ3).

Fine-Grained Soils In the case of fine-grained soils, the deviator stress has a primary influence on the resilient modulus which decreases non-linearly with increasing applied deviator stress, when all other factors are kept constant, and therefore constitutive models are primarily established between the resilient modulus and the deviator stress (Fleming et al., 2003). Li and Selig (1994) categorize these models into five different types as follows: Bilinear model The bilinear model was proposed by Thompson and Robnett (1976) and takes the following form:

Er = k1 + k 2 q

when q < qi

(26)

Er = k 3 + k 4 q

when q > qi

(27)

where qi is the deviator stress at which the gradient of the resilient modulus changes and k1, k2, k3 and k4 are constants which depend on the soil type and its physical state.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Power models Mossazadeh and Witczak (1981) formulated a power model to represent the behaviour of three fine-grained soils as follows:

Er = kq n

(28)

Where k and n are constants which depend on soil type and physical state. Brown et al. (1975) suggested an effective stress version of the model to represent the behaviour of silty clays, in which the resilient modulus is also a function of, p'0, the initial (geostatic) mean effective normal compressive stress. Their modified model is given by:

⎛ p′ ⎞ Er = k ⎜⎜ 0 ⎟⎟ ⎝ q ⎠

n

(29)

In the work reported by Burrow et al. (2007a) mentioned earlier which investigated a means of calibrating a FE model by using field deflections obtained from a FWD, Brown’s model was used to represent the performance of a silty clay subgrade. Using the FE model the parameters k and n in Equation 29 were determined for various subgrade layers beneath the ballast and are given in Table 2 below.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

317

Table 2. Back analysed parameters (after Burrow et al., 2007a) Subgrade 1 (Firm CLAY)

Base modulus, K (MPa) Exponent, n Poisson’s ratio, ν Layer thickness (m) Base modulus, K (MPa) Exponent, n Poisson’s ratio, ν Layer thickness (m) Base modulus, K (MPa) Exponent, n Poisson’s ratio, ν Layer thickness (m)

Subgrade 2 (Soft to Firm CLAY)

Subgrade 3 (Very soft to soft CLAY)

123 1.1 0.49 0.50 30* 1.1 0.49 1.50 6* 1.1 0.49 0.25

Semilog model The semilog model was suggested by Fredlund et al. (1977) for a moraine glacial till as follows:

M r = 10 ( k −nq )

(30)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Raymond et al. (1979) used a similar model to successfully represent the behaviour of Leda clay.

Hyperbolic model Drumm et al. (1990) proposed the hyperbolic model as a match for data for fine-grained Tennessee soils and it takes the following form:

Er =

K + nq q

(31)

Octahedral model The Octahedral model in which the resilient modulus is a function of the octahedral normal and shear stresses, σOCT and τOCT respectively, was proposed by Shackel (1973):

Er = k

n σ OCT m τ OCT

(32)

Li and Selig (1994) carried out an investigation of the first four models described above, which involved determining how well each model fitted to data available in the literature. The Octahedral model was not considered as it was regarded as being too difficult to apply. From their analysis they found that all of the models may be considered to fit the data considered by choosing suitable parameters (such as k and n). However, they found that the bilinear model

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

318

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

gave the best representation, followed by the power, semilog and hyperbolic models respectively.

Composites As described above, large amount of research effort has focused on determining the properties of the individual constituent layers of the track substructure in isolation. However, the interaction of these layers is also important. For example, in the case of a conventional granular trackbed layer which is supported by a softer subgrade, the deflections will be partially controlled by the load spreading ability of the granular material which controls the level of stress transmitted to the subgrade. However, the reaction of subgrade to the stress transmitted will influence the amount of load spreading that can occur. Thus the layer interaction affects the stress distribution which in turn affects the total elastic and plastic strains that are developed within each layer and hence their response to those stresses and vice versa (Fleming et al., 2003). In many routine highway design methods an attempt is made to take into account the non-linear behaviour of granular materials values by using empirical relationships between the modulus of the layer in question and that of the underlying one (Gomes Correia, 2004; Loizos, 2004). Such relationships are typically of the form:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

En = kEn+1

(33)

where En+1 is the resilient modulus of the underlying layer and k is typically between 2 and 4. Whilst these types of relationships are simple and may be easily incorporated in any railway track design procedure they do not take into account the influence of the water content and quality of the constituent material (Gomes Correia, 2004; Loizos, 2004). Loizos (2004) describes research to provide more accurate relationships, to those described by Equation 33 above, which take into account the modulii and thicknesses of the layers above and below those in question.

Field Work and Laboratory Testing For the models described above appropriate procedures should be used to determine the various model parameters. Most models of the type described above have been developed using cyclic load triaxial test apparatus (Gomes Correia, 2004). However, under the passage of a moving wheel load an element of material in the track substructure is subject to a complex regime consisting of vertical, horizontal and shear stresses (see Figure 5). This regime is more closely represented under laboratory conditions using the Hollow Cylinder Apparatus (HCA) which, unlike a conventional triaxial apparatus, allows the normal and shear stresses to be controlled to simulate more closely the in-situ regime. Whilst permanent deformations determined using conventional techniques, when compared to using the HCA, may be significantly underestimated (Gräbe and Clayton, 2003), research by Chan (1990) demonstrated that resilient strains are unaffected by the phenomenon and that the principal planes of strain remain coincident with those of stress. The work demonstrates that the relatively simple resilient modulus models derived from triaxial tests, as described above, rather than more complex apparatus can be used for structural design (Brown, 1996).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

319

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5. Stresses induced under a moving wheel load (after Chan and Brown, 1994).

Laboratory testing of small elements, however, raises questions concerning the preparation of reliable specimens and whether they are representative of in-situ conditions (Brown, 1996). Consequently, field testing, though more expensive than laboratory testing, plays an important part in the design process as it allows material properties to be determined under representative conditions. To this end, dynamic deflection tests such as single point load tests, deflection basin tests or multiple axle vehicle load tests can be used. An example of a methodology which may be employed to determine the resilient moduli of layers beneath a railway track using FWD device together with a dynamic FE model of the device has been mentioned above and useful summaries of appropriate tests may be found in Brough et al. (2003) and Selig and Waters (1994).

Design Criteria Two considerations should be considered in design, appropriate track stiffness and the protection of the subgrade. As mentioned previously, it is important to ensure that the stiffness of the track is within acceptable limits. The ability of the track substructure to deform elastically (i.e. in a resilient manner) facilitates the transfer of train induced loads from the wheel, via the rail to sleepers through the ballast, sub-ballast and finally into the subgrade. The entire system should possess a stiffness which limits rail displacements on the one hand, but is not so stiff to cause load concentrations to occur. The former can cause

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

320

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

damage to subgrade, whilst the latter can damage the track superstructure. In conventional railway track, the optimum design of such a system should involve a gradual decrease in stiffness from the ballast, through the sub-ballast to the subgrade. The achievable and uniformity of stiffness of each layer and thus of the overall system is a function of material properties, the thickness of the various layers and the quality of construction. The design criteria related to track stiffness, k, can be expressed as follows:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

k all ≤ k p ≤ k aul

(34)

where kall and kaul are lower and upper acceptable limits of track stiffness respectively. Additionally, it is important to ensure that the track substructure is appropriately designed as it becomes progressively weakened through the cumulative effect of traffic induced repetitions of stresses and strains. As ballast lends itself to maintenance the main objective of the design procedure is to protect the subgrade. In the subgrade, the primary modes of trafficinduced deterioration are subgrade attrition by the ballast, progressive shear failure, massive shear failure and an excessive rate of settlement through the accumulation of plastic strain (Selig and Waters, 1994). These modes are mostly associated with fine grained soils. Subgrade attrition occurs as a result of relative movement of ballast and subgrade at the ballast-subgrade interface. The usual method of preventing its occurrence is to place a layer of sand of appropriate thickness directly between the ballast and subgrade. Progressive shear failure occurs where cyclic stresses in the subgrade are high enough to cause it to be sheared and remoulded and overstressed soil is squeezed sideways from beneath the track and upwards to cause a form of bearing capacity failure. It is less of a problem with coarse grained materials which possess high values of internal friction such that the increase in shear strength associated with applied normal stress exceeds the increase in associated shear stress (Li and Selig, 1998a). Massive shear failure can occur due to the weight from the train, track superstructure and unbalanced portions of the substructure. However, as progressive failure usually occurs at stress levels below that causing massive failure it governs performance and therefore design. Massive shear failure is only likely to be problematic when, for example, heavy rainfall or flooding cause the subgrade to have an unusually high water content. An excessive rate of settlement through plastic deformations may cause a ballast pocket to form as a result of the vertical component of progressive shear deformation, deformations caused by progressive compaction, or consolidation of the subgrade layer (Li and Selig, 1996; Selig and Waters, 1994). Consequently, the design problem as far as the subgrade is concerned, provided a sufficiently thick sand blanket is included in the construction, can be regarded as identifying, and putting limiting values on, the stresses, strains and deflections that are the major causes of an of excessive rate of settlement and progressive shear failure. Li and Selig (1998a) suggest that subgrade performance is influenced by (i) vertical strain at top of subgrade and (ii) vertical plastic deformation. Both of these factors are controlled by levels of shear stress (or deviator stress). In the case of (i) the shear stress of interest is that at the top of the subgrade, whilst for deformation the shear stresses throughout the subgrade are important. Accordingly, the design objective is to ensure that the cumulative effects of the

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

321

repetitions of shear stresses throughout the subgrade do not cause excessive progressive shear failure or excessive rates of settlement to occur before the end of the design period. The design criteria to protect the subgrade can be expressed as follows (Li and Selig, 1998a):

ε p ≤ ε pa (to prevent progressive shear failure)

(35)

ρ ≤ ρ a (to prevent excessive plastic deformation)

(36)

and

Noting that D

ρ = ∫ ε p ds

(37)

o

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

where εp and εpa are the actual total cumulative and allowable plastic strains, ρ and ρa are the actual and allowable plastic deformations, respectively, at the subgrade surface over the design period. D is the depth of the subgrade.

Material Performance under Repeated Loading In an analytical design process measures of material performance are used to determine appropriate limits to stresses, strains or deflections. These, in conjunction with stresses, strains and deflections predicted to occur over the lifetime of the railway track are used to formulate the design. Ideally, such measures should be determined under conditions which closely match the in-situ regime. If tests are conducted in the laboratory then these should enable the effect of the rotation of principle stresses to be taken into account. Whilst this effect has been shown to be insignificant in determining resilient properties (see above), permanent deformations may be greatly underestimated when principal stress rotation is ignored. Gräbe and Clayton (2003), for example, found that the axial plastic strain may be underestimated by as much as between 1.6 and 3.2 times depending on the clay content of the material. Granular Materials For granular materials, permanent deformation is a function of both the cyclic deviator stress and the confining pressure and it has been shown to increase with the logarithm of the number of cycles of applied stress. Accordingly, performance models for granular materials take the following form:

⎛1⎞ ε c = a⎜ ⎟ ⎝N⎠

b

(38)

where εc is the permissible vertical compressive strain at the point of interest (usually a layer interface) for N load repetitions; a and b are coefficients determined from experimentation. Similarly, for ballast the vertical compressive strain is also often expressed as a function of the permanent strain, ε1, after the first loading cycle and number of cycles, N as follows:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

322

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

ε c = ε1 (1 + C log N )

(39)

where C is a material constant and is typically between 0.2 and 0.4 (Selig and Waters, 1994). Selig and Waters (1994) report laboratory and field trials undertaken by the Office for Research and Experiments (ORE), of the International Union of Railways to characterise ballast settlement. The ORE’s results suggested that the settlement is a function of the initial porosity of the sample, n, the deviator stress and the number of load cycles as follows:

ε c = 0.082(100n − 88.2)(σ 1 − σ 3 ) 2 (1 + 0.2 log N )

(40)

Paute et al. (1993) suggest a model to predict permanent axial strain, εp in granular materials as a function of N as follows:

ε p = ε p (100) + ε *p ( N )

(41)

⎡ ⎛ N ⎞−β ⎤ ε ( N ) = A⎢1 − ⎜ ⎟ ⎥ ⎢⎣ ⎝ 100 ⎠ ⎥⎦

(42)

and * p

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

where εp(100) = accumulated permanent axial strain during the first 100 loading cycles ε∗p(N) = additional permanent axial strain for N > 100 A and Β are regression parameters. Further analysis of this model by Lekarp et al. (1996) suggested that the model is generally successful in predicting permanent strain. A limiting deviator stress, known as the shakedown limit, has been found to exist below which the accumulation of plastic strain is stable and above which the accumulation increases rapidly. For example, Brown (1996) describes data from the literature which demonstrate that insignificant plastic strains develop if the peak repeated stress ratio is always less than 70% of that required to cause static failure. However, research continues to define the boundary between stable and unstable behaviour (Werkmeister, 2001).

Fine-Grained Materials For fine-grained materials permanent deformation has been shown to be a function of the number of loading cycles, soil stress history and drainage conditions. In addition, as with granular materials, it is recognised that a critical level of repeated deviator stress, known as the threshold stress, exists above which the rate of accumulation of deformation increases rapidly. At deviator stress levels below the threshold stress deformation has been found to increase with the logarithm of the number of cycles whilst at deviator stress levels above the threshold the rate of accumulation of deformation increases exponentially. This behaviour has been found to be related to the material stress history and water content, and thus shear strength (Fleming et al., 2003). Brown (1996) suggested that the quotient of the applied shear stress by the soil’s shear strength is the principal factor influencing permanent deformation. From experiments on an over-consolidated silty clay, Brown found that the threshold stress

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

323

was at a deviator stress 1.3 times the value of static yield over the range of initial effective stresses studied. A number of models have been developed to predict cumulative plastic strain under repeated loading. A commonly used one is a power model of the form (Li and Selig, 1996; Monismith et al., 1975):

ε p = AN b

(43)

where εp is the percentage cumulative plastic strain, N is the number of repeated load applications; and A and b are two parameters related to the stress state and material properties. In order to take into account soil physical state and type Li and Selig (1996) proposed a modified version of Equation 43 as follows:

ε p = a(

σd m b ) N σs

(44)

Where a, m and b are material parameters, σd is the deviator stress and σs is the soil static strength.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Track Stiffness Variability Despite careful design and construction procedures a railway track superstructure is often built on a non-homogeneous substructure leading to significant changes of track stiffness within short distances. These may be due to the presence of pile decks, embankments, bridges, and transition zones between ballasted and slab track. Also, the track stiffness can change very quickly at switches and turnouts, especially at crossings (frogs), insulation joints and where there are hanging sleepers. An example of a 3 km section of railway track is shown in Figure 6. In the upper part of the figure the stiffness variation along the track is shown, whilst the lower part shows the associated settlement over a two year period. The change of stiffness between the embankment and bridge is approximately four fold, from about 40 kN/mm (at km 11.40) to 160 kN/mm (at km 11.45). From Figure 6, also it can be seen that large variations in track stiffness correlate with large track settlements at 9.4 km (due to a bridge) and at 11.4 km and 11.7 km (due to an embankment and light-weight fill respectively). Figure 7 shows the track stiffness variation along a 25m section of a track. It can be seen from the figure that the track stiffness varies with a (spatial) frequency corresponding to the sleeper distance (that here is approximately 0.65m). i.e. the track is stiffer when measured above a sleeper than between two sleepers. It is also interesting to note that the track stiffness is much lower in the vicinity of the three sleepers near km 149.807. This is most probably due to a lack of ballast support at the location (i.e. the sleepers are hanging). A possible cause is considered to be the presence of an insulated joint which has induced irregularities in the wheel/rail contact force causing increased loading on the sleepers and associated vibrations leading to deterioration of the ballast bed below the sleepers. Thus, the associated track

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

324

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

1211.5 11 10.5 10

Position along the track [km]

9.5

0-

1-

2-

3-

4-

-9

-000407 -001109 -010320 -011024

Longitudinal level, rms value over 20 meters, 2000 2001

5-

9.5

0 --

20-

60-

80-

100-

120-

--

140-

40-

Longitudinal level [mm]

160-

9

Bridge

10

--------Stiffness v = 20 km/h, f = 5.7 Hz -- reinforcement No --- weight fill Light Pile-deck, bridge --

10.5

11

-------along the track [km] Position

11.5

Continuous stiffness measurements, WEst coast line in Sweden, east track w37 2001 20 km/h 5.7 Hz

12-

deterioration is believed to have started at a point of discontinuous track stiffness at the rail joint.

-

Stiffness [kN/mm]

-260 -240 Track stiffness [ kN/mm ]

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 6. Left figure: track stiffness (axle load divided by track deflection) along railway track as measured by the Banverket track stiffness measurement trolley. Right figure: four measurements over a two year period of the longitudinal level of the track (positive downwards, meaning that a large peak in the curve indicates a local settlement of the track).

-220 -200 -180 -160 -140 - 149.79

-149.795

-149.8 - 149.805 Position along the track [km]

-149.81

149.815 -

Figure 7. Local stiffness variation along railway track. Stiffness variation due to sleeper passages can be seen, as well as a sharp reduction in of track stiffness for three sleepers at 149.807 km. Measured by the Banverket track stiffness measurement car (Berggren, 2006).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

325

Previous research A great deal of research has been conducted on track stiffness variation, it causes and possible means of mitigation. The following section presents a short review of some pertinent research focusing on the following aspects:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

o o o o

Random track stiffness variation The effect of rail joints Switches and turnouts Reducing sources of track stiffness variation

Random Track Stiffness As discussed previously, track stiffness varies in a random manner spatially along the railway track. To investigate contributing factors, Mahmoud and Eltawil (1992) modelled the track as a beam resting on elastic supports and found that the response of the beam is highly dependent upon the modulus of subgrade reaction (i.e. on track stiffness). Naprstek and Fryba (1995) carried out similar analyses using a beam resting on a Winkler foundation; the stiffness of which was a random function of the length coordinate. Oscarsson (2001, 2002 and 2003) investigated the influence of stochastic properties of the track structure and to obtain sufficient statistical information from the track structures, full-scale in-field measurements and laboratory measurements were carried out. The rail pad stiffness, the ballast stiffness, the dynamic ballast-subgrade mass (a discretized equivalent mass), and the spacing between sleepers were all assumed to be random variables. The influence of scatter on the maximum contact force between the rail and the wheel, the maximum magnitude of the vertical wheelset acceleration, and the maximum sleeper displacement were studied and quantified in terms of probabilities and standard deviations. Andersen and Nielsen (2003) used a simple track structure with randomly varying support stiffness. In their analysis the vertical support stiffness was assumed to be a stochastic homogeneous field consisting of small random variations around a deterministic mean value. Response spectra were obtained and compared with from numerical solutions obtained from finite element simulations. Wu and Thompson (2000) treated the sleeper spacing and ballast stiffness as random variables and explored their effects on the rail vibration. It was shown that the pinned-pinned resonance phenomenon may be suppressed by the random sleeper spacing, but a randomly varying foundation had no significant effect on the average noise generated by the track. The effects of varying geometry and foundation stiffness are particularly significant when a train moves onto a bridge abutment. To minimise the rate of track settlement growth Hunt (1997) suggested that in the vicinity of bridge abutments the track should have carefully prepared variations in foundation stiffness. Li and Davis (2005) state that remedies intended to strengthen the subgrade between the bridge should be designed to produce consistent and acceptable track stiffness between the bridge and the approach in order to be effective. Nordborg (1998) found that in comparison with surface roughnesses the track support irregularities may be a significant excitation mechanism up to 100 Hz and vibration levels increase with train speed.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

326

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Rail Joints The vertical bending stiffness of a rail joint is generally much lower than that of the rail and so a passing wheel generates larger deflections in the joint region than elsewhere, leading to increased wheel forces and accelerated track deterioration. In order to investigate the effects of joints on track stiffness variation, Koro et al. (2004) used a discretely supported Timoshenko beam and finite elements to predict the impulsive wheel-track contact force excited by the wheel passage on the rail joint. Different train speeds and gap sizes at the joint were simulated in the study. They found that rail joints contribute greatly to track deterioration, the settlement of ballast and the failure of other track components. The research was continued by Suzuki et al. (2005) who investigated measures to reduce ballast settlement, finding that in many cases soft rail pads could be effective. Rolling contact fatigue and plastic deformations at insulated rail joints were the subject of research carried out by Kabo et al. (2006). Switches and Turnouts A switch contains several irregularities both in stiffness and in inertia. For example, the bending stiffness of the switch rail differs from that of the stock rail, the lengths and distances between sleepers vary, the crossing (the frog) is stiffer in bending and has a larger mass than the surrounding rails. Andersson and Dahlberg (1998 and 2000) investigated, by use of a numerical model, the load impact at the crossing nose when a wheel moves (at the frog) from the wing rail to the nose. It was found that the severity of the load impact depends on variations of track stiffness, mass distribution, and geometric irregularities at the crossing. Zarembski et al. (2001) and Zarembski and Palse (2003) performed theoretical formulation, analytical studies, and field tests and concluded that the impact load at a crossing could be virtually eliminated by an appropriate transition. Zhu (2006) studied the effect of varying stiffness under the switch rail of a high-speed turnout. Results from his work showed that the elasticity under the switch rail could improve effectively the vertical wheel-rail interaction dynamics when the train passes from the stock rail to the switch rail. Kassa (2007) used mathematical models to simulate the dynamic trainturnout interaction and compared the results with field measurements. Reducing Sources of Track Stiffness Variation Elastomeric products, such as rail-pads, under-sleeper pads (USP), and sub-ballast mats (SBM), and geogrid (or geotextile) reinforcements, can be used to construct a suitable transition zone with the desired variation of stiffness and geometry. Track settlement in the transition zone has been studied numerically by Guiyu et al. (2004), and the influence of tensile-reinforcements on track settlement was investigated by Monley and Wu (1993). Fullscale simulation of geogrid reinforcement for railway ballast was performed by Brown et al. (2007) using a specially developed test rig. In Johansson et al. (2006) the influence of undersleeper pads on dynamic train–track interaction was investigated. They used two numerical models, valid for different frequency intervals, to study wheel/rail contact forces, rail bending moments, rail vibrations (displacements, velocities, and accelerations), sleeper vibrations, and loads on sleepers. Frequency-dependent material properties of rail pads, USP, and ballast/substructure were modelled using viscoelastic spring-dampers that were calibrated with respect to measured data. It was found that USP influence dynamic train/track interaction mainly in the frequency range 0 – 250 Hz. Loy (2006) investigated the use of

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

327

under-sleeper pads to reduce the variation in the static rail deflection in turnouts. He found that by putting sleeper pads of specific stiffness in different sections of a turnout, the track stiffness could be adjusted so that vertical rail deflections could be smoothed effectively. In Anon. (2006) various track transition designs were reviewed and analysed and a number of techniques were proposed to improve track performance by providing a transition to smooth the stiffness interface between dissimilar track types.

Modelling Dynamic Interaction between Train and Track This section describes some of the recent work carried out at the Mechanical Engineering department at the University of Linköping, Sweden to investigate variations in track stiffness using a model of the dynamic interaction of trains and track (Witt, 2008; Dahlberg, 2006; Lundqvist, 2005; Lundqvist and Dahlberg, 2005). Three studies are described, one to investigate optimum values of track stiffness in a transition between a soft and stiff section of track; another examines the use of under sleeper pads to achieve optimal stiffness for a track whose stiffness is randomly varying. The third considers hanging sleepers. For the three cases the objective was to minimise the dynamic component of the contact force between the wheel and rail. The model

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Train and Track Model The model consists of a finite element track model comprised of 3-D fully integrated solid elements as shown in Figure 8 with the material properties and other parameters as given in Table 4. The track model is composed of one rail (symmetry with respect to the centre line of the track is assumed), rail pads, sleepers, under sleeper pads, and the ballast/substructure bed. The rail was modelled as a standard UIC60 rail, and the rail pads as an elastic material of defined stiffness. Table 4. Components and properties of materials used in the model Component Vehicle Wheel and ½ axle Wheel Car body Superstructure Rail Rail pads Sleeper Substructure (Ballast + subgrade)

Property

Value

Weight Speed Weight

7.358 kN 90 m/s 100 kN

Type Stiffness Mass Spacing

UIC 60 300 kN/mm 125 kg (1/2 sleeper) 0.6 m 100 MPa (stiff section) 30 MPa (soft section) 0.1 2, 500 kg/m3 1m

Modulus of elasticity Poisson’s ratio Density Depth

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

328

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Figure 8. Train/track model consisting of rigid wheel, rail, rail pads, rigid sleepers, under sleeper pads below the ten central sleepers (not shown in the figure), and ballast/substructure. Symmetry with respect to the centre line of the track is assumed.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

A moving wheelset is applied to the track model to simulate the load from one axle of a train. The wheelset is modelled as a rigid body moving at speed, v, with a constant load representing the weight of the car body. Inertia from the un-sprung mass (i.e., from the wheel and the axle) is taken into account by including the wheel mass and half of the axle mass. To avoid wave reflections at the boundaries of the limited model, non-reflecting boundary conditions have been used. These prevent artificial stress wave reflections generated at the boundaries from being reflected back into the model and adversely affecting the analysis. The non-reflecting boundary conditions absorb the shear and pressure waves so that no reflections will occur at the boundaries, but still allow bending waves in the rail to be reflected.

Train/Track Interaction In the FE-program used in the study the contact force between two contacting bodies of the structure (for example between wheel and rail or between sleeper/USP and ballast) was calculated using a penalty method described by Belytschko et al. (2000). In the penalty algorithm, one of the contact surfaces is defined as the master surface and the other as a slave surface. If contact is obtained between a slave node and the master surface, the slave node attempts to penetrate the master surface. However, as the slave nodes are constrained to slide on the master surface after contact (they must remain on the master surface), the penalty algorithm will introduce normal interface springs between the penetrating nodes and the contact surface. The spring stiffness matrix (from the interface springs) is then assembled into the global stiffness matrix. The stiffness of the interface spring is the minimum of the master segment stiffness and the slave node stiffness. The magnitude of the interface force is thus proportional to the amount of penetration. If there is no contact (slave node does not penetrate), nothing is done. With this contact algorithm, it is possible to simulate loss of contact and recovered contact between wheel and rail and between sleeper/USP and ballast bed.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

329

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Optimisation The finite element model was built-up using the pre-processor TrueGrid (Truegrid, 2001) and the train/track interaction problem was solved using the commercially available finite element software LS-DYNA (Hallquist, 2006). The advantage of the software is that it automatically makes the time step small so that high-frequency variations are well represented. Optimisation was achieved using the associated software optimisation package LS-OP (Stander, 1999) and is described in Lundqvist (2005). Optimal Ballast / Subgrade Stiffness In order to investigate an optimal stiffness of the ballast / subgrade in the transition zone the model spanned a length of 45 sleepers. This size of model was selected so that boundary effects would not disturb the track responses investigated in the central portion of 25 sleepers. In the model the substructure (ballast, sub-ballast and subgrade) was modelled as a continuum with elastic material properties. Longitudinally, the ballast bed was divided into several different sections. Two sections at the ends of the model, spanning 15 sleepers, were used to represent stiff and soft sections of the track respectively (see Table 4). The central part of the model consisted of five shorter sections, each of three sleeper spans. The stiffnesses of the short sections were varied to determine an optimum combination for the two cases when the load is travelling from stiff to soft track and from soft to stiff track, respectively. The optimal stiffness of the transition zone determined from the study, for both travelling directions, can be seen in Figure 9. When moving from stiff to soft track, the ideal stiffness change in the transition zone was found to be smooth in the beginning and at the end of the zone, with a more rapid stiffness change in the central part of the zone, see Figure 9(a). The optimal modulii shown in Figure 9 (from left to right) are 100, 93, 82, 71, 45, 38, and 30 MPa respectively. For the study of the wheel load moving from soft to stiff track, the ideal change in stiffness of the transition zone was found to be almost linear as shown in Figure 9(b), with optimal values of stiffness of the five sections of 30, 40, 50, 60, 70, 80, and 100 MPa respectively. The wheel/rail contact force (travelling from stiff to soft track) is shown in Figure 10. A rapid decrease in the contact force is noted when the wheel enters the soft region. This implies a downwards motion of the wheel, and when this downward motion comes to an end, there is a large increase of the contact force. It can be seen that the large amplitude in the contact force that is obtained when there is no transition zone has almost disappeared after the optimisation. Only small variations of the contact force occur at every small change of stiffness in the transition zone. Going from soft to stiff track is worse than going from stiff to soft and the wheel/rail contact force variation is then larger than the variation shown in Figure 10 (see Lundqvist, 2005). The study also demonstrated that if the transition zone is optimised for one direction of travel, then the results are almost as good for trains running in the opposite direction. Consequently, this suggests that if the transition zone is optimised for a particular direction of travel, then the transition when travelling in the opposite direction is almost as smooth as if the transition zone had been optimised for that direction.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

330

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 9. Optimal stiffness of the transition zone for the two cases (a) going from stiff to soft track (left figure, Young’s modulus E going from 100 to 30 MPa), and (b) going from soft to stiff track (right figure).

Figure 10. The wheel/rail contact force before and after track stiffness optimisation. Train (wheelset) travelling from stiff to soft track.

Under-sleeper pad stiffness For the study of the effect of under-sleeper pad stiffnesses on overall track stiffness variation (as represented by the variation in wheel / rail contact force) a numerical model consisting of 30 sleepers was used. The model contained three sections each of 10 sleepers. The sleepers in the central section had 20 mm thick under-sleeper pads. The two other sections at either end of the model, consisted of a soft and stiff section respectively (see Table 4). Under-sleeper pads were placed under each of the 10 sleepers in the stiff section and their stiffness was varied in order to achieve as smooth a transition as possible between the soft and stiff part of the track.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

331

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Shear Modulus of USP [GPa]

In order to keep the number of optimisation variables low, the same stiffness of the USP was given to two adjacent sleepers, so that the number of optimisation variables was five. The shear modulus G of the USP material was selected as the parameter to be optimized, and its lower limit was set to G = 10 MPa. From the study, the optimal values of the shear modulus of the USP material are shown in Figure 11 where it may be seen that the first two USP (on the stiff part of the track) should have a relatively low stiffness. Perhaps surprisingly, the modelling shows that the stiffness of the USPs under sleepers three and four should be considerably greater than for the first two sleepers, whilst those under sleeper five to eight should be similar in value but less than the previous two USPs. The last two sleepers should have the stiffest pads of all. The wheel/rail contact force is shown in Figure 12 both with and without the presence of USPs. A large irregularity can be seen at time t = 0.12 s, where the track stiffness changes from soft to stiff. Having optimal values of the USP stiffness this irregularity is almost completely eliminated, as shown by the second curve in Figure 12.

200

100

1

2

3

4

5

6

7

8

9

10

USP No Figure 11. Optimised values of shear modulus of USP material.

In order to investigate the robustness of the optimal solution shown in Figure 11, the model was run with two other distributions of USP stiffness. The two other stiffness distributions investigated (without optimisation) had the following stiffnesses respectively; 10, 100, 125, 150, and 175 GPa, and; 10, 150, 150, 150, 150 GPa. It was found that these two stiffness distributions gave almost the same result as the optimised distribution in Figure 11. This suggests that as long as the two first under-sleeper pads are very soft (by one order of magnitude), the stiffness of the following eight pads does not influence the result very much and therefore it is practical to use USPs to reduce effectively the stiffness variation, for example, at the transition from a “soft” embankment to at “stiff” concrete construction such as a bridge.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

332

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 12. Wheel/rail contact force for track without USP and with five optimised stiffnesses of USP. Transition from soft to stiff track occurs approximately at time t = 0.12 s.

Hanging Sleepers Hanging, or voided, sleepers occur because of differential track settlement, often due to variations in ballast stiffness. This results in some sleepers not being fully supported to the extent that contact with the ballast bed may be lost. Where hanging sleepers occur the vertical track stiffness becomes very low and this in turn causes high train / track interaction forces which may result in an increased rate of track settlement. One of the tasks of the sleepers is to distribute train induced loads to the ballast. However, in sections of the track with hanging sleepers, the sleepers adjacent to those which are hanging carry an increased load. This in turn may overstress the ballast and cause differential settlement to occur. Thus accelerating track settlement further. To investigate this phenomena one fully supported sleeper between two hanging ones was modelled and results were obtained when no USP, a USP of medium stiffness and a stiff USP, respectively, were placed under the hanging sleeper (Witt, 2008). The stiffness of the medium and high stiffness USPs were 400kN/mm and 3000 kN/mm respectively. Hanging sleepers were modelled by reducing the Young’s modulus of the ballast from 100 MPa to 0.1 MPa under the hanging sleepers. The results from the analysis are shown as a wheel / rail contact force diagram in Figure 13. The first hanging sleeper is reached after 0.132 s, the supported sleeper after 0.139s and the second hanging sleeper after 0.146 s. In Figure 13 it may be seen that the contact force decreases in front of the first hanging sleeper for all 3 cases, to 87, 85 and 89 kN without, with stiff and with medium stiffness USPs respectively. Generally the contact force with stiff USPs is lower throughout and reduces faster than with no USPs or USPs with medium stiffness. With USPs of medium stiffness, the contact force oscillates with higher amplitudes with a longer time period after the wheel has passed the sleepers. This demonstrates that USPs can be used to effectively reduce the effects of hanging sleepers, in particular the wheel rail contact forces, however the choice of stiffness of the USP is very important.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

333

Figure 13. Vertical wheel/rail contact force with one supported sleeper between two hanging ones.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Track Stiffness Measurement The measurement and assessment of the vertical track stiffness is an important tool for determining the structural causes of poorly performing sites (Esveld, 2001; Ebersöhn and Selig, 1994; Sussmann et al., 2001). Berggren (2005), for example, shows how soft soil properties can be estimated from dynamic stiffness measurements. Many structural problems first manifest themselves as irregularities in track geometry. However, their cause is often related to the structural performance of the substructure and so they cannot be diagnosed correctly by track geometry measurements alone and require in addition the track stiffness to be measured (Ebersöhn and Selig, 1994; Read et al., 1994; Selig and Li 1994; Sussmann et al., 2001). Sussmann et al. (2001) and Brough et al. (2003) relate different kind of track problems with track stiffness and recommend possible solutions for each problem (see Table 5). They suggest that the cause of unacceptably low values of track stiffness are mostly likely to be an indication of a weak subgrade as the properties of the subgrade most influence the value of track stiffness. In some cases the weakness may be attributed to poor drainage. However, by the time the problem has fully manifested itself the redesign and reconstruction of the track substructure may be required to reduce the traffic induced stresses in the subgrade to acceptable levels. Alternatively, methods of soil reinforcement to increase the strength and stiffness of the subgrade can be used. In some cases low track stiffness values can also result from fouled or dirty ballast that prevents adequate support for track loading. This problem occurs especially when the ballast and subgrade deteriorate in the presence of water due to train induced repeated loading. The resulting migration of fines into the ballast and subsequent formation of wet spots, can lead to

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

334

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

a reduction in stiffness of the track support system and hence loss of track geometry. Whilst replacing, or cleaning, the ballast may improve the problem in the short term, longer term solutions may require; the improvement of the drainage; the use of appropriate materials at the subgrade/ballast interface to reduce pumping and the migration of fines, and; in particularly problematic cases redesign and reconstruction. For the problem of an unacceptable variability in track stiffness Sussmann et al. (2001) suggest that potential solutions include the design of rail seat pad stiffness, the appropriate design of the substructure and the use of under ballast mats as discussed above. Sometimes very large deflections may occur under load indicating the presence of hanging sleepers or loose rail seat fasteners. These types of problems can usually be fixed by repairing fasteners, the appropriate tamping or stone blowing of the ballast, and undercutting when the void is due to fouled ballast that deforms easily under a load (Sussmann et al., 2001).

Table 5. Relation between stiffness and track problem and recommended maintenance (after Susmann et al., 2001). Parameter Low track stiffness

Variable track stiffness

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Voided sleepers

Problem Poor or weak subsoil or fouled ballast Variable track support (stiffness or modulus of subgrade) Fouled ballast, local settlement, poor fastener condition

Maintenance / Rehabilitation Substructure design, Stabilize subgrade Matching rail seat pads, substructure design, ballast mats Inspect fasteners, tamp, stoneblow, undercut

As mentioned previously an additional consideration, although one which occurs less often, are problems associated with high stiffness which can lead to faster deterioration of the track and its components due to higher dynamic loads. In such cases several methods may be used to reduce the stiffness including the installation of soft pads and resiliently mounted sleepers (Hildebrand, 2001; Remington, 1987). Soft pads, however, may have an undesired side effect of increased noise radiation.

Noise Radiation The noise radiated by the track is a function of, amongst other things, the receptance of the rail and particularly the stiffness of the rail pad between the rail and the sleeper (Jones and Thompson, 2001). For noise, only frequencies in the audible band are of interest (i.e. 20 – 20000 Hz). The upper limit as seen from the track is 2000 – 5000 Hz (Hildebrand, 2001). There is, obviously noise at higher frequencies, for example wheel and braking noise, but the track (rail) is not the dominant source. Whilst soft pads may reduce track stiffness effectively they can increase noise emanating from the rail as their use causes the rail effectively to become uncoupled from the sleeper. This minimizes noise from the sleeper but enables the rail to vibrate more freely so that waves can travel over a greater distance, increasing the noise from the rail. Conversely, with stiff pads the contribution from the rail is reduced but that from the sleepers is increased (Jones and Thompson, 2001). Ground borne vibration problems are associated with subgrade soils of low stiffness and /or a clear resonance as these tend to propagate vibrations more effectively. Smekal et al.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

335

(2003) give an example of how such potentially problematic sites may be identified. In their work they show how low frequency soil properties can be identified with the help of stiffness measurements at different frequencies.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Measuring Track Stiffness There are a number of methods which may be used to measure the vertical track stiffness. Perhaps the most important distinction is between those that make measurements at discrete intervals whilst static, and those that measure stiffness continuously whilst moving (i.e. rolling measurement). Static measurements are more widely used, often for research purposes, whereas rolling measurement techniques are in their infancy. As mentioned previously, track stiffness is a function of frequency and it is necessary to select an appropriate device for measurement depending on the frequency of interest. The static and low frequency dynamics of the track are mostly related to geotechnical and geodynamical issues. Measurements of track stiffness at these frequencies may be very useful for investigations related to the bearing capacity of the subgrade, ground borne vibrations and some soft-soil related problems. High frequencies relate to problems associated with noise and train-track impact forces. Static Measurement Amongst the number of static devices available for measuring track stiffness four are described. These are by the use of simple instrumentation, the impact hammer, the FWD and track loading vehicles. A very simple method of measuring track stiffness can be achieved by instrumenting any number of sleepers, and or rails, with displacement transducers or accelerometers and measuring the response during the passage of a train. The associated stiffness can then be calculated for that track section if the axle load is known. For improved accuracy, where dynamic loads are taken into consideration, the load from the train can also be measured with the help of strain gauges on the rail web, or on the sleeper. The typical results from such a measurement are load – deflection diagrams, where the stiffness can be identified with any of the above definitions of stiffness. The impact hammer is a hand held device which is used to hit the rail or sleeper (Rasmussen and Man, 2000; Read, 1994). The hammer head is equipped with a force transducer to measure the impulse, and an accelerometer is attached to the rail head or the sleeper. The transfer function between the impulse force of the hammer and acceleration of the rail is calculated (often it is also double integrated to get receptance instead of accelerance). Typically, a frequency interval of 50 – 1500 Hz can be covered using the hammer depending on the material used for the top of the hammer head. Rubber gives lower frequencies than a metal top for example. Since frequencies below 50 Hz are not recorded, impact hammer tests are most suited to problems associated with noise, vibration and wheelrail contact forces. The FWD is a device which is most often used to measure the stiffness of the track structure excluding the rails (Burrow et al., 2007b) (see Figure 14). The standard FWD device consists of a mass that is dropped from a known height onto rubber buffers mounted on a footplate. The resulting impact is measured by a load cell on the centre of the plate and velocity transducers are used to determine surface velocity at various distances, d, from the footplate (see Figure 15). The velocities are integrated to give vertical displacements. For

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

336

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

railway tracks in the UK, the device is designed to apply a 125 kN load to a sleeper, disconnected from the rails, via a 1.1 m long loading beam shaped to distribute the load to both ends of the sleeper. This loading system is considered to produce a load pulse which is similar to that applied by a single axle of a train travelling at high speed (Sharpe and Collop, 1998). The magnitude of the applied load is measured in the centre of the loading beam and the velocity transducers are positioned on the loaded sleeper and on the ballast at various distances from the centre of the beam (see Figure 15).

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 14: Falling weight deflectometer

Figure 15: Schematic of the falling weight deflectometer

The track stiffness is calculated from the load and deflections measured at some of the geophones, depending on the application. For example, if the stiffness of the ballast and subballast layers is to be established the track stiffness is often determined from the load and

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

337

deflections measured immediately under the load and from those measured by geophones 1000 mm from the load as follows:

k=

62.5 kN/mm/sleeper end (d 0 − d1000 )

(45)

Typical outputs from the FWD are given in Figures 16 and 17. Figure 16 shows the measured deflections as a function of distance along the track for geophones positioned 0, 300, 1000 and 1500 mm respectively from the impact load, while Figure 17 shows the track stiffness determined using Equation 45. 4

d0 d300

3.5

d1000 d1500

Deflection (mm)

3

2.5

2

1.5

1

0.5

0 0.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

80.00

90.00

Distrance (m)

130

Effective Stiffness - 07/06/03 120 110 100 90 Stiffness (kN/mm/sleeper end)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 16: FWD deflections

80 70 60 50 40 30 20 10 0 0.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

80.00

90.00

Distance (m)

Figure 17: FWD track stiffness

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

338

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

TLV A Track Loading Vehicle (TLV) uses its own weight to load the track with the help of hydraulic jacks. Usually the rail heads are loaded, but the sleeper can be loaded also with the rails decoupled. Depending on the equipment different loads can be applied. A number of railway infrastructure companies own and operate TLVs. These include the Swedish TLV which has a weight of 49 tons and can load each rail statically up to 150 kN and excite dynamically up to 200 Hz (see Figure 18). It can also measure lateral track stability / stiffness (Köhler, 1999). Other devices include those developed in the USA by the Transportation Technology Center, Inc. (TTCI) (Thompson et al., 2001) and the DECAROTOR (Chorus and Zarembski, 1982); the South African BSSM (Ebersöhn, 1997) and a modified tamper described by Esveld (1980).

Figure 18: The vertical and lateral hydraulic actuators of the Swedish TLV (Berggren, 2005b).

The main advantage of a standstill TLV compared to rolling measurements is that the preload, dynamic load and frequency range can be varied to a greater degree. However, the process is more time consuming and requires the railway track to be closed.

Rolling Measurement If standstill measurements have been used mainly for research purposes, rolling measurements have the potential to be used on a more regular basis for maintenance purposes. Whilst there are several different systems for measuring the vertical track stiffness along the track, most measure the displacement under one or two axles caused by the weight on the axles and the track flexibility. With knowledge of the static axle loads, the track stiffness can be calculated. In case of a two-axle system, the axle loads are different and the lightest loaded axle is used to remove the effect of track irregularities on the stiffness measurement.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

339

Measurement Principles The vertical track stiffness measured by each device is unlikely to be identical for a number of reasons as follows: • •

• •

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

•

Static preload: The static preloads applied are different and are therefore likely to result in different stiffness values being recorded for the same section of track. Excitation frequency / speed: Equipment using a static running wheelset to load the track will excite the track with a range of frequencies which is a function of the speed of the vehicle. As the measuring speed increases, so will the frequency content. Since the dynamic track stiffness is not constant with frequency, the stiffness determined is likely to differ (see Equation 4). Spatial resolution: The different measurement techniques may have different spatial resolutions. Model dependency: The devices measure the deflection of the rail different distances away from the wheelset. Where the deflection is not measured directly under the wheelset a model for the rail bending has to be used in order to calculate the rail deflection under the wheelset. These models are approximations of reality and can introduce uncertainty and related errors. Degree of influence from track irregularities: Track geometry irregularities, especially those associated with the vertical alignment, can influence the stiffness measurements since the displacement transducers used in the equipment in most cases measure a combination of deflection due to track flexibility and displacement due to track geometry irregularities. Wheel out-of-roundness and wheel flats introduce similar disturbances.

Devices A number of organisations have developed rolling devices to measure track stiffness. Some of these are summarised below (Berggren, 2005b).

People’s Republic of China The China Academy of Railway Sciences was one of the first organisations to develop a system for continuous track stiffness measurement (Wangquing et al., 1997). Their system, which travel at speeds of up to 60 km/h, uses two track geometry chord measurement systems with different loading applied to each of the measurement axles (Figure 19). The light-weight car has a weight of 40 kN and is used to reduce the effect of track geometry irregularities on the stiffness measurement recorded by the heavy-weighted car. A weight of 40 kN was chosen as it was found to be sufficient to reduce the effects of voided sleepers. The load of the heavy-weighted car can be varied between 80 and 250 kN enabling the nonlinear characteristics of the same section of track to be investigated by repeating with different loads.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

340

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Figure 19. Principle of Chinese track stiffness measurements (after Wangquing et al., 1997)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

On railway tracks with a speed limit 160 km/h, a track stiffness between 65 – 100 kN/mm (one rail) obtained using the device was found to be optimal.

TU Delft, Netherlands TU Delft’s High Speed Deflectograph (HSD) makes use of laser doppler sensors attached to a moving railway vehicle, travelling at speeds of up to 130 km/h, to measure the rail bending velocity (Rasmussen et al., 2002). An inertial unit (3-axle gyro and accelerometer) is used as an input to a servo system to control the laser position so that the laser rays are perpendicular to the rail. At the moment the system exists in concept only for railways although a fully working prototype for roads has been developed. The HSD has a number of advantages over other rolling devices including: a) the effect of track geometry irregularities on the measurement of track stiffness is much less than when displacement transducers are used, although the effect of hanging sleepers still contributes to the rail bending velocity; b) the rail bending velocity increases with train speed and as a result higher trains speeds are likely to produce more accurate results. TTCI, USA TTCI’s track loading vehicle (TLV) has been developed to measure both lateral and vertical stiffness at standstill and when moving at speeds of up to 16 km/h (Li et al., 2002; Thompson et al., 2001). For rolling vertical stiffness measurements the TLV for static measurements is coupled with an empty car. The TLV has a fifth wheelset (loaded bogie) mounted underneath the vehicle centre, that can be loaded hydraulically (both vertically and laterally) with vertical loads between 4 – 267 kN. A load of 178 kN is applied to the test axle of the static TLV. If two separate runs are used to differentiate the supports between the ballast and the subsoil, a light test axle load of 44 kN is used for the second run. The deflection is measured with the help of laser sensors, yielding a chord measurement of rail bending deflection (see Figure 20).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

341

Figure 20. Rail bending deflection measurement with lasers, yielding chord values (after Li et al., 2002).

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Measurements are also made under the empty car, which is also equipped with a centre loaded bogie with pneumatic actuators capable of applying a nominal load of 9 kN.

Figure 21. Example of results from TTCI stiffness measurements (Li et al., 2002).

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

342

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Typical results from the TLV are shown in Figure 21. In the upper part of Figure 21 total deflection is shown for two cases when the TLV has loads of 44 kN and 178 kN respectively. In the middle figure the difference between these two deflections is shown and this represents a measure of subgrade stiffness since the deflection under the 178 kN load is that due to the entire track system, whilst that under the 44 kN load corresponds (approximately) the response of the rail, pad, sleeper and ballast. The lower figure is a transformation (with help of the theory of beam on elastic foundation) from contact deflection (middle figure) into track modulus.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Banverket, Sweden The Swedish vehicle, known as the Rolling Stiffness Measurement Vehicle (RSMV), is a modified two-axle freight wagon. The vehicle dynamically excites the track using two oscillating masses above one of the ordinary wheel axles as shown in Figure 22 and 23. Track stiffness is calculated from the measured force and acceleration (Berggren et al. 2005).

Figure 22. The measurement equipment in the RSMV (vertically moving masses above measuring axle, contained in steel cages) 0.

The RSMV can measure the dynamic stiffness at frequencies of up to 50 Hz with a static axle load of 180 kN (or more) and a maximum dynamic axle load amplitude of 60 kN. Measurements at higher speeds (up to 60 km/h) with up to 3 simultaneous sinusoidal excitation frequencies or more detailed investigations at lower speeds (below 10 km/h) with noise excitation can also be performed.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

343

-0

350

-5

300

-10

250

-15

200

-20

150 38.8 38.9

39 39.1 39.2 39.3 39.4 Position along the track [km]

-25 39.5 39.6

-10

0-

-20

-5

-30

-10

-40

-15 -20

-50 -60

Clay depth [ m ]

400

Clay depth [ m ]

Track stiffness phase [ degrees ]

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Track stiffness magnitude [ kN/mm ]

Figure 23. Measurement principle (one side only) of RSMV (Berggren et al. 2005a).

-25 38.8 38.9

39 39.1 39.2 39.3 39.4 Position along the track [km]

39.5 39.6

Figure 24. Repeatability test of stiffness magnitude (total) and phase – six measurements on the same track with a speed of 40 km/h and excitation at 11.4 Hz and dynamic load of 2x20 kN (solid lines). Standard deviation: 3.3 kN/mm, 1.3 degrees. Depth of clay layer is also indicated in the figure (xmarked dashed line) (Berggren et al. 2005a).

The measurement principle used by the RSMV has shown very good repeatability. Figure 24 shows the results from a test carried out on 800 metres of track whose stiffness was measured on 6 separate occasions with the same speed and excitation frequency. Also the depth of the clay layer is plotted showing clear correlation with expected substructure behaviour.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

344

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Czech Republic The Czech stiffness measurement equipment, called the SKMT, has been developed by Czech Technical University of Prague and the Commercial Railway Research Ltd (KZV) (Vymetal and Turek, 2001). It is essentially a chord-based deflection measurement system that determines the deflection from two different loads (two runs are required over the track). The chord measurements are taken using an adapted tamping machine fitted with trusses fitted on five special axles. The angles of rotation between adjacent trusses are measured with Linear Variable Displacement Transducers (LVDTs) and converted to track deflection. The vertical force is applied to both rails where the original lifting mechanism of the tamping machine was located. Two measuring runs are made on the same track, one without additional loading and the other with a static vertical load of 80 kN and the theory of a beam on an elastic foundation is used to obtain a stiffness value.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

University of Nebraska, USA The University of Nebraska at Lincoln (UNL) in the USA has developed a chord-based deflection measurement system to effect track stiffness measurements as shown in Figure 25 (Arnold et al., 2006; Norman et al., 2004). The technique uses line-lasers to measure relative rail deflection between the bogie and the rail (McVey et al., 2005). The measurement principle is shown in Figures 26 and 27. The relative deflection is measured using two lasers and a camera that measures the distance, d, between the two lines and as the sensor moves with respect to the rail surface, the distance between the laser lines changes. The Winkler model is used to relate the measured deflections to track modulus/stiffness.

Figure 25. Measurement vehicles (Lu et al., 2007)

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

345

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 26: Rail deflection / Sensor measurement of UNL-stiffness equipment (Norman et al., 2004)

Figure 27. Sensor geometry of UNL stiffness equipment (Norman et al., 2004)

France The French Portancemeter is a stiffness monitoring tool designed by CETE-Normandie Centre for road structures and is currently being adapted for use on railways (Hosseingholian et al., 2006). The device applies a dynamic load to the track via a vibrating wheel suspended by a spring and a damper. For road testing 10 kN of static weight, 0.5 mm of theoretical amplitude at 35 Hz is used. For rail testing these characteristics will be changed with increased loads (both static and dynamic) and the capability to alter the frequency.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

346

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Switzerland SBB Swiss railways, Schweizerische Bundesbahnen (SBB), has developed a device (see Figure 28) which is similar to the Chinese and TTCI equipment and uses two geometry measurement systems (Soldati, 2006).

Figure 28: Swiss track stiffness measurement vehicle (Soldati, 2006)

Lateral Track Stiffness Measurement

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Lateral track stiffness is associated with track stability (resistance against track buckling) and whilst it has not been discussed in this chapter in any detail, it is an important parameter. Both the Swedish TLV and TTCI’s TLV are capable of its measurement

Conclusion The vertical track stiffness is an important measure of the structural performance of the railway track and is useful for helping to determine why railway track may be performing poorly. New track, especially that built for high speed lines, must be designed and constructed to appropriate standards so that the track stiffness is within an acceptable range of values. If the stiffness is lower than the acceptable range of values, excessive track displacements can occur and on the other hand unacceptable track deterioration may take place when the stiffness is higher than ideal. This chapter described an analytical approach to railway substructure design which considers track stiffness and the protection of the subgrade. It has the potential to cater for changes in traffic conditions and materials. A number of design procedures are readily available in the literature and it is hoped that this chapter will help the reader to select and modify, where appropriate, a suitable procedure to give a realistic design for a given set of conditions. Further the issue of variable track stiffness was addressed in the chapter. It was shown that deviations in track stiffness can induce variations of the wheel/rail contact force which will contribute to track structure deterioration and differential track settlement. This may in turn give rise to unsupported sleepers. The track degradation increases the rate of track deterioration leading to further track quality related

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

347

problems. Modelling techniques were described and used to investigate techniques which can be used to alleviate track stiffness variations. It was demonstrated that a transition zone between track sections of different stiffness can be created to obtain a smooth transition between the two sections. The optimal stiffness variation in the transition zone depends on the travelling direction, but it is not very sensitive to it. Also, under sleeper pads with nonoptimised stiffnesses can significantly reduce the wheel/rail contact force variation. The optimal transition zone can be built by using elastomeric products, such as under sleeper pads and/or sub ballast mats, to construct a tailor-made transition zone with desired stiffness variation and geometry. A thorough understanding of the physical mechanisms causing track deterioration, and understanding of the relationship between the track design parameters and the long-term track maintenance requirement would therefore imply that an optimised (or at least an improved) ballasted track can be constructed. The measurement of the vertical track stiffness and the correct understanding of how measurement procedures calculate stiffness are important so that the correct diagnosis of the causes of track geometry problems can be diagnosed. To this end chapter concluded with a comparison of state of the art techniques which may be used to measure track stiffness. Methods which measure the vertical track stiffness discretely and those which effect its measurement continuously were compared and contrasted.

Acknowledgements Track modelling and calculations to study changes in track stiffness were carried out by Andreas Lundqvist (2005), Rikard Larsson (Lundquist et al. 2006) and Stephen Witt (2008) at Linköping University, Sweden.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

References Aknin, P. and Chollet, H. (1999). A new approach for the modelling of track geometry recording vehicles and the deconvolution of versine measurements, Vehicle System Dynamics Supplement 33, 1999. Andersen L and Nielsen S R K (2003). Vibrations of a track caused by variation of the foundation stiffness. Probabilistic Engineering Mechanics, Vol 18, pp 171-184, 2003. Andersson C and Dahlberg T (2000). Load impacts at railway turnout crossing. Vehicle System Dynamics, Vol 33(Supplement), pp 131-142, 2000. Anon. (2006). Research Result Digest 79 - Design of track transitions. Transportation Research Board, TRB, 500 Fifth Street, NW, Washington, DC 20001, USA. Arnold, R., Lu, S., Hogan, C., Farritor, S., Fateh, M. and El-Sibaie, M. (2006). Measurement of Vertical Track Modulus From a Moving Railcar, Proceedings of the AREMA 2006 Annual Conference, Louisville, KY, September 17-20, 2006. Belytschko T, Liu W K, and Moran B (2000). Nonlinear Finite Elements for Continua and Structures. John Wiley & Sons Ltd, Chichester 2000.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

348

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Berggren, E. G. (2006): Measurements of track stiffness and track irregularities to detect short waved support conditions, Proceedings from RailFound 06 – International conference on Railway Track Foundation, Birmingham 11-13 September 2006. Berggren, E. G., Jahlénius, Å., Bengtsson, B. E. and Berg, M. (2005a). Simulation, Development and Field Testing of a Track Stiffness Measurement Vehicle, Proceedings of 8th International Heavy Haul Conference, Rio de Janeiro, 13-16 June 2005. Berggren, E.G. (2005b). Dynamic Track Stiffness Measurement – A New Tool for Condition Monitoring of Track Substructure, Licentiate thesis TRITA AVE 2005:14, Royal Institute of Technology (KTH), Stockholm 2005.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Berggren, E.G., Jahlénius, Å., and Bengtsson, B. E. (2002). Continuous track stiffness measurement: An effective method to investigate the structural conditions of the track, Proceedings from Railway Engineering Conference, London 2002. Blair, and Chan A.H. (2006). Modelling Track Deterioration – A Brief Review of Past and Current Trends. RailFound06, the First International Conference on Railway Foundations. The University of Birmingham, UK. Boyce, H.R. (1980). A non linear model for the elastic behaviour of granular materials under repeated loading. Int. Symp. on Soils under Cyclic and Transient Loading, pp. 285 – 294, Swansea , U.K. Brandl, H. (2001). Geotechnics of rail track structures. In Geotechnics for Roads, Rail Tracks and Earth Structures (A.G. Correia and H. Brandl, eds.). The Netherlands. pp 47 – 68. A.A. Balkema Publishers. Brough, M.J., Ghataora, G.S., Stirling, A.B., Madelin, K.B., Rogers, C.D.F. and Chapman, D.N. (2003). Investigation of Railway Track Subgrade. l: In-situ Assessment. Proceedings of the Institution of Civil Engineers: Transport, 156. Brown S F, Kwan J, and Thom N H (2007). Identifying the key parameters that influence geogrid reinforcement of railway ballast. Geotextiles and Geomembranes (2007), in press (doi. 10.1016/j.geotexmem.2007.06.003). Brown S.F, Lashine A.K.F. and Hyde A.F.L. (1975). Repeated load triaxial testing of a silty clay. Geotechnique, 1975, 25, No 1, 95-114. Brown, S.F. (1996). 36th Rankine Lecture: Soil Mechanics in Pavement Engineering, Geotechnique, 46(3). pp 383-426. Brown, S.F., and Pell, P.S. (1967). An experimental investigation of the stress, strains and deflections in layered pavement structure subjected to dynamic loads. Proceedings of the 2nd International Conference on Structural Design of Asphalt Pavements, pp 487-504. Ann Arbor, USA. Burrow, M.P.N., Bowness D. and Ghataora, G.S. (2007a). A comparison of railway track foundation design methods. Proc. IMechE Part F: J. Rail and Rapid Transit. Vol 221, No F1, pp 1 – 12. Burrow, M.P.N., Chan, A. H.C, and Shein, A. (2007b). Falling weight deflectometer based inverse analysis of ballasted railway tracks. Geotechnical Engineering, Proceedings of the Institution of Civil Engineers. 160 July 2007 Issue GE3. pp 169–177.

Chan, F.W.K and Brown, S.F. (1994). Significance of principal stress rotation in pavements. XIII International symposium on soil mechanics and foundation engineering, New Delhi, India.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

349

Chorus, J. and Zarembski, A. M. (1982). Track Strength Vehicle Testing on High Curvature Mainline Track, Rail International 7, July 1982. Dahlberg T (2007). Railway track stiffness variations. A literature review. Report for the REVA project (Reduction of the track vertical stiffness variability). Unpublished. October 25, 2007. Dahlberg, T (2006). Track issues. Chapter 6 in Iwnicki S (editor). Handbook of Railway Vehicle Dynamics. CRC, Taylor & Francis, Boca Raton, ISBN-13. 978-0-8493-3321-7. Dahlberg, T. (1995). Vertical Dynamic Train / Track Interaction – Verifying a Theoretical Model by Full-Scale Experiments. Vehicle System Dynamics. Vol. 24 pp 45 – 57. Drumm, E.C., Boateng-Poku, Yl and Pierce, T.J. (1991). Estimation of subgrade resilient modulus from standard tests. Journal Geotech. Engineering, ASCE, 116(5), 774 – 789. Ebersöhn, W. (1995) Substructure Influence on Track Maintenance Requirements, PhD Dissertation, Department of Civil and Environmental Engineering, University of Massachusetts at Amherst 1995. Ebersöhn, W. (1997). Track Maintenance Management Philosophy, Proceedings from 6th IHHA Conference, Cape Town 1997. Ebersöhn, W. and Selig, E. T. (1994). Track Modulus Measurements on a Heavy Haul Line, Transportation Research Record 1470, TRB, National Research Council, Washington D.C. 1994.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Eisenmann, J. (1992), Railway track foundation dynamics. Railway Engineering International Edition. No. 2. pp 17 - 20 Esveld, C. (1980) Track Stiffness Measurements Using an Adapted Tamping Machine, Rail International, February 1980. Esveld, C. (2001). Modern Railway Track, Second Edition. MRT-Productions, Zaltbommel, The Netherlands. Ewins, D. J. (1984). Modal Testing: Theory and Practice, Research Studies Press LTD, Taunton, Somerset, U.K. 1984. FRA (2004). FRA report RR04-03, A Preliminary Design of a System to Measure Vertical Track Modulus from a Moving Railcar, January 2004, http://www.fra.dot.gov/downloads/-Research/-rr0403.pdf. Fredlund D.G., Bergan, A.t and Sauer, E. K. (1977). Relationship between resilient modulus and stress conditions for cohesive subgrade soils. Transportation Research Record, 642, pp 73 – 81. Frost, M.W., Fleming, P. R. and Rogers C.D.F. (2003). Cyclic triaxial test on clay subgrades for analytical pavement design. In Transportation geotechnics, proceedings of a symposium held at Nottingham Trent University, Thomas Telford Ltd., London. Gomez Correia A. (2004). Evaluation of mechanical properties of unbound granular materials for pavements and rail tracks. ISSMGE TC3 International Seminar on Geotechnics in Pavement and Railway Design and Construction NTUA - Athens, 16-17 December 2004. Gomez Correia, A., Hornych, P. and Akou, Y. (1999). Review of models and modelling of unbound granular materials, Int. Workshop on Modelling and Adv. Testing for Unbound Granular Materials, Lisboa, Baleema, Rotterdam. pp 3 -15.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

350

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Gräbe, P.J. and Clayton, C.R.I. (2003). Permanent deformation of railway foundations under heavy axle loading. International Heavy Haul Association. Specialist technical session. Dallas, Texas. Hallquist J O (2006). LS-DYNA Theory Manual, Livermore Software Technology Corporation, 2876 Waverly Way, Livermore, California 94550. Hildebrand R. (2001). Countermeasures against Railway Ground and Track Vibrations, PhD Dissertation, Royal Institute of Technology (KTH), Stockholm, 2001. Hornych P., Kazai, A. and Piau, M. (1998). Study of the resilient behaviour of unbound granular materials. Proceedings, BDRA 98, Vol. 3, pp 1277 – 1287, Trondheim, Norway. Hosseingholian, M. Froumentin, M. and Robinet, A. (2006). Feasibility of a Continuous Method to Measure Track Stiffness. Proceedings from Railway Foundations conference. Birmingham 2006.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Huille, J.P. and Hunt, G.A. (2000). Track Stiffness: A Tool for Maintenance? Railway Engineering 2000, London, U.K. Hunt H E M (1997). Settlement of railway track near bridge abutments. Proc. Instn Civ. Engrs, Transport, Vol 123(1), pp 68-73, 2004. Johansson A, Nielsen J C O, Bolmsvik R, Karlström A, and Lundén R (2006). Under sleeper pads – influence on dynamic train-track interaction. Proceedings of 7th International Conference on Contact Mechanics and Wear of Rail/Wheel Systems (CM2006), Brisbane, Australia, September 24-27, 2006. Jones, C.J.C. and Thompson, D.J. (2001). Means of Controlling Rolling Noise at Source, Ch. 6 from Noise and Vibration from High-Speed Trains, Editor V. Krylov, Thomas Telford Ltd, 2001. Jönsson P-A and Stichel S (2007). On the influence of Freight Traffic Operational Conditions on Track Deterioration Cost. Proc International Heavy Haul Association, 2007 Specialist Technical Session “High Tech in Heavy Haul”, June 11-13, 2007, Kiruna, Sweden. IHHA Forest Hills Court, Virginia Beach, VA USA. Kabo E, Nielsen J C O, Ekberg A (2006). Prediction of dynamic train-track interaction and subsequent material deterioration in the presence of insulated rail joints. Vehicle System Dynamics, Vol 44(Supplement), 718-729, 2006. Kassa E (2007). Dynamic Train-Turnout Interaction - Mathematical modelling, Numerical Simulation and Field Testing. PhD Thesis, Dept of Applied Mechanics, Chalmers Univ of Tech, Gothenburg, Sweden, ISBN 978-91-7385-003-2. Köhler, J. (1999) Track Loading Vehicle – Description, Internal report: Banverket BMA 98/22k, Borlänge 1999. Koro K, Abe K, Ishida M, and Suzuki T (2004). Timoshenko beam finite element for vehicletrack vibration analysis and its application to jointed railway track. Proc. Instn Mech Engrs, Part F. Journal of Rail and Rapid Transit, Vol 218, pp 159-172, 2004. Lekarp. F., Isacsson, U. and Dawson A. (2000). State of the Art: 1. Resilient response of unbound aggregates. Journal of Transportation Engineering, ASCE, 126(1). pp 68 – 75. Lekarp. F., Richardson I.R., and Dawson A. (1996). Influences on permanent deformation behaviour of unbound granular materials. Transportation Research Record, TRB 1547, pp 68 – 75.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

351

Li D and Davis D (2005). Transition of railroad bridge approach. Journal of Geotechnical and Geoenvironmental Engineering, Vol 131(11), pp 1392-1398, 2005. Li, D. and Selig, E.T. (1994). Resilient modulus for fine-grained subgrade soils. Journal of Geotechnical Engineering, ASCE, 120(6). pp 939 – 957. Li, D. and Selig, E.T. (1996). Cumulative plastic deformation for find grained soils. Journal of Geotechnical Engineering, ASCE, 122(12). Li, D. and Selig, E.T. (1998a). Method for railroad track foundation design. I: Development. Journal of Geotechnical andd Geoenvironmental Engineering, ASCE, 124(4). Li, D. and Selig, E.T. (1998b). Method for railroad track foundation design. II: Applications. Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 124(4). Li, D., Shust, W. and Kalay, S. (1997). Measurement of Lateral Track Strength using AAR’s Track Loading Vehicle, Proceedings from 6th IHHA Conference, Cape Town 1997.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Li, D., Thompson, R. and Kalay, S. (2002). Development of Continuous Lateral and Vertical Track Stiffness Measurement Techniques, Proceedings from Railway Engineering Conference, London 2002. Loizos, A. (2004). Subgrade and unbound materials in highway and airport pavement engineering. ISSMGE TC3 International Seminar on Geotechnics in Pavement and Railway Design and Construction NTUA - Athens, 16-17 December 2004. Lopez Pita A, Teixeira P F, and Robuste F (2004). High speed and track deterioration. The role of vertical stiffness of the track. Proc. Instn Mech Engrs, Part F. Journal of Rail and Rapid Transit, Vol 218(F1) 2004, pp 31-40. Loy H (2006). Sleeper pads in turnouts – Optimizing the static rail deflection. Proc 14.th Nordic Seminar in Railway Technology (14.e Nordiska seminariet i Järnvägsteknik), April 26-27, 2006, Institute of Technology, Linköping University, Linköping, Sweden. S. Lu, S., Hogan, C., Minert, B. Arnold, R. Farritor, S. Seeger, W. Oliver, M. Lees, H. Thornton, D. Fateh, M. and Carr G.(2007). Vertical Track Modulus Testing On BNSFs Heavy Haul Coal Line, Proceedings of the AREMA 2007 Annual Conference, Chicago, IL, September 9-12, 2007. Lundqvist A (2005). Dynamic train/track interaction – Hanging sleepers, track stiffness variations, and track settlement. Linköping University Studies in Science and Technology, Thesis No. 1159, Dept of Mechanical Engineering, Linköping, Sweden, 2005. Lundqvist A and Dahlberg T (2005). Load impact on railwat track due to unsupported sleepers. Proc. Instn Mech Engrs, Part F. Journal of Rail and Rapid Transit, Vol 219 (F2), pp 67-77. Lundqvist A, Larsson R, and Dahlberg T (2006). Influence of railway track stiffness variations on wheel/rail contact force. Proc Workshop on TRACK for High-Speed Railways, Oct 12-13 2006, Civil Engineering Dept, University of Porto, Portugal, pp 6778. Mahmoud A A and Eltawil M A (1992). Beams on random elastic support. Applied Mathematical Modelling, Vol 16(6), pp 330-334, 1992. McVey, B., Norman, C., Wood, N., Farritor, S., Arnold, R., Fateh, M. and El-Sibaie, M. (2005). "Track Modulus Measurement From A Moving Railcar", Proceedings of the AREMA 2005 Annual Conference, Chicago, IL, September 25-27, 2005.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

352

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Meissonnier, F (2000). European research for an optimised ballasted track. Final report, synthesis part. European Union, Contract number. BRPR-CT97-0455. Project number. BE96-3263. Monismith, C.L., Ogawa, N., and Freeme, C.R., (1975). Permanent deformation characteristics of subgrade soils due to repeated loading. Transportation Research Record, TRB, 537. Monley G J and Wu J T H (1993). Tensile reinforcement effects on bridge-approach settlement. Journal of Geotechnical Engineering, Vol. 119(4), 1993. Moossazadeh, J. and Witczak, M.W. (1981). Prediction of subgrade modulus for fine-grained subgrade soils that exhibits nonlinear behaviour. Transportation Research Record, No 810. pp 9 -17. Naprstek J amd Fryba L (1995). Stochastic modelling of track and its substructure. Vehicle System Dynamics, Vol 24(Suppplement), pp 297-310, 1995. Nordborg A (1998). Parametrically excited rail/wheel vibrations due to track-support irregularities. Acta Acoustica (Stuttgart), Vol 84(5), pp 854-859, 1998. Norman, C., Farritor, S., Arnold, R., Elias, S.E.G (2003). Preliminary Design of a System to Measure Track Modulus from a Moving Railcar, Proceedings from Railway Engineering Conference, London 2003.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Norman, C., Farritor, S., Arnold, R., Elias, S.E.G., Fateh, M. and El-Sibaie, M. E. (2004). Design of a System to Measure Track Modulus From a Moving Railcar, Proceedings of the AREMA 2004 Annual Conference, Nashville, TN, May 16-21, 2004. Oscarsson J (2001). Dynamic train-track interaction. linear and non-linear track models with property scatter. PhD thesis, Dept of Solid Mechanics, Chalmers University of Technology, Göteborg, Sweden, ISSN 0346-718X, 2001. Oscarsson J (2002). Dynamic Train-Track Interaction. Variability Attributable to Scatter in the Track Properties. Vehicle System Dynamics, Vol 37(1) pp 59-79, 2002. Oscarsson J (2003). Dynamic train-track-ballast interaction with unevenly distributed track properties. Vehicle System Dynamics, Vol 37(Supplement), pp 385-396, 2003. Pappin, J.W., Brown, S.F. and O’Reilly, M.P. (1992). Effective stress behaviour of saturated and partially saturated granular material subjected to repeated loading, Géotechnique Vol. 42, No. 3, pp 485 – 497. Paute, J.L., Hornych, P. and Benaben, J.P. (1993). Repeated load triaxial testing of granular materials in the French network of laboratories des Point et Chausséess. Proceedings of the European Symposium on Flexible Pavements, Euroflex 1993, Laboratório Nacional de Engenharia Civil, Lisborn, Protugal. pp 81 – 95. Rasmussen, S. and Man, A. P. (2000), The High Speed Deflectograph and the Hammer Excitation Test: Two Measurement Techniques for Determining Track Performance, Rail Engineering International, Edition 2000 Number 1. Rasmussen, S., Krarup, J. A. and Hildebrand, G. (2002) Non-contact Deflection Measurement at High Speed, 6th International Conference on the Bearing Capacity of Roads, Railways and Airfields, Lisbon 24-26th June, 2002, Portugal. Raymond, G.P., Gaskin, P.N. and Addo-abedi, F.Y. (1979). Repeated compressive loading of Leda clay. Canadian Teotechn. Journal 16(1), pp 11 -20.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Track Stiffness Considerations for High Speed Railway Lines

353

Read, D., Chrismer, S., Ebersöhn, W. and Selig, E. T. (1994). Track Modulus Measurements at the Pueblo Soft Subgrade Site, Transportation Research Record 1470, TRB, National Research Council, Washington D.C. 1994. Remington, P. J., Kurzweil, L. G. and Towers, D. A. (1987). Low-Frequency Noise and Vibration from Trains, ch. 16 from Transportation Noise Reference Book, ed. Nelson, P. M. Butterworth & Co. Ltd, 1987. Schwartz, C.W. (2004). Computational issues for geotechnical aspects of pavements and rail beds. ISSMGE TC3 International Seminar on Geotechnics in Pavement and Railway Design and Construction NTUA - Athens, 16-17 December 2004. Selig E T and Li D (1994). Track modulus. its meaning and factors influencing it. Transportation Research Record, No 1470, ISSN 0361-1981, 1994. Selig, E.T. and Waters, J.W. (1994). Track geotechnology and substructure management. Thomas Telford Ltd., London. Shackel, B (1973). The derivation of complex stress-strain relations. Proceedings 8th International Conference on Soil Mechanics and Foundation Engineering, Moscow. pp 353 – 359. Sharpe P. and Collop A. C (1998). Trackbed investigation: a modern approach. In Proceedings of the 1st International Conference on Maintenance and Renewal of Permanent Way and Structures, London (FORD M.C. (ed.)). Engineering Technics Press, Edinburgh, 1998, 37–43. Shust, W., Li, D., and Lundgren, J. (1999) Summary of Track Strength Measurements in North America, Proceedings from Track Technology UIC Conference, Beijing 1999.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Smekal, A., Berggren, E., Hrubec, K. (2003). Track Substructure Investigations using Ground Penetrating Radar and Track Loading Vehicle, Proceedings from Railway Engineering Conference, London 2003. Soldati, G. (2006). EMW Kiesen 2006 – Gleiseinsenkungen mit dem Einsenkungsmesswagen Achslast 20 t, SBB-report, Bern 18 Jan 2007 (in German). Stander N (1999). LS-OPT User’s Manual. Livermore Software Technology Corporation, 1 edition, Livermore, 1999. Statens Järnvägar – Minnesskrift 1906 – 1931, Stockholm 1931. Stewart, H.E. and O’Rourke, T.D. (1988). Load factor method for dynamic track loadings. Journal of Transportation Engineering, ASCE, 114 (1). Sussmann, T. R.; Ebersöhn, W. and Selig, E. T. (2001). Fundamental Nonlinear Track LoadDeflection Behavior for Condition Evaluation, Transportation Research Record 1742, Paper No. 01-2916, 2001. Suzuki T, Ishida M, Abe K, and Koro K (2005). Measurement on dynamic behaviour of track near rail joints and prediction of track settlement. Quaterly Report of Railway Technical Research Institute, QR of RTRI, Vol 46(2), pp 124-129, 2005. Thompson, M.R. and Robnett, Q.L. (1976). Final Report, resilient properties of subgrade soils. FHWA-IL-UI-160, University of Illinois, Urbana, Ill. Thompson, R., Marquez, D. and Li, D. (2001). Track Strength Testing using TTCI’s Track Loading Vehicle, Railway Track & Structures, December 2001.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

354

Michael Burrow, Paulo Fonseca Teixeira, Tore Dahlberg et al.

Truegrid manual, version 2.1.0 (2001), XYZ Scientific Applications, Inc. Livermore, Ca, USA. Ullidtz, P. (2002). Analytical tools for pavement design. Key note address, 8th International Conference on Asphalt Pavements, Copenhagen. Uzan, J., Witczak, M.W., Scullion, T. and Lytton, R .R. (1992). Development and validation of realistic pavement response models. Proceedings of the 7th International conference on Asphalt Pavements, vol. 1. pp 334 – 350. Nottingham, U.K. Vymetal, A. and Turek, J. (2002), Continuous Measurement of Track Stiffness Using a Rolling Load, Proceedings from the RTSA CORE conference on Railway Engineering, Wollongong, Australia 10 – 13 November 2002.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Wangqing, W., Geming, Z., Kaiming, Z. and Lin, L. (1997). Development of Inspection Car for Measuring Railway Track Elasticity, Proceedings from 6th International Heavy Haul Conference, Cape Town 1997. Werkmeister, S., Dawson, A.R., and Wellner, F. (2001). Permanent deformation behaviour of granular materials and the shakedown theory. Transportation Research Record, TRB, 1757. Witt, S (2008). The Influence of under sleeper pads on railway track dynamics. Report LIUIEI-A—08/00442-SE (unpublished). Linköping University, Sweden Woldringh, R.F. (2001). Notes on specifications for railway embankments in high speed lines. In Geotechnics for Roads, Rail Tracks and Earth Structures (A.G. Correia and H. Brandl, eds.). The Netherlands. pp 262 – 270. A.A. Balkema Publishers. Wu T X and Thompson D J (2000). Influence of random sleeper spacing and ballast stiffness on the vibration behaviour of railway track. Acta Acustica (Stuttgart), Vol 86(2), pp 313321, 2000. Zarembski A M and Palse J (2003). Transitions eliminate impact at crossing. Railway Track and Structures, Vol 99(8), pp 28-30, 2003. Zarembski A M, Palse J W, and Katz L (2001). Reduction of Dynamic Wheel/Rail Impact Forces at Grade Crossings Using Stiffness Transitions. American Society of Mechanical Engineers, Rail Transportation Division (Publication) RTD, v 21, Rail Transportation 2001, 2001, p 97-116 Zarembski, A. M., and Choros, J. (1979). On the Measurement and Calculation of Vertical Track Modulus, Bulletin of the American Railway Engineering Association, Bulletin 675, Volume 81, Nov – Dec 1979. Zhu J Y (2006). On the effect of varying stiffness under the switch rail on the wheel-rail dynamic characteristics of a high-speed turnout. Proc. Instn Mech Engrs, Part F. Journal of Rail and Rapid Transit, Vol 220(F1), pp 69-75, 2006.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

In: Railway Transportation Editor: Nicholas P. Scott

ISBN: 978-1-60692-863-9 © 2009 Nova Science Publishers, Inc.

Chapter 11

ULTRASONIC MONITORING OF THERMAL STRESS IN CONTINUOUS WELDED RAIL D. Vangi1,a and A. Virga2,b Dipartimento di Meccanica e Tecnologie Industriali University of Florence, via S. Marta, 3 - 50139 Firenze, Italy

Abstract

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Ultrasonic systems are since long time used to measure stress on materials, utilizing the acoustoelastic effect. Research has also been carried out in railway engineering, basically to estimate stress on the rails. So far a widely used commercial system, based on ultrasonics or on other methods, does not exist. In this chapter a methodology for monitoring thermally induced loads on continuous welded rails is described, based on the use of ultrasonic waves. The technique allowed an estimation of rail’s neutral temperature and instantaneous longitudinal loads, by means of a new data elaboration method. A complete monitoring system was built and run for about two years on a 3-km track. The method proved to be expensive and time-consuming, if a large amount of railway track is to be monitored, because of the cost of the instrumentation and the necessity of rail adjustment. To find a solution to these problems, a new study was undertaken, aiming to design and test a portable ultrasonic device capable of measuring stress on rails. When using instruments of this kind, several problems must be faced, like those arising from the variation of contact characteristics between ultrasonic probes and rail surface. To solve these problems a double couple probe instrument was designed and built: it proved capable to allow for variation in couplant thickness and type, contact pressure and surface roughness. A calibration with respect to stress and temperature for some of the materials commonly used for rails was carried out. The influence of material’s internal structure and residual stress was also investigated.

a b

E-mail address: [email protected], phone/fax +390554796505 E-mail address: [email protected], phone +390554796297

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

356

D. Vangi and A. Virga

Introduction To improve overall railway track performance, rails are, in most cases, laid out as continuous welded rail (CWR). Its installation consists in welding the rails together to constitute a seamless structure: a rail is defined CWR if it is made of at least about 120m without joints. The main advantages that can be attained (basically due to the elimination of joints) can be resumed as follows: • • • •

reduced track maintenance; reduced trains maintenance; higher traveling speed; improved ride comfort.

On the other hand operating a CWR track requires a careful installation and a constant control of its tensional state, to avoid that extreme temperature can break the rail during cold seasons or cause its buckling (due to elastic instability) during hot seasons, particularly dangerous since its occurrence can easily cause trains’ derailment. Particularly important are the following operations: •

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

• • •

accurate adjustment of internal longitudinal forces during the installation; in this operation a very important parameter is set, called neutral temperature, that is the temperature at which thermally induced loads in the rail are null (zero stress temperature); continuous control of track geometry and ballast conditions; periodical check of longitudinal and transverse loads, especially near critical points (such as curves, bridges, crossings, acceleration and deceleration zones, etc.); repetition of adjustment whenever necessary.

Adjustment is a very important operation in the installation and maintenance of CWR tracks. It basically consists in the following steps: • • • • • •

cutting of the rail if adjusting an existing track; unfastening of the bolts of a sufficient length; removing of anchoring devices and what else prevents the rail from moving freely; in case of adjustment with natural heating, the correct rail temperature is waited for (installation neutral temperature); otherwise a deformation is imposed to the loosened piece of rail, corresponding to the difference between the actual temperature and the desired neutral temperature; welding and fastening of the bolts.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail

357

Sources of Stress in Rails

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Stress in rail can be originated from different sources; first of all, during manufacturing, processes of roller straightening produce residual stress that can reach very high levels (fig. 1); The exact amount and distribution of residual stress is of difficult determination, since it depends on the processes used. What is certain is that integration over the whole cross sectional area yields a null resultant, so that no longitudinal load is produced. Besides it can modify substantially during the life of the rail, because of wear.

Figure 1. Example of residual stress distribution in the cross sectional area for a new rail.

Welding operations during the installation of the rails produce stress as well, but only on local scale, so that no contribution to the longitudinal load is added. The most important source of stress is represented by thermal loads. A free rail undergoing a variation of temperature changes its length proportionally; if it is prevented from free thermal expansion, very high loads are produced. If the anchoring devices (sleepers and fasteners) can be regarded as perfect restraints, any Celsius degree of difference from neutral temperature will produce a stress of about 2.5 MPa, that is, in a UIC60 rail (with a cross section of 7,686 mm²) a load of 18.6 kN/°C. So, if the neutral temperature is known, loads on the rail can be easily estimated by measuring the actual temperature. Unfortunately neutral temperature is not constant, since it can vary, especially in critical points, in a way not easily predictable. This is due to movements of the rails, caused by train passages, uneven temperature distribution, movements of the ballast, etc. From this issues stems the necessity to know what is the actual thermal set-up of the rail, that is what its real neutral temperature is. From the first applications of CWR tracks the necessity of a reliable, non-destructive measurement of thermally induced longitudinal forces has been well known. Evaluating internal axial forces in continuous welded rails can be done by a simple temperature measurement only if no longitudinal or lateral movement has occurred in the track. A

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

358

D. Vangi and A. Virga

destructive method consists in cutting the rail and measure the displacement of its free end. A widely used method consists in checking the movements of the track by engraving two marks on the rail, at a certain distance (e. g. 120 m) and recording their position with respect to some fixed reference along the railway line. If ΔS is the relative displacement (in mm) of the two marks, the change in neutral temperature can be estimated with:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

ΔT = ΔS

83 L

where L is the distance between the control marks. If the marks get closer, neutral temperature lowers. Such method, though simple, cannot reach high precision, also because the references may move as well; moreover lateral movements escape from detection. Several innovative methods have been studied and experimented, some leading to practical application. Techniques based on Barkhausen effect and strain gages have been experimented. Methods based on the use of transverse displacement technique, either in vertical or in lateral directions, have also been investigated. They consist in measuring the force required to lift or move laterally a rail section. Some attempts with ultrasonic techniques were made [1, 2]; such methods are based on the property of ultrasonic waves to change their traveling speed in relation with the stress state (acoustoelastic effect) [5]. Research on several types of ultrasonic waves and measurement direction has been carried out, for example ultrasound with EMAT transducers [6] and with several kinds of polarized waves [2]. However an absolute measurement of internal stress is made quite difficult by the presence of huge residual stress (Figure 1); moreover other aspects play a role, such as material texture, grain size and above all temperature, whose influence over ultrasound speed compares to that of stress. Shear waves with horizontal polarization, with or without EMAT transducers, though theoretically promising, seems to be hardly applicable because of their little acoustoelastic constants and shortness of wave path within the rail, whose consequence is a low sensitivity to stress. The present chapter describes the installation and management of a remotely controlled CWR monitoring system, relying on ultrasonic techniques. It is based on the use of critically refracted longitudinal waves (LCR) propagating along the rail’s longitudinal axis; LCR waves propagates under the surface and are similar to Rayleigh's waves, but with the same speed as longitudinal waves. They have better features, since they have greater acoustoelastic constants in comparison with both Rayleigh's and shear waves, and moreover show low attenuation while propagating in metals. So a longer path can be used (for instance on the rail web, along the longitudinal axis) amplifying time of flight variation induced by acoustoelastic effect.

Ultrasonic Waves Traveling Speed and Its Correlation with Stress Components In the most general case the stress state on the surface of a generic loaded mechanical component is biaxial, and relative change in ultrasound speed is a linear combination of principal stresses. Let us consider longitudinal waves and indicate with VLx and VLy the speed along two orthogonal directions. It can be written:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail

359

VLx − VLx0 = K xσ x + K yσ y VLx0 (1)

VLy − V

0 Ly

VLy0

= K yσ x + K xσ y

where Kx can be defined as the sensitivity to stress of a wave propagating along the stress direction and Ky the sensitivity of a wave propagating in a perpendicular direction. Parameters Kx and Ky are in relation with the acoustoelastic constants. The exponent 0 indicates a zero stress condition. Wave speed measurement involves the exact determination of path length. In practice it is sufficient to consider its traveling time (time of flight, tof, in the following referred as t), that can be directly measured. If the path length is constant equations (1) can be written as follows: 0 t Lx − t Lx = K xσ x + K yσ y t Lx

(2)

t − t Ly 0 Ly

t Ly

= K yσ x + K xσ y

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

For metals Ky is much lower than Kx (about ten times); disregarding its contribution and allowing for the stress along the wave path only:

σ=

1 ⎛ t0 − t ⎞ ⎟ ⎜ K ⎜⎝ t ⎟⎠

If the ultrasonic probes are permanently connected to the part under inspection, their mutual distance varies because of elastic deformation. From (1), under the same hypothesis:

V −V 0 = K (σ − σ 0 ) V0

(3)

Path length can be expressed as:

⎛ σ −σ 0 ⎞ L = L0 ⎜1 + ⎟ E ⎠ ⎝

(4)

where L0 is the initial path length and E Young’s modulus. Since V = L/t and V = L0/t0 using (3) and (4) yields:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

360

D. Vangi and A. Virga

L L0 L0 ⎛ σ − σ 0 ⎞ L0 ⎛ σ − σ 0 ⎞ − ⎜1 + ⎟− ⎜1 + ⎟t0 − t t t0 t ⎝ E ⎠ t0 ⎝ E ⎠ = = = K (σ − σ 0 ) L0 L0 t t0 t0

(5)

and if subscript 0 refers to a null stress condition (σ0 = 0):

⎛ σ⎞ ⎜1 + ⎟t0 − t ⎝ E⎠ = Kσ t Ultrasonic waves traveling speed is also a function of temperature [3]. This property must be accounted for when a variation of temperature is expected. Let us suppose a constant stress state; sensitivity to temperature can be defined as follows:

t 0 (1 + α (T − T0 )) − t = K T (T − T0 ) t

(6)

where α is the coefficient of thermal expansion. If both stress and temperature vary, from (5) and (6):

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

t0 (1 + α (T − T0 ) + (σ − σ 0 ) / E ) − t = K (σ − σ 0 ) + K T (T − T0 ) t

(7)

It must be expected that some of the quantities influencing ultrasonic wave traveling speed have also an influence on stress and thermal sensitivities. So for each material K and KT must be determined. Besides material texture and grain size will affect time of flight [4] and an influence on acoustoelastic constants is also ascertained. A great influence is shown by temperature, comparable to that produced by stress, so that it seems impossible to exploit the acoustoelastic effect without accounting for temperature changes. Moreover the influence of couplant and transducers must be considered. The total wave path includes, though of little extension, segments inside the probes and the couplant; since in such media traveling speed is low the time spent in crossing them can be considerable. If probes and couplant undergo a change of temperature a contribution to tof variation must be expected as well.

Calibration A calibration campaign, aiming to determine coefficients K and KT, was carried out. Several pieces of rail 1.5 m long were placed on a testing machine (MTS 810, full scale 500 kN) where different stress levels were produced and temperature was controlled by means of a climatic chamber, so that stress ranged from – 50 MPa to + 50 MPa and temperature from + 60 °C to –12 °C. Meanwhile the ultrasonic signals were acquired and recorded.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail

361

Figure 2. Rail used for calibration.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

On the pieces of rail (fig. 2) was installed a couple of custom-made, transmitting/receiving ultrasonic transducers, designed and produced on purpose (fig. 3) whose core is a 2 MHz piezoelectric crystal.

Figure 3. Custom-made ultrasonic probe.

The acoustoelastic behavior and the influence of temperature were investigated for two types of material, grade 700 and 900A according to UIC 860 standard. For each signal the relative time of flight with respect to a reference signal was calculated using cross-correlation, and equations (5), (6) and (7) were applied to determine sensitivity to stress and temperature. A preliminary data elaboration has revealed that a slight, but not negligible, dependence of K and KT with temperature is evident. So it was chosen to express them as follows:

K = a + b(T − T0 ) K T = A + B(T − T0 )

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

(8)

362

D. Vangi and A. Virga

so that equation (7) can be rewritten as:

t 0 (1 + α (T − T0 ) + (σ − σ 0 ) / E ) − t = (a + b(T − T0 ))(σ − σ 0 ) + ( A + B(T − T0 ))(T − T0 ) t (9) where subscript 0 indicates reference values. Equation (9) represent a surface correlating stress and temperature with time of flight; for each piece of rail coefficients a, b, A and B were determined with a best-fit method, and by averaging a single surface for each material was obtained (Table 1 and Table 2).

Table 1. Calibration coefficients for rail grade 700. test id 1 2 3 4 mean

a (MPa-1) -2.05E-05 -2.05E-05 -2.03E-05 -2.05E-05 -2.05E-05

b (MPa-1°C-1) -1.07E-08 -7.87E-09 -1.64E-08 -1.15E-08 -1.16E-08

A (°C-1) -1.75E-04 -1.75E-04 -1.70E-04 -1.76E-04 -1.74E-04

B (°C-2) -2.01E-07 -2.62E-07 -2.26E-07 -2.78E-07 -2.42E-07

Table 2. Calibration coefficients for rail grade 900A.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

test id 5 6 7 8 mean

a (MPa-1) -2.04E-05 -2.05E-05 -2.06E-05 -2.00E-05 -2.04E-05

b (MPa-1°C-1) 6.39E-09 -4.17E-09 -1.21E-08 -4.60E-09 -3.62E-09

A (°C-1) -1.13E-04 -1.62E-04 -1.58E-04 -1.64E-04 -1.49E-04

B (°C-2) -2.01E-07 -9.74E-08 -1.65E-07 -9.93E-08 -1.41E-07

Load Measurement in CWR Ultrasonic probes permanently fixed to the rail allow solving several problems; in a generic instant time of flight can be expressed as:

t = t ID + Δt RS + ΔtT + Δt F + Δt TX + Δt CPL

(10)

where: ΔtID time of flight in ideal conditions (no external loads, no residual stress, ideal material); ΔtRS variation of time of flight due to residual stress; ΔtT variation due to temperature change; ΔtF variation due to longitudinal forces; ΔtTX due to the influence of material texture; ΔtCPL due to the influence of couplant.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail

363

In another instant:

t ' = t ID + Δt RS + Δt 'T + Δt ' F + ΔtTX + Δt CPL

(11)

If it can be assumed that the influence of couplant, texture and residual stress is invariable, subtracting equation (10) from (11):

t '−t = Δt = Δt 'T −ΔtT + Δt ' F −Δt F

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

thus only the influence of force and temperature must be allowed for, since any other one is automatically compensated. A remotely controlled monitoring system was built and run for about two years; it was made of 14 sections, each comprising 4 couples of transmitting/receiving ultrasonic probes, like those used during calibration (Figure 3) and with a thermometer added; all the sections were served by a custom-made pulser/receiver device, in bundle with an ADC digitizer provided with RS232 interface. The sections (Figure 4) are linked together and with a remote control station via a 6-wire cable (4 for data and 2 for energy supply). Following a request from the control station the chosen transmitting probe is excited and the ensuing signal from the receiving probe is acquired and digitized at a sampling rate of 20 MS/s and then sent together with temperature. The signal is then cross-correlated to the reference signal and the resulting time of flight is stored. All probes are cyclically excited and the acquisition of data from all the 56 points is completed in about 40 minutes. The whole process is automatically controlled by means of PC software.

Figure 4. Layout of CWR monitoring plant.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

364

D. Vangi and A. Virga

Applying equation (9) to evaluate stress requires the acquisition of a reference signal at a known level of stress and temperature; such reference helps to compensate the errors following the difference between the actual rail material and the one used during the calibration (regarding texture, residual stress, etc.) and the uncertainty about couplant thickness and the effective distance between the transducers. Such signals were acquired during the adjustment, taking advantage of the fact that the rail must be free of moving and unloaded (σ0 = 0); at the same time temperature and tof are acquired and recorded (T0 and t0). If Tn is the neutral temperature of the rail, under the hypothesis that any movement is impeded, thermal longitudinal stress produced by temperature T can be computed as:

σ = Eα (Tn − T )

(12)

During the adjustment neutral temperature is set to the desired value, referred to as adjustment temperature (Tad); as long as neutral temperature is equal to the adjusted one, stress on the rail can be evaluated by using equation (12), with Tad in place of Tn. Unfortunately, as said, neutral temperature usually changes: in such case equation (9) holds true, and stress can be calculated:

σ =

(t − t 0 ) + (T − T0 )[t ( A + B(T − T0 )) − t 0α ] t 0 / E − t (a + b(T − T0 ))

(13)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

with coefficients a, b, A and B obtained by calibration. Figure 5 shows how stress evaluated with equation (13) evolves in an adjusted rail during a period of few days. It is compared with the stress calculated with equation (12), assuming that neutral temperature is still equal to adjustment temperature (Tad = 32 °C in this case). The two plots match enough well, with some discrepancy under particular conditions.

Figure 5. Nominal and measured stress vs. time.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail

Δσ Temperature

 

    

    

 

T (°C)

measured



-

 

nominal

(MPa)



365

  

  

                      

Time (h)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 6. Difference of nominal and measured stress vs. time.

In Figure 6 is plotted the difference σnominal – σmeasured between the stress calculated with eq. (12) and eq. (13), together with temperature. It can be noted that such difference shows the same period as temperature, and that it is greater when temperature changes faster. The error in stress thus evaluated basically stems from the difference in thermal behavior between the rail and the probes and from the fact that the actual volume of material crossed by the ultrasonic wave lies a few millimeters under the surface, while the thermometer is in contact with the rail outside. Such difference is greater in moments when the gradient of temperature is higher and affects the use of both eq. (12) and (13). As an example Figure 7 shows the plots of the temperature acquired in different points of a rail exposed to the weather of a typical summer day from 12PM for 24 hours. The difference between T1 and T2 can give a clue about the thermal gradient in the rail web where the waves propagate.

Figure 7. Temperature in different points of a rail in a typical summer day.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

366

D. Vangi and A. Virga

As a consequence a single instantaneous evaluation of stress can be affected by errors, mainly due to problems in the exact measurement of temperature. Let us now focus our attention on the difference of the actual neutral temperature Tn from the one set during the adjustment (Tad). It seems preferable to appraise the thermal set-up of the rail in terms of neutral temperature rather than in terms of stress, since the latter changes during the day (following temperature) while the former remains constant. Inserting eq. (12) in eq. (13) and rearranging, neutral temperature as a function of temperature and tof can be expressed as:

Tn = T +

Δt + [A + B (T − T0 )]t (T − T0 ) − t 0α (T − T0 ) α [t 0 − (a + b(T − T0 ))Et ]

(14)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 8. Difference between neutral temperature Tn and adjusted temperature Tad.

Figure 8 shows the difference between neutral temperature Tn calculated with eq. (14) and adjusted temperature Tad; variations on a daily basis are clearly visible. As said this is due to problems in measuring temperature. If the variation of time of fight during a whole day is considered (Figure 9) it can be observed that the part of the curve relative to moments with a positive temperature gradient (by daytime in this case) is different from the one relative to moments with a negative gradient (by night). That is to say that the whole curve shows a sort of hysteresis, mainly due to errors in the measurement of temperature. Such observation suggested considering all the values acquired in a whole day, rather than a single acquisition, and led to the use of least-squares lines as a means to minimize errors. Such lines are directly influenced by the thermal set-up of the rail, as can be seen from Figure 10: temperature influences tof variations in a different way before and after the adjustment. Least-squares lines have slope and intercept depending on the relation between longitudinal load and temperature in a given point: to each temperature a time of flight corresponds, as a function of temperature itself and the resulting stress level. If the rail doesn’t move its thermal set-up stays unchanged and day after day the least-squares lines remain quite the same. Observing Figure 11 the following remarks can be made:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail

367

Figure 9. Time of flight vs. temperature during a whole day.

After the adjustment

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Þ Û

Before the adjustment

Figure 10. How adjustment influences tof-temperature relation.

• •

in condition of constant thermal set-up (the rail doesn’t move respect to soil even if temperature changes) the lines maintain rather unchanged both slope and intercept; if a modification in the thermal set-up intervenes (for instance following an adjustment or because the rail moved) to a given temperature corresponds a different stress and a different time of flight; this means that neutral temperature has modified. If, in spite of some movement, hypothesis of perfect restraints still holds, leastsquares lines move parallel to each other, maintaining the same slope, because only a constant contribution is added to the relation between stress and temperature. To a change of neutral temperature corresponds only a change of intercept, while slope remains constant;

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

368

D. Vangi and A. Virga •

if during the day a modification of the thermal set-up intervenes, for instance because some restraints yield and the rail moves as can happen following a strong range of temperature, then the daily Δt-T curve is no longer straight and the least-squares line changes its slope. An example is given in Figure 11: during the procedure of adjustment, when rail anchoring devices are removed, the slope is different, since the relation between temperature and longitudinal load is different.

After the adjustment

Ô

Ñ

Before the adjustment

Ñ

During the adjustment

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 11. Variation of least-squares lines following track movements.

Before the adjustemnt

After the adjustment

Figure 12. Difference between neutral temperature Tn and adjusted temperature Tad.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail

369

Least-squares line’s slope can change also in consequence of particular thermal cycles (for instant in case of strong thermal asymmetry) or for numerical reasons, if temperature range is too little. This offers a way to identify daily thermal cycles that cannot be reliably used to calculate neutral temperature. The difference between the actual neutral temperature and adjustment temperature can be found applying eq. (14) to the least-squares lines, possibly around neutral temperature, where a difference in slope would be less influential. Doing so a few days before and after the adjustment gives the plot shown in Figure 12. In Figure 13 the variation of neutral temperature is shown together with minimum and maximum daily temperature; after a major change due to adjustment (on the 6th day), a little change in neutral temperature can be observed following days with sudden modification of the weather (for instance after 31st and 50th day).

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 13. Influence of thermal range on neutral temperature.

As said a commonly adopted method for controlling CWR consists in the observation of how marks engraved on the rail move respect some fixed reference points alongside the railroad and how they change their mutual distance. By means of the monitoring system just described, the relative movement between transmitting and receiving probes can be calculate as:

Δl = α (Tn − Tad ) ⋅ l

where l is the distance between the probes (about 0.6 m). Let us consider two adjoining control points (A and B) distant L from one another; if it can be supposed that any movement of each intermediate point is linearly distributed, then an average rail movement can be expressed as follows:

⎡ (T − T ) + (Tn − Tad )B ⎤ ΔL = ⎢ n ad A ⎥αL 2 ⎦ ⎣ Figure 14 shows the relative movement calculated for two control points at a distance of 105 m. A comparison with common CWR check methods is thus possible.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

370

D. Vangi and A. Virga

Relative movement (mm)

8 6 4 2 0 -2 -4 -6 -8 0

20

40

60

80

100

Time (d) Figure 14. Average relative movement between control points A and B.

Portable vs. Permanent CWR Monitoring System

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The permanent monitoring system just introduced, together with an appropriate data management procedure, revealed itself capable of a sufficiently reliable control of internal stress on CWR. Ultrasonic transducers permanently fixed to the rail, also through the use of a reference signal, help to overcome problems of couplant variability, material’s difference and uncertainty about the actual distance between the probes. On the other hand the following problems must be underlined: • • •

• •

if a complete monitoring of a railroad line is needed, many control points must be installed; a wired link must be laid down alongside the railroad to connect each control point to the remote control station; alternatively a wireless radio system can be necessary; transducers, cables and electronic devices operate in a rather hostile environment, with strong magnetic fields and exposed to all sort of weather; instruments reliability has proved a delicate issue; to acquire a reference signal, adjusting/de-stressing of the rail can be necessary; if because of metal’s wear internal stress pattern modifies, it is possible that a new reference signal must be acquired.

As a consequence it is probable that the system will become expensive and cumbersome. This considered a portable system or instrument would be very interesting; moreover the possibility of avoiding adjustment should be considered. Following such directions further research has been carried out. However it must be pointed out that, while with fixed probes the influence of couplant, texture and residual stress is invariable, now all this factors must be allowed for.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail

371

The main obstacle to the use of a movable ultrasonic instrumentation based on time of flight measurement lies on the necessity of ensuring a sufficient repeatability of acoustic coupling, that is compensating or allowing for the term ΔtCPL in equation (10). This can be obtained adopting a good and constant surface finish, checking contact pressure between probes and rail and using a couplant with steady acoustic properties and constant thickness.

Figure 15. Double LCR transducer.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Alternatively a double ultrasonic transducer can be used, made of two piezoelectric discs and apt to transmit and receive LCR longitudinal waves (Figure 15). The wave path outside the plexiglas structure is very short and crosses twice the couplant layer. The slot prevents the wave from shortcutting through the plexiglas. Two probes like this were applied at both ends of an invar alloy bar at a distance of 0.5 m; invar alloy, having a coefficient of thermal expansion lower than rail steel (14·10-7 °C-1), offers greater dimensional stability. The instrument thus obtained is applied longitudinally to the rail web (Figure 16).

Figure 16. Portable device fitting to the rail.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

372

D. Vangi and A. Virga

Figure 17. Different wave paths acquired with the portable device.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Three different signals have to be acquired, relevant to three different paths (Figure 17): • • •

path 1, through plexiglas and couplant around the first probe; path 2, through plexiglas and couplant around the second probe; path 3, through plexiglas, couplant and metal, from the first probe to the second.

If the two secondary paths (1 and 2) can be assumed to be symmetrical, then net time of flight can be calculated as:

t net = t path 3 −

t path1 + t path 2 2

(15)

Such a device is intended for the achievement of the following features: • • •

automatically allowing for couplant type and quantity, thus compensating contact pressure as well; automatically allowing for contact surface roughness; eliminating temperature induced time of flight variations on probes and couplant.

Several test sessions were carried out to verify operation and assess reliability. Figure 18 shows results about compensating the influence of contact pressure. The signals relevant to Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail

373

the three paths described in Figure 17 were acquired while increasing contact force. Signals were cross-correlated to the first one that, for this reason, has tof variations always equal to zero. While contact force increases the layer of couplant gets thinner and tof relative to path 1, 2 and 3 decreases, while net time of flight, calculated with eq. (15) remains rather unchanged. Figure 19 describes how the device behaves as to positioning. Each time the probes were put in contact with the surface after having substituted the couplant: partial paths show a great variability, whereas net path is little affected. Satisfactory results were also obtained as regards couplant type and surface roughness.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 18. Compensation of effect of contact force variability.

Figure 19. Some repeatability tests.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

374

D. Vangi and A. Virga

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

A new calibration was carried out with the purpose of determining sensitivity to stress and to temperature for some rail material; since the distance between the two probes can be considered constant the following equations must be used in place of, respectively, eq. (6) and eq, (5):

t0 − t = K T (T − T0 ) t

(16)

t0 − t = K (σ − σ 0 ) t

(17)

Figure 20. Sensitivity to temperature for a grade 900 rail obtained with the portable device.

Figure 21. Sensitivity to temperature for a grade 900 rail obtained with fixed probes.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail

375

As an example in Figures 20 and 21 is shown coefficient KT obtained as slope of leastsquares line in the two cases for a 900 rail free from external loads; the difference is mainly due to the fact that with the portable device temperature effects on the probes are automatically allowed for. In Figures 22 and 23 sensitivity to stress K is compared; the slight difference is due to a difference in wave path length (in fact for fixed probes path inside plexiglas contributes to total tof). One of the features of the portable instrument is that it allows an absolute measurement of LCR wave speed. If for a given material the acoustoelastic behavior is known, then an absolute measurement of stress is possible. In fact eq. 3 can be rearranged as follows:

V = kσ + V0

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

with constant k and V0 function of material, texture and temperature. Such measurement includes also residual stress. As an example in Table 3 absolute time of flight for two pieces of rail of the same nominal material (900) is shown for a series of tests; the pieces have the same temperature and no external load. A difference of about 120 ns is obtained, and if it would be imputed to a difference in residual stress only, applying eq. (17) to rail B with rail A as a reference (with K = -1.465·10-5 MPa-1) a Δσ of about 95 MPa is obtained. Since the pieces are made of the same nominal material but don’t belong to the same lot such difference must be imputed also to texture.

Figure 22. Sensitivity to stress for a grade 900 rail obtained with the portable device.

A confirmation of the effect of residual stress and material texture is found examining different parts of the same rail (Table 4). Rather high tof differences are obtained just moving in different points of the cross section of the same rail (for instance moving from foot to head), so that to use an absolute measurement of tof a calibration for each zone should be done.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

376

D. Vangi and A. Virga

Figure 23. Sensitivity to stress for a grade 900 rail obtained with fixed probes.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Table 3. Comparison of absolute time of flight between two rails of the same nominal material.

test id 1 2 3 4 5 mean

Rail A Rail B t (ns) 87844.0 87963.0 87841.0 87965.0 87839.0 87964.5 87845.0 87968.0 87844.5 87968.0 87842.7 87965.7

Difference (ns) 119.0 124.0 125.5 123.0 123.5 123.0

If a complete characterization of residual stress and texture is not possible, a differential measurement can nevertheless be made, by applying eq. (17). The meaning of eq. (17) is that if a reference signal is acquired at a known stress level, then stress in a generic condition can be obtained by measuring time of flight, constant K being function only of material. As a conclusion it seems quite difficult to perform an absolute measurement of external loads on a rail unless a complete characterization of material’s behavior as to ultrasonic speed is obtained. Otherwise a reference value of tof is needed in condition of known external load. For CWR load measurement this implies that an operation of adjustment can hardly be spared, unless an independent way of obtaining a single reference load measurement is available. However if a reference condition cannot be recorded the method is nonetheless suitable for monitoring the stability of the rail; in fact a reference signal can be acquired, even in a condition that has an unknown neutral temperature. Such signal will be used as a starting point for subsequent monitoring: any major change or tendency to change of neutral

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Ultrasonic Monitoring of Thermal Stress in Continuous Welded Rail

377

temperature will produce a variation of the relationship between tof and temperature, suggesting some intervention or a new adjustment.

Table 4. Comparison of absolute time of flight measured in different zones of the same rail. Web test id 1 2 3 4 5 mean

87844.0 87839.5 87850.0 87853.0 87852.0 87847.7

Foot t (ns) 88557.5 88556.0 88550.0 88554.5 88556.5 88554.9

Head 88232.5 88221.5 88222.0 88237.5 88224.0 88227.5

Difference (ns) Web/Foot Web/Head 713.5 388.5 716.5 382.0 700.0 372.0 701.5 384.5 704.5 372.0 707.2 379.8

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Conclusion Continuous welded rail (CWR) represents an important means for improving overall railway track performance by eliminating joints, thus allowing faster and more comfortable traveling together with reduced maintenance. On the other hand its reliability and safety depend on the accuracy with which it is installed and maintained. One of the more challenging goals regarding CWR is the possibility of measuring actual thermal induced loads, since the relationship between temperature and internal load is subject to change. Several techniques have been attempted, and among the others some based on ultrasonics. This chapter presented some experimental work regarding installation and operation of a permanent CWR monitoring system, based on the use of LCR ultrasonic waves; the plant, with custom ultrasonic probes and instrumentation, have been running for about 2 years on a 3-km railroad; it was remotely controlled and allowed, also through a proper daily data elaboration, the estimation of neutral temperature, from which an estimation of internal load is possible. Such monitoring system revealed to be rather expensive and its instrumentation subject to failures, being installed in a hostile environment. A portable instrument was then designed and tested, resulting in a prototypal device provided with two couple of LCR ultrasonic probes; it proved capable of allowing for uncertainty in contact pressure, couplant variability and surface roughness, assuring a satisfactory repeatability. Calibration respect some common rail materials was also carried out.

References [1] Bray, D. E.; Egle, D. M. Application of the acousto-elastic effect to rail stress measurement, Mater Eval 1979, 37 (4), 41-46, 55. [2] Szelazek J. (1998). Monitoring of thermal stresses in continuously welded rails with ultrasonic technique. NDTnet, 3 (6). [3] Salama, K. Relationship between temperature dependence of ultrasonic velocity and stress. Rev Prog Quant Nondestruct Eval, NDT Int 1984, 4, 1109-1119.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

378

D. Vangi and A. Virga

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

[4] Szilard, J. Examining the grain structure of metals, In Ultrasonic testing; Szilard, J.; John Wiley & Sons: New York, NY, 1982; pp 217-261. [5] Luthi, T. Determination of biaxial and triaxial stress distribution using ultrasonics, NDT Int 1990, 23 (6), 351-356. [6] MacLauchlan, D. T.; Burns, L. R.; Alers, G. A. Measurement of stress in steel structures with SH waves EMATS. Rev Prog Quant Nondestruct Eval, NDT Int 1987, 7B, 13991404.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

In: Railway Transportation Editor: Nicholas P. Scott

ISBN: 978-1-60692-863-9 © 2009 Nova Science Publishers, Inc.

Expert Commentary

EDMOND FORTIER AND THE LAGOS TRAM (1908) Liora Bigon1 and Frank Brown2 1

2

Holon Institute of Technology, Israel The University of Manchester, Great Britain

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Abstract The French photographer Edmond Fortier, who lived and worked in Dakar (Senegal) from the 1900s, produced more than 3,000 pictures/postcards during his extensive trips in French West Africa. While his photographs are celebrated today in scholarly and commercial circles, his 1908 series of twenty-two Lagos pictures has not so far attracted the attention it deserves. Partly representing views from the Lagos steam tram and partly depicting the tramline as a background feature, this photographic series constitutes one of the rare pieces of contemporary evidence that relates to the tram. Operating between 1902 and 1933, the history of this line has hardly been researched as well. This commentary reflects on the historiography of these two relatively unknown episodes in colonial West Africa: that of Fortier’s visit in Lagos and the subsequent series of photographs; and that of the Lagos steam tramway which served as their background.

On the Lagos Tram The first main railway lines in West Africa were constructed during the establishment of the official colonial regimes there from the late nineteenth century – following the Berlin Conference (1884-5) and the imperial ‘scramble’ for Africa – until World War I. This was the time when national rivalry between the European countries was projected into their colonial territories in Africa. This meant, however, that direct railway connections between two adjacent territories did not normally exist in West Africa, where each of these territories was governed by a rival colonial regime. Considering Fortier’s visit, the passage referred to was thus rather of metaphoric or representational nature, and concerned a single railway: the Lagos steam tramway.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

380

Liora Bigon and Frank Brown

The Nigerian Railway (connecting Lagos with the hinterland of Ibadan by 1901, as far as Sahelian Kano by 1912) constituted the raison d’être of the Lagos Tram. Yet, what was unusual about this tram in contrast to the history of railway transport in colonial West Africa, is that it served as a municipal line, running within the four-square-miles of Lagos Island. Its main purpose, beyond creating a link for merchandise between Lagos Island’s wharves and the Nigerian Railway terminus on the mainland, was to convey passengers and travellers on the Island before the advent of wheeled road transport [1]. In colonial West Africa in general, railway projects were not normally conceived as passenger lines, circulating within a relatively small area to serve its European and African residents. Their main purpose – in this part of the continent where white-settler colonies were never established due to climatic conditions – was rather the safe and speedy conveyance of raw materials to the ports for export. Only in the 1930s did road transport compete with that of the railway, and by this period roads in West Africa were mainly designated to feed the railway [2]. Having commenced his service in Lagos (Nigeria) as a municipal engineer in 1949, N.S. Miller published his short enquiry on the Lagos Tramway within a decade [3]. This modest work, which was carried out in his spare time, constitutes a rare historical source, particularly as all the original files concerning this tram at the Nigerian Railway Corporation were destroyed in the summer of 1956. Guided by his “boyhood enthusiasm for things mechanical”, some rumours about the line, surviving pieces of the track itself and some archival records kept by British Crown Agents, Miller created a vivid historical portrait of the Lagos tram, operating between 1902 and 1933. In addition to this material, Miller’s work is loaded with technical and operating details about the locomotives and each engine, along with the rolling stock, the number of passengers and the distance between stations and tariffs. Most of the figures also consist of tram-engine models, line drawings and visual reconstructions. The actual three-coach tram appears in one figure only: a contemporary photograph shows it coming from the mainland towards the Island over the old Carter bridge, the first bridge in Lagos. It remains only to point out the revival of this yet uncovered issue in the history of colonial Nigeria, this time from the perspective of cultural history [4]. The special position of Lagos as a Crown Colony and a chief lieu de colonisation in British West Africa from the 1850s provides us with several directions of exploration that are impossible elsewhere in the region. Following the pax Britannia by the end of the nineteenth century, Lagos comprised a preferred place for the resettlement of a few thousand emancipated slaves mostly from Brazil, Cuba, Bahia and Sierra Leone. These people, who were mostly of a Yoruba origin and locally named ‘Saro’ and ‘Amaro’, were involved in trade, business and in colonial administration and became a local elite group, economically and intellectually [5]. Amongst them there were popular figures that used the tramcar, for pleasure or either for taking local children for joyrides and paying their fares. Many members of this bourgeois group were the first to possess Western objects, including cameras. As many of their descendants still live on Lagos Island and play a key role in regional politics, it is not impossible that complementary evidence regarding this tram may be found in the form of stories, memoirs and family photographs. Local newspapers founded in Lagos from the 1860s by the Saro elite were used as a platform for discussion and critique of town planning, technical and innovative issues. This source of information could be enlightening. To balance the colonial points of view, there is a need for further work drawing on primary and photographic evidence. This kind of evidence is probably still extant in the archives of British

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Edmond Fortier and the Lagos Tram (1908)

381

firms whose headquarters are in some cases still based in London, Leeds, Liverpool and Manchester.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

On Edmond Fortier Of all colonial photograph producers and cliché-makers in both French and British West Africa, Edmond Fortier is the most famous. Born in 1862 in the Vosges, France, he had already spent several years in Saint-Louis in Senegal before his departure in 1900 to Dakar, the capital of the federation of French West Africa. As a photographer and a publisher of postcards, his extensive trips in French West Africa (Senegal, Guinea, Mali, Ivory Coast and Benin) yielded more than three thousand prints before 1910, and a similar number of reprints in the ensuring period [6]. Fortier’s postcards, which can be found in private and archival collections all over the world, and are sold through the internet, could be classified into a few main themes. Noticeable amongst them, for instance, are: ethnic and racial types; representations of women; scenes of social and working life; and views of places. Indeed, these postcards must be understood within the genre of colonial photography as products of certain historical, cultural and ideological conditions. As such, documentation of indigenous cultures tended to stress their ‘bizarre’ or ‘barbaric’ aspects, a process of ‘othering’ that assisted in defining the coloniser’s ‘civilised’ ‘self’/landscapes [7]. There are more questions than answers regarding Fortier’s visit to Lagos in 1908, apparently his sole visit to a British colonial sphere. What were the circumstances of his visit? Was he asked to accompany French officials? For how long? The result, as we know, was a series of twenty-two postcards. The subject of this series is contemporary urban landscapes from the Island, each photograph being marked by the street, area or quarter in Lagos where it was taken. There are two facts of interest here. First, the exact location from which Fortier took his picture corresponds with the tram stations in Lagos as they appear on a map reconstructed by Miller. Second, the tramway is shown in the background and the foreground of many of these pictures. Though Fortier’s journey from Dakar to Lagos was most probably by ship rather than by train, it yielded a semi-official body of evidence (in the form of postcards) that delineates relations between a French and a British territory. Each territory was situated at the furthest point of what is defined as ‘West Africa’. Their capital cities (Dakar/Lagos) served as the main ports and competed as the most important places in each colony. Fortier’s visual testimony is also outstanding because photographs of the Lagos tram are relatively rare. Most of the photographs that were taken by British officials and employees in colonial Lagos depict their own residential spheres and gardens, with a preoccupation with the variety of flora. Photographs of the public sphere of the expatriate community, such as the Government House and gardens, the Marina area and the Racecourse area, were also extremely popular amongst these officials and employees [8]. This background makes Fortier’s Lagos series even more distinctive, as he did not ignore the streetscapes that were viewed from those stations that were situated within the indigenous living areas of the Island [Figure 1] – e.g. Igbosere Road, Acharawu Street, Docemo Street and Massey Street.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

382

Liora Bigon and Frank Brown

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 1. Fortier’s postcard showing Acharawu Street, Lagos, 1908 (private collection).

Figure 2. Fortier’s postcard showing the station in Abidjan, the 1910s. (Courtsey of Archives nationales du Sénégal).

Yet, Fortier’s Lagosian series is less distinctive compared to his other photographs of railway stations within urban areas in French West Africa, such as that of the station in Abidjan, Ivory Coast [Figure 2]. In both cases, scenes from everyday urban life of the majority of the African residents were captured along the tramlines. These were non-staged scenes, in contrast to Fortier’s famous African ‘types’ – a colonialist genre per se. Indeed, a comparison between the semi-official evidence of Lugard’s photographs from the 1910s that

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Edmond Fortier and the Lagos Tram (1908)

383

depict intercity railway transport in British West Africa and those of Fortier taken in the French territories shows similar features [9]. In both cases, a recurring scene is that of the terra incognita of the ‘bush’ that was crossed by a tramline, or, preferably, of a train crossing a bridge on the savanna’s background. [Figure 3] In other words, the imperialist imagery in both cases was inflamed by the very presence of the railroad, symbolising the genius of Western engineering projects, the civilising mission and the mise en valeur of the African terrains [10].

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 3. Fortier’s postcard showing the Kayes-Niger line’s passage over Badinko Bridge in presentday Mali, the 1910s. (Courtsey of Archives nationales du Sénégal).

In the academic literature, however, the images of Fortier are usually discussed in the context of the history of Dakar and colonial photography, orientalism and the influence of his African nudes on Picasso’s primitivism. These two interwoven historical episodes – Fortier’s 1908 visit in Lagos and the Lagos Tram itself – remain unclear. This invites the launching of a new research programme which will focus on the cultural dimensions in the history of railway projects in colonial West Africa (as well as elsewhere in the colonial world), combining colonial photography and railway historiography.

References [1] In its final phase this tram served as a sanitary line, removing night-soil in pails towards ‘Dejection Jetty’ in the southeastern edge of the Island – a function which is also unusual with regard to the function of other colonial lines in tropical Africa. [2] It is too obvious to recount here that the provision of modern transport, particularly the railway, was primarily aimed at the expansion of the economic ambitions of the European métropole in West Africa. This is true in terms of trade and commerce and the development and exploitation of both agricultural (cash crops) and mineral

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

384

[3] [4] [5]

[6]

[7]

[8]

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

[9]

[10]

Liora Bigon and Frank Brown resources there from the late nineteenth century. Before the 1930’s, however, the colonial authorities in Nigeria took active measures to suppress road transport, especially amongst the Africans, but after this time road construction was revised at the expense of the railway due to a growth of primary production. See, for instance, Headrick, D. (1981). The tools of Empire: Technology and European Imperialism in the nineteenth century. New York: Oxford University Press, pp. 192-204; Home, R. K. (1974). The influence of colonial government upon urbanisation in Nigeria (unpublished Ph.D. thesis, University of London), pp. 100-109 Miller, N. S. (1958). Lagos steam tramway, 1902-1933. Lagos: n.p. See, for more, Bigon, L. (2007), Tracking ethno-cultural differences: the Lagos steam tramway, 1902-1933, Journal of historical geography, 33, 596-618. For more about the history of the ‘Amaro’ and the ‘Saro’ groups in Lagos see: Cole, P. D. (1975). Lagos society in the nineteenth century. In A. B. Aderibigbe (Ed.), Lagos: the development of an African city (pp. 27-58). London: Longman; Mann, K. (1985). Marrying well: marriage, status, and social change among the educated elite in colonial Lagos. Cambridge: Cambridge University Press. A few hundred of Fortier’s postcards are found in the collections of the Archives nationales du Sénégal in Dakar, and in the Centre des archives d’outre-mer in Aixen-Provence. See also http://home.wxs.nl/~kreke003/ homeng.htm For more about the genre of colonial photography and its possible classifications with relation to French West Africa see Prochaska, D. (1991). Fantasia of the photothèque: French Postcard Views of Colonial Senegal. African arts, 24, 40-47. This is clear, for instance, from the examination of the relevant photographic collections of the Royal Commonwealth Society of the University of Cambridge – most of the photos were given as contributions by the families of these employees. Frederick J.D. Lugard (1858-1945) served as the first high commissioner of the protectorate of Northern Nigeria (1900-1906) and the governor general of Northern and Southern Nigeria (1912-1919). While sections of the Nigerian Railway were documented by him, the Lagos Tram is virtually absent from his photographic collection (placed in Rhodes House, Oxford). While there is research on the symbolic dimensions of railway transport in French West Africa (see for instance, Conklin, A. (1997). A Mission to civilize: The Republican idea of Empire in France and West Africa, 1895-1930. Stanford: Stanford University Press, pp. 38-72), equivalent research on this phenomenon in British West Africa is rather meager.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

In: Railway Transportation Editor: Nicholas P. Scott

ISBN: 978-1-60692-863-9 © 2009 Nova Science Publishers, Inc.

Short Communications A

SECURITY RISK MANAGEMENT OF RAILWAY TRANSPORTATION SYSTEMS Francesco Flammini1,2,a and Nicola Mazzocca2,b 1

ANSALDO STS - Ansaldo Segnalamento Ferroviario S.p.A. Via Nuova delle Brecce 260, Naples, Italy 2 Università di Napoli “Federico II”, Dipartimento di Informatica e Sistemistica Via Claudio 21, Naples, Italy

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Abstract Railway operators and suppliers can count on well established techniques for safety assurance. However, the recent terrorist strikes have shown that security is also an important issue to be addressed by system engineers. While safety refers to the possible hazardous effect of the system upon the external environment, security studies the effect of the external environment upon the system, including natural as well as malicious threats. Terrorists are not the only adversaries: vandals, thieves and other perpetrators use to attack mass transit rail based system on a daily rate. Procedures and protection mechanisms are therefore needed for the safeguard of the infrastructure against external threats. This paper describes a systemic approach to the security risk management of railway infrastructures. The analysis addresses both methodological and technological means, showing that while a correspondence can be found between safety and security taxonomies, novel approaches are needed to cope with security specific issues.

Keywords: Security, Risk Analysis, Critical Infrastructure Protection, Railway.

1. Introduction Throughout the years, we have assisted to a continuous development of the techniques used for reliability and safety assurance of high-integrity computer systems, including railway control devices. Rigorous standards and guidelines exist to assist engineers in providing the a b

E-mail address: [email protected]; [email protected]. E-mail address: [email protected]

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

386

Francesco Flammini and Nicola Mazzocca

demonstration of RAMS (Reliability Availability Maintainability Safety) properties (see e.g. [[18]]). However, recent events have caused a sort of paradigm shift in the research community working on critical systems: while engineers felt confident about their ability of demonstrating safety properties, they were not yet prepared to provide a formal (and possible quantitative) evidence of system security. Therefore, the attention of many researchers has moved from the evaluation of the negative effects of the system on the external environment (safety) to the effects of the external environment on the system (security). It is evident that security has also a safety impact: malicious attacks (e.g. terrorist strikes) can be aimed at disrupt the infrastructure or control systems in order to cause a catastrophic failure, that is a failure with possible consequences on human beings. Several attempts have been made in the recent years to provide a parallelism between safety and security taxonomy and evaluation approaches [[20]][[21]]. However, these attributes should be better framed into the dependability taxonomy [[22]], which does not leave any ambiguity on the meaning of the involved terms. As an example,

Figure 42 reports the definition of threats according to such taxonomy. Physical security involves natural or human-made external faults. Figure 43 provides a reference context diagram showing the interactions of the control system with external entities which can act as sources of random as well as intentional threats.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 42. Threat definition according to dependability taxonomy.

Figure 43. Context diagram showing external dependability threats for a railway control system.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Security Risk Management of Railway Transportation Systems

387

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Dependability attributes of systems can be predicted using several more or less formal approaches, including model-based techniques, which can be either analytic (e.g. fault trees) or simulative (e.g. fault injection) and provide qualitative as well as quantitative results. Unfortunately, security attributes can not be evaluated using the same formalism or models adopted for RAMS evaluation. For instance, the widespread techniques of Reliability Block Diagrams (bottom-up approach) and Fault Trees (top-down approach) are not suitable to model dependant basic events, which are common in security related scenarios (e.g. simultaneous exploitation of flaws). Therefore, different approaches are needed for security analyses [[21]]. In this paper we provide the description of a methodology which can be adopted to perform a security risk management of railway infrastructures from a systemic point of view. Risk Management (sometimes indicated with Risk Analysis) includes Risk Assessment (i.e. evaluation of current risk indices) and Risk Mitigation (i.e. reduction of risk acting on protection mechanisms), in a possible iterative process. The method is multi-disciplinary and involves data coming from experts of different fields, above all statistics and sociology. Therefore, we will concentrate on the core aspects of the approach. Details of specific points which are not addressed in this paper can be found in the referenced works, especially in [[17]]. The rest of this paper is organized as follows. Section 2 describes a preliminary qualitative analysis which is needed for the subsequent quantitative steps; Section 3 presents the risk assessment methodology and the reference models used for the analysis; Section 4 describes the evaluation process for risk mitigation, including the tool-based support for the design of protection mechanism; Section 5 provides an example application of quantitative risk analysis; finally, Section 6 draws conclusions and provides some hints about future developments.

2. Preliminary Security Analysis In this section we present some organizational aspects related to the risk management process, which are important for the success in the application of the methodology. In the first phase, some relevant points should be clarified, by answering the following qualitative questions (divided into main categories): - Objectives • Which are the threats I need to protect against? This is important to screen all the external faults (e.g. natural disasters) which for any reason do not need to be accounted in the analysis. • Which are the boundaries of the systems I need to protect? This allows to understand where does the system end and the “external environment” begins. - People • Who are the professionals needed to perform the analysis?

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

388

Francesco Flammini and Nicola Mazzocca

•

It is important to group all the people who are necessary to perform the analysis (experts in the railway domain, system operators, designers/developers, psychologists, etc.). How do I proceed to the analysis? The analysis can not be performed by a single person, but it needs (at least for some phases) to unite all the experts around a table in order to collect and compare their judgments in brainstorming sessions.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

- Documentation • How is the system made? Analysts need to know both the physical and logical architecture of the system, including access points for public and personnel, technical rooms, current protections, functional dependencies, geographical locations of assets, etc. • What is the value of the assets included in the system? This data allows to answer the questions related to the cost needed to substitute an assets after a damage (a budgetary estimation can be useful). • What is the frequency of common threats? Historical data related to the occurrence of micro-criminality (aggressions, thefts, vandalism, etc.) and the provided damage (money needed to restore the functionality, effect on service operation, etc.) is useful, if available (data related to similar systems can be used as well if a strong analogy can be demonstrated). - Surveys • Which I need to watch and take pictures of? Regardless of their detail, documents are not enough to perform risk analysis. Site surveys are required to collect data of the actual condition of the installations and how they are managed. • How I collect data from site surveys? Checklists can be very helpful in performing site surveys. Each point in the checklist corresponds to an aspect to take into account, including interviews to service personnel. Sometimes, sites are not easily accessible. In these cases, other means can be adopted (e.g. satellite maps). - Tools • What do I need to perform the analysis? Analysts should take note in advance of everything they need to perform the analysis, including know-how (e.g. railway norms and regulations) and tools (e.g. instruments for site surveys, software to automate some steps of the analysis, etc.). After the above points have been clarified, the analysis can proceed to the following step: risk assessment.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Security Risk Management of Railway Transportation Systems

389

Figure 44. Example information sources for security analyses: pictures, design lay-outs, maps.

3. Security Risk Assessment With reference to a specific threat, the quantitative risk R can be formally defined as follows:

R = P ⋅V ⋅ D .

(1)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Where: • P is the frequency of occurrence of the threat, which can be measured in [events / year]; • V is the vulnerability of the system with respect to the threat, that is to say the probability that the threat will cause the expected consequences (damage); • D is an estimate of the measure of the expected damage occurring after a successful attack, which can be expressed in euros [€]. The vulnerability V is an adimensional parameter, since it represents the conditional probability:

P ( success | threat ) .

(2)

Therefore, a quantitative way to express the risk associated to a specific threat is to measure it in lost euros per year: [€ / year]. The overall risk can be obtained as the sum of the risks associated to all threats. Despite of the simplicity of (1), the involved parameters are not easy to obtain. The analysis involves both procedural and modeling aspects. Procedural aspects include brainstorming sessions, site surveys, design review, statistic data analysis, expert judgment, etc. Formal modeling languages which can be used to analytically compute P, V and D include Attack Trees, Bayesian Networks, Stochastic Petri Nets and possibly other formalisms which are able to take into account the uncertainty inherently associated to the risk as well as the possibility of strategic attacks [[7]]. In fact, the three parameters feature an inter-dependence which should be modeled, too. A possible starting point for the risk assessment is a database of threats, characterized by:

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

390

Francesco Flammini and Nicola Mazzocca -

Threat identifier; Short description of the attack scenario (including the adversary category, required tools, etc.); Threat category (e.g. vandalism, theft, sabotage, terrorism, flooding, etc.); Site (geographical reference).

Similarly it is possible to define databases for assets, including their relevant attributes, like values (i.e. cost to restore) and interdependencies. In the security analysis, it is important to consider physical as well as logical threats, since they are both important and strictly correlated. In fact, a physical intrusion can facilitate an attack to information integrity (e.g. by accessing an unsupervised client), while a logical attack (e.g. by remotely exploiting a flaw in the computer-based train control system) can cause a safety incident. A model-based estimation of P, V and D can be based on a set of reference models as described in the following.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

3.1. Reference System Models The first models which are needed are the ones representing the aspects of interests of the system under analysis. These models will be helpful in determining the most relevant parameters in risk evaluation (frequency, vulnerability and consequences). Models can be more or less formal. For instance, a schematic block-model of a station is a completely informal model. Let us start from the estimation of the P parameter. Whereas statistics can be extracted from historical series of similar events, the estimation is straightforward. When this is not possible, P should be obtained as a correlation of the accessibility of the attack and of its expected consequences (e.g. value of the asset to be stolen in case of theft, predicted casualties in case of a terror strike), possibly using a Bayesian estimation (see below). This serves as a sort of “attractivity” index. The estimation of vulnerability V is a more difficult task. It should account for the obstacles (e.g. physical barriers) encountered by the adversary in reaching its objective and the ones preventing protections mechanisms to early detect the threat (e.g. ground morphology impacting on sight distance). The adversary succeeds in its attack when the time to reach the objective and perform the attack is less than the time needed for surveillance and protection mechanisms to react and defeat the attacker (see Figure 45). Hence, it is difficult to provide a general model for vulnerability evaluation, but the formalism of Timed and Stochastic Petri Nets can be advantageously adopted to model time-related dynamics of security threats also accounting for the uncertainty of some parameters. For the estimation of consequences (D parameter), we propose here a semi-formal model based on the most important view of the Unified Modeling Language (UML): Class Diagrams [[23]]. Class Diagrams allow for a representation of the (physical or logical) entities of the system under analysis, including all their relationships (e.g. physical distance, functional dependencies, hierarchical aggregation, etc.). An example UML Class Diagram for ERTMS/ETCS (European Railway Traffic Management System / European Train Control System) is reported in Figure 46.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Security Risk Management of Railway Transportation Systems

Deter

Detection and Verification

391

Delay The Response Cycle

Detection Point

Critical Response Point

Figure 45. Evaluation of Vulnerability.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

A combination of Fault and Event Trees can be adopted to model risk, according to the scheme reported in Figure 47, which needs to be replicated for each asset, each failure mode and each attack scenario (physical or logical). Such model can be automatically derived from a class diagram like the one reported in Figure 46, whenever quantitative attributes are employed. In order to use a single modeling formalism, Bayesian Networks can be adopted, which also ease the sensitivity analysis to evaluate the effects of parameters’ uncertainty. The extensions for Decision Networks can be advantageously employed in order to evaluate the Risk indices and perform cost-benefit analyses.

Figure 46. An example UML Class Diagram for a railway system compliant to ERTMS/ETCS.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

392

Francesco Flammini and Nicola Mazzocca

Other assets' attractivity

Attractivity Likelihood of attack

Intrinsic robustness Accessibility

Aggregated asset failure Existing protections

Asset failure Dependant asset failure

Component asset failure

Influencing asset failure Event Tree Fault Tree

Figure 47.A modeling scheme for quantitative vulnerability assessment.

4. Security Risk Mitigation

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Protection mechanisms are able to reduce the risk by having three main effects: • Protective, aimed at the reduction of V • Deterrent, aimed at the reduction of P • Rationalizing, aimed at the reduction of D Therefore, by quantifying the listed effects it is possible to estimate the risk mitigation, considering any combination of threats and protection mechanisms. A possible way to compute risk mitigation is to associate threats and protection mechanisms by means of threat categories and geographical references, namely “sites”. A site can be considered as a particular kind of critical asset (actually, an aggregate asset), sometimes defined as “risk entity”. Each threat happens in at least one site and, homogonously, each protection mechanism protects at least one site. For a railway infrastructure, a site can be an office, a bridge, a tunnel, a parking area, a platform, a control room, etc. In the assumption that: • • • •

Threat T belongs to category C; Threat T happens in (or passes through) site S; Protection M is installed in site S; Protection M is effective on threat category C;

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Security Risk Management of Railway Transportation Systems

393

then it can be affirmed that M protects against T. Basing on the above definitions, it is possible to express the overall risk to which the system is exposed as follows:

RT = ∑ Ri ⋅ ∏ (1 − EPji ⋅ COV j ) ⋅ (1 − EDji ⋅ COV j ) ⋅ (1 − ERji ⋅ COV j ) . i

(3)

j

Where: • RT is the total mitigated risk; •

Ri is the initial risk associated to threat i (computed according to (1));

•

EPji is an estimate of the protective effect of mechanism j on threat i;

•

EDji is an estimate of the deterrent effect of mechanism j on threat i;

•

ERji is an estimate of the rationalizing effect of mechanism j on threat i;

•

COV ji is a measure of the coverage of mechanism j (e.g. percentage of the physical area or perimeter of the site).

The values of parameters expressing coverage and effectiveness are in the range [0..1]. The formula can be validated by attempts using sample data and boundary analysis: for instance, when both the coverage and one of the effectiveness parameters are set to 1, the risk is mitigated to 0, as expected; on the opposite, if either the coverage or all the effectiveness parameters are set to 0, the risk is not mitigated at all. The cost/benefit index can be defined simply as the balance between the investment on security mechanisms and the achieved risk mitigation:

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

EB = risk reduction − total investment in security = (R T − ∑ Ri ) − ∑ C j . i

(4)

j

Where: • EB is the Expected Benefit, which can be positive or negative; • C j is the cost of the protection mechanism j, obtained considering all the significant costs (acquisition, installation, management, maintenance, etc.). Therefore, the return on investment can be obtained from the expected benefit EB considering the cost of the invested capital (which depends on the rate of interest, the years to pay-off, possible external funding, etc.). Expressions (3) and (4) need to be computed starting from a database of attack scenarios, sites, protection mechanisms and related significant attributes. The management of such data and the computation of results are performed by an automatic tool which will be described in detail in next section.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

394

Francesco Flammini and Nicola Mazzocca

4.1. Evaluating the Effectiveness of Protection Mechanisms A tool has been developed which automatically manages risk data and evaluates risk and benefit indices starting from input data. The tool has been named simply Q-RA (Quantitative Risk Analysis), to be pronounced as [kura] (sounding like the Italian for “cure”). •

A list of protection mechanisms, characterized by: - Protection mechanism identifier; - Short description of the mechanism; - List of threat categories on which the mechanism is effective; - Expected protective ( EPji ), deterrent ( EDji ) and rationalizing ( ERji ) -

effectiveness; Estimated coverage (COV); Site (geographical reference); Annual cost (acquisition, management, maintenance, ecc.).

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

A database is used in order to store and correlate the input data. Data referring to economic aspects is also managed (number of years to dismiss, rate of interest, etc.). Parameters can be chosen using average or worst case considerations. Sensitivity analysis can be performed acting on input data ranges in order to evaluate the effect of uncertainty intervals upon the computed results and possibly defining lower and upper bounds. The tool elaborates data according to the relationships defined in the database (in particular, using the common attributes of site and threat category) and the mathematical models of (3) and (4), providing: •

The risk associated to each threat ( Ri ) and the overall risk ( RT );

• • •

The total risk reduction considering all the threats; Annual cost of the single protection mechanism and of the whole security system; Annual cost/benefit balance (EB).

5. Example Application Let us consider a case-study of a railway or subway station. The following threats against the infrastructure should be considered: • • • •

1

Damage to property and graffitism (vandalism); Theft and aggressions to personnel and passengers (micro-criminality) Manumission and forced service interruption (sabotage) Bombing or spread of NBCR1 contaminators (terrorism)

Nuclear Bacteriologic Chemical Radiologic.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Security Risk Management of Railway Transportation Systems

395

Table 1. Attack scenarios considered in the example application. Threat ID

Exp. Exp. asset Service D [k€] D [k€]

Threat Description

Threat Category

1

Graffitism

Vandalism Station Ext.

60

0.9

0.5

0

2

Theft of PCs

Theft

Tech. Room

4

0.8

8

6

3

Glass break

Vandalism Station Ext.

12

1

0.5

0

4

Bombing

Terrorism Expl.

Platform

0.01

1

600

300

5

Hacking

Sabotage

Tlc Server

2

0.8

0

10

6

Gas Attack

Terrorism Chem.

Platform

0.01

1

10

150

Vandalism

Hall Platform

70 50

1 1

0.1 0.1

0 0

Physical Sabotage

Platform

4

0.9

5

0

7 8

Furniture Damage Infrastruct.D amage

Site

Est. P Est. [# / year] VINIT

Table 2. Protection mechanisms considered in the example application.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Prot. ID

Countermeasure Description

1

Alarmed Fence

2

Volumetric Detector Videosurveillance (internal)

3

Acq. Cost [k€] 10

5 150

Manag. Cost Site COV [k€ / year] 1 Station Ext. 0.9 Station Int. (night) 1 Tech. 1 Room 20 Hall, 0.95 Platform

4

Chem. Detector

50

2

5

Intrusion Detection System Explosive Detector

1

0.5

Tlc Server

1

50

2

Station Int. (*)

1

6

Platform

0.9

Threat Categories

EP

ED

ER

Vandalism Theft P. Sabotage Theft

0.9 0.9 0.9 0.8

0.3 0.3 0.3 0.6

0.2 0.2 0.2 0.2

Vandalism Theft Sabotage Terrorism Expl. Terrorism Chem. Terrorism Chem. L. Sabotage

0.4 0.6 0.6 0.4

0.6 0.6 0.6 0.3

0.3 0.3 0.8 0.6

Sabotage Terrorism Expl.

0.4 0.3 0.6 0.6 0.2 0.4 0.9

0

0

0.8 0.4 0.1 0.8 0.1 0.1

(*): detectors are physically installed near turnstiles, but the protection is effective on the whole station internal.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

396

Francesco Flammini and Nicola Mazzocca

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Figure 48. Q-RA output data presentation for the example application.

Let us consider the example scenarios reported in Table 1 and the protection mechanisms listed in Table 2, both referring to a specific station. It is assumed that the values are obtained by analyzing historical data of successful and unsuccessful attacks before and after adopting specific countermeasures (such data is usually available for comparable installations). The expected damage relates to the single attack and it is computed by predicting the expense needed to restore the assets and the possible consequences of service interruption (no human injury or loss is considered). The estimated annual cost of the protection mechanisms also accounts for maintenance and supervision, while acquisition and installation costs are accounted separately. Please note that the effect of protection mechanisms may vary according to threat category. Furthermore, all the specified values should not be considered as real. The choice of real values requires an extensive justification using the preliminary analysis and the reference models introduced in Sections 2 and 3. Figure 48 reports the results of the example application computed by the tool. In the assumptions of the example, the positive expected benefit resulting from the adoption of the protection mechanisms clearly justifies the investment, the total benefit being 36722 €/year.

6. Conclusions Infrastructure security has become a major issue in modern society. In this paper we have presented a model-based approach to perform risk assessment and management of railway transportation systems. The method includes both qualitative and quantitative aspects and its main aim consists in minimizing the subjectivity in the risk management process. The method is presently being experimented on rail-based mass transit systems, including subways. Future developments will be aimed at integrating in a single cohesive framework all the steps of the methodology, including survey data management and the engineering of reference models.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Security Risk Management of Railway Transportation Systems

397

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

References [1] Asis International: General Security Risk Assessment Guideline http://www.asisonline.org/guidelines/guidelinesgsra.pdf (2008) [2] Broder, J.F.: Risk Analysis and the Security Survey. Butterworth-Heinemann (2006) [3] Garcia, M.L.: Vulnerability Assessment of Physical Protection Systems. ButterworthHeinemann (2005) [4] Lewis, T.G.: Critical Infrastructure Protection in Homeland Security: Defending a Networked Nation. John Wiley (2006) [5] Meritt, J. W.: A Method for Quantitative Risk Analysis, http://csrc.nist.gov/nissc/1999/ proceeding/papers/p28.pdf (2008) [6] Moteff, J.: Risk Management and Critical Infrastructure Protection: Assessing, Integrating, and Managing Threats, Vulnerabilities and Consequences. CRS Report for Congress, The Library of Congress (2004) [7] Nicol, D.M., Sanders, W.H., Trivedi, K.S.: Model-based evaluation: from dependability to security. In Dependable and Secure Computing, IEEE Transactions on, Vol.1, Iss.1, pp. 48-65 (2004) [8] SANDIA National Laboratories: A Risk Assessment Methodology for Physical Security. White Paper, http://www.sandia.gov/ram/RAM%20White%20Paper.pdf (2008) [9] Srinivasan, K. Transportation Network Vulnerability Assessment: A Quantative Framework. Southeastern Transportation Center – Issues in Transportation Security (2008) [10] U.S. Department of Transportation: The Public Transportation Security & Emergency Preparedness Planning Guide. Federal Transit Administration, Final Report (2003) [11] U.S. Department of Transportation: Transit Security Design Considerations. Federal Transit Administration, Final Report (2004) [12] Wilson, J. M., Jackson, B.A., Eisman, M., Steinberg, P., Riley, K.J.: Securing America's Passenger-Rail Systems. Rand Corporation (2007) [13] F. Flammini, S. Marrone, N. Mazzocca, V. Vittorini: “Evaluating the Hazardous Failure Rate of majority voting computer architectures by means of Bayesian Network models”. In: Risk, Reliability and Societal Safety - Aven & Vinnem (eds), Proceedings of ESREL’07, Stavanger, Norway, June 25-27, 2007: pp. 1715-1721 [14] G. De Nicola, F. Flammini, N. Mazzocca, A. Orazzo: “Model-based functional verification & validation of complex train control systems: an on-board system testing case-study". In: Archives of Transport - International Journal of Transport Problems, Vol. XVI, No. 3-4, December 2005: pp. 163-176 [15] F. Flammini, A. Gaglione, N. Mazzocca, V. Moscato, C. Pragliola: “Wireless Sensor Data Fusion for Critical Infrastructure Security”. To appear in: Proc. International Workshop on Computational Intelligence in Security for Information Systems, CISIS’08, Genoa, Italy, 23-24 October, 2008 [16] F. Flammini, A. Gaglione, N. Mazzocca, C. Pragliola: “DETECT: a novel framework for the detection of attacks to critical infrastructures”. To appear in: Proc. European Safety & Reliability Conference 2008 (ESREL’08), Valencia, Spain, 22-25 September 2008

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

398

Francesco Flammini and Nicola Mazzocca

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

[17] F. Flammini, A. Gaglione, N. Mazzocca, C. Pragliola: “Quantitative Security Risk Assessment and Management for Railway Transportation Infrastructures”. To appear in Proc. International Workshop on Critical Information Infrastructures Security, CRITIS’08, Frascati (Rome), Italy, October 13-15, 2008. [18] CENELEC 2000. EN 50126 Railways Applications – The specification and demonstration of Reliability, Maintainability and Safety (RAMS) [19] Gary Stoneburner, Toward a Unified Security-Safety Model. In IEEE Computer, vol. 39, no. 8, pp. 96-97, Aug., 2006. [20] Andreas Pfitzmann, Why Safety & Security should and will merge. Invited Talk for 23nd Intern. Conference SAFECOMP 2004, Potsdam, September 2004, LNCS 3219, Springer-Verlag, Heidelberg 2004, 1-2 [21] Nicol, D.M.; Sanders, W.H.; Trivedi, K.S. Model-based evaluation: from dependability to security. In Dependable and Secure Computing, IEEE Transactions on, Vol.1, Iss.1, Jan.-March 2004 Pages: 48- 65 [22] J.-C. Laprie (Ed.), Dependability: Basic Concepts and Terminology, Dependable Computing and Fault-Tolerance, 5, 265p., Springer-Verlag, Vienna, Austria, 1992. [23] Object Management Group – Unified Modeling Language: http://www.omg.org/uml

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

In: Railway Transportation Editor: Nicholas P. Scott

ISBN: 978-1-60692-863-9 © 2009 Nova Science Publishers, Inc.

Short Communications B

STUDY ON RAILWAY ACCESSIBILITY MODE CHOICE: A CASE STUDY ON URBAN RAILWAY IN BEIJING Huang Shana, Guan Hongzhib and Yan Hai Transportation Research Center Road and Transportation Lab, Beijing University of Technology, P.R.C.

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Abstract Effective use of urban rapid railway systems requires that the railway systems be effectively connected with other transportation modes. Under such conditions, a higher level of accessibility can be achieved. This paper analyzes different attributes between arrival and departure access modes through a trip survey in Beijing. Using this data, this paper sets up the nested logit model access mode choice paradigm and analyzes the influence of three factors: access time, access cost, and access distance on the access mode choice. It is confirmed to be effective based on the combined estimated methodology. It is shown that these three factors are all negatively correlated with the access mode choice. It also appears more sensitive to the bus access fee. The conclusions provide an analysis tool for urban railway planning and construction.

Key words: railway access model; arrival access mode; departure access mode; nested logit model; data fusion.

Introduction Effective use of urban rapid railway systems requires that the railway systems be effectively connected to other transportation modes so that they are accessible. To some extent, the connection may even influence an individual’s travel mode. So it is necessary to study the railway connection system to make the whole transit system more convenient and more effective. a b

E-mail address: [email protected]. E-mail address: [email protected] Phone number: 0086-13661202904. Fax number: 010-67391509. Address: Beijing University of Technology num 100 PingLeYuan ChaoYang district Beijing P.R.C

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

400

Huang Shan, Guan Hongzhi and Yan Hai

Various railway accessibility problems have been analyzed in prior literatures. Korf and Demetsky[1]divided subway stations into five types and analyzed subway station access mode choices. Bates[2] classified station accessibility based on service areas covered by buses and intercity railways and discussed how to improve existing access facilities. Dickins[3] expounded the utilization and potential of light rail facilities based on survey data of 51 cities in Europe and North America. His analysis included proposals for locations, attributes, as well as sizes of access facilities. Access facilities were the focus of several domestic studies [4, 5] , in which access time and access distance were investigated and system evaluations [6] carried out from a planning point of view. Guan Hongzhi has developed two different Logit models based on the survey conducted in Beijing to analyze the mode choice behavior on the arrival and departure railway station respectively. This paper also focuses on the mode choice behavior analyze Based on a choice behavior survey conducted in Beijing, this paper analyzes factors that influence choices made by tripmakers when using urban railways. The Nested Logit model is then used to develop a railway access mode choice model to organized both arrival and departure access process together. Meanwhile the influence and the elasticity of each variable was analyzed to illustrate various impact of these variable to the access mode choice and show how travelers make choices about using urban railway systems, which would influence urban railway accessibility policies.

Data

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

The survey was conducted in Beijing, and in order to analyze the preference of railway users, it included inquiries about gender, age, occupation, income, and car ownership, as well as questions related to their travel characteristics such as trip purpose, departure time, vehicle used, and access time and cost. The trip purpose was divided into 7 groups as go to work, go to school, business, recreation, go shopping, drop at friends and others. Other than railways, other traffic modes were buses, walking, taxis, bikes, private cars, and others [7].

Figure 1. Sketch of Survey Locations.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Study on Railway Accessibility Mode Choice

401

Access Modes

mode distribution proportion (%)

The station access modes were grouped into arrival and departure modes as shown in Figure 2 where the differences between the proportions of the two modes are small. Walking is the dominant access mode, making up over 70 percent of the total count in either group. Bus ranks as the secondary traffic mode with a proportion of about 20 percent each. The proportions for the other traffic modes are relatively insignificant (Figure 2). 80 70 60 50 40 30 20 10 0

70 72

arrival departure 23 22 3 2 walking

bicycle

bus

2 3

0 0

2 2

taxi

private

others

car

mode Figure 2. Access Mode Distribution.

The access time is defined as the total time that railway users spend on both arriving at and departing from the stations. The ratio of the access time to the entire trip time generally ranges from 0.2 to 0.6 (over 60%) as shown in Figure 3.

choice proportion (%)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Proportion of Access Time And Trip Time

30 24.37

25

22.34

20 15

13.2

12.18

10

8.63

6.6

5.08

5

7.61

0 0-0.1

0.110.2

0.210.3

0.310.4

0.410.5

0.510.6

0.610.7

>0.7

ratio of the access time to the total trip Figure 3. Distribution of Ratio of Access Time To The Total Trip.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

402

Huang Shan, Guan Hongzhi and Yan Hai

Access Fee As shown in Figure 4, 72% of the users do not pay any fare for arriving at or transferring from the railway stations, 20% pay less than 2 RMB for their access fare, which is mainly for bus access. This agrees with the distribution in Figure 2 in which walking and buses were the two main access modes.

80

72

70 proportion(%)

60 50 40 30 20

14

10

6

4

2

3

1

3

4

〉4

3

1

3

3

4

〉4

0 0

1

Acces s fare(RMB))

(a) Arrival 80

72

70 proportion(%)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

60 50 40 30 15

20

6

10 0 0

1

2

Acces s fare(RMB)

(b) Departure Figure 4. Effect Of Access Fee.

Access Distance Figure 5 shows that the access distance differs significantly between arrival and departure modes. The arrival mode is only 1.4 km, 2.0 km short of that for departure mode.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Study on Railway Accessibility Mode Choice

403

access distance(km)

2.5 2.0

2.0 1.4

1.5 1.0 0.5 0.0

arrival

departure

Figure 5. Comparison Of Access Distance.

Bus Access Time Also, Figure 6 shows that the access time varies greatly between arrival and departure modes when using buses to access railway terminals. The mean arrival mode of 11 minutes is much shorter than the mean departure mode of 25 minutes.

bus access time(min)

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

30 25

25 20 15

11

10 5 0 arrival

departure

Figure 6. Comparison Of Bus Access Time.

The two figure analysis above indicated the fact that people will show preference on rail mode when it is easier to arrive at rail station from origin. Moreover, this also shows that there is significant difference between arrive and departure access mode.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

404

Huang Shan, Guan Hongzhi and Yan Hai

Methodology The logit model is commonly used to model the behavior of mode choice. This paper adopts the NL (nested logit) model. The normal equation of NL model is as follows[8]:

Pn (tm) = Pn (t | m)Pn (m) exp(V(t|m)n ) exp[μ(Vmn + Id )] = Rmn ⋅ M ∑exp(V(t′|m)n ) ∑exp[μ(Vm′n + Id )] t′=1

(1)

m′=1

Where:

Pn (tm) —the probability of the nth traveler choosing tm; Pn (m) —the probability of the nth traveler choosing m; Pn (t | m) —the probability of the nth traveler choosing r under the condition of choosing m first; V ( t | m ) n —the fixed part of the utility function of t under the condition of choosing m;

Vmn —the fixed part of the utility function of m; I d —the inclusive value;

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

μ —the scale parameter.

The structure of the NL model is shown in figure7. It has four branches of access modes: bus, walking, taxi, and bike, as well as two limbs: arrival and departure. What is noticeable is that the relationship between arrival and departure is not mutually exclusive but parataxis/juxtaposed, so the estimation methods will be a little different from the normal NL model.

Figure 7. Access Model Structure.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Study on Railway Accessibility Mode Choice

405

According to the preceding statistical analysis, there is little difference between arrival and departure in terms of the proportion for each mode, but the variation of bus access time and access distance is considerable. The utility functions of arrival and departure, respectively, is shown as follows:

and

U f = V f + ε f U r = Vr + ε r

(2)

V f = θ f 0 + ∑θ fk X k , Vr = θ r 0 + ∑θ rk X k

(3)

k

k

Where: X k —the variance vector that influences the traveler and;

θ f 0 θ r 0 θ fk θ rk —the coefficient of the parameter. The first assumption is that the traveler’s mechanism in choosing his access mode is the same for both arrival and departure, which means:

θ fk = θ rk In addition, we assume that the random components of the utility functions have different levels of randomness, which can be expressed by [9, 10]:

Var[ε f ] =

1

ω2

Var[ε r ]

(4)

Where: ω is an unknown parameter. Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Because

ε f and ε r are independently distributed with zero means, the same as what is

assumed in the ML model, it can be estimated with MLE(maximum likelihood estimation). The ML functions are as follows: N

Lf (θ , ω) = ∑∑δn[ωVfn − ln ∑exp(ωVf )] n=1 f N

Lr (θ , ω) = ∑∑δrn[Vrn − ln ∑exp(Vf )] n=1 r

Where:

(5)

f′

(6)

f′

δ fn , δ rn is the choice of the traveler in the access mode of arrival and departure,

that is :

⎧1 : traveler n chose mode i ⎩0 : otherwise

δ in = ⎨

θ is the unknown vector for the parameter.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

406

Huang Shan, Guan Hongzhi and Yan Hai

Based on these assumptions, the parameter can be estimated by the following sequential estimation method: Step1: Define V fn = θ f 0 +

∑θ

k

X nk ; and then estimate the arrival model in function 3 from

k

the arrival data to obtain θˆ f 0 and

∧

ωθ k .

Step2: Define Vrk = θ r 0 + γ

∧

∑ ωθ

k

X nk , and then estimate the departure model in function 4

k

from the departure data to obtain γˆ and θˆr 0 . Calculate ωˆ = 1

ˆ ˆ ˆ γˆ , θ k = ωθ k / ω

Step 3: Multiply X fnk by ωˆ to obtain a modified arrival data set. Pool the departure data and the modified arrival data; and then estimate the two models jointly to obtain θˆk .

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Result of the Model Application of the survey data in the Logit railway access mode choice model for arriving at and departing from the stations gave the results listed in Table 1. In order to explain its validity and superiority, this paper also set up another model that was just a simple ML model with data for both arrivals and departures. According to the correlation analysis and the research in the transportation mode choice, we select three variables: access time, access fee, and access distance to establish the choice model.

Table 1. The Parameter Estimates Of The Access Model model1 combined model Coefficient t-statistics

Constant(walk) Constant(bike) Constant(bus) Access fee Access time Access distance Scale parameter ω

ρ2

3.233 -2.4782 1.3092 -0.7812 -0.5335 -4.1627 2.1681

2.7743 -2.4631 1.7351 -5.5936 -5.4789 -7.8584 5.8500 0.8267

model2(ML) Coefficient t-statistics

8.6656 1.7322 3.3009 -0.1945 -0.1298 -4.1486 ——

7.1451 1.6724 3.5798 -3.7942 -1.4398 -7.5715 —— 0.8032

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Study on Railway Accessibility Mode Choice Comparing model1 with model2, the value of

407

ρ 2 of model 1 improved from 0.8032 to

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

proportion

0.8267. And the value of t test of access time, the most important variable to influence the behavior of people’s access mode choice, changed from -1.4398 to -5.4789, which is significant under confidence level of 95%. From this point of view, model1 is more effective than model2. The coefficient for the three variables - access time, access fee, and access distance - are all negative, which means the utility will decrease with the increase of any these three variables. The finding is in accordance with the widely-accepted belief that the more timeconsuming, the more expensive, and the longer distance access to the metro station a trip is, the greater possibility that one will switch to an alternative mode of transportation. Furthermore, the value of the coefficient reflects each variable’s impact on the utility function.. An interesting result is that the coefficient for access time is nearly 80 times as that for access fee. Therefore, this ratio can be used to reflect the value of travel time—around 80RMB/h, which is much higher than the actual value of travel time for Beijing. This is because the average income of rail users is higher than the average income for all of Beijing, which is clearly shown in Figure 8. Compared to all users(all the survey data), the proportion of mid income rail users is larger, while the proportion of high income group is equal.

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

28%

42%

high income mid income

48%

rail user

62%

low income

all user

Figure 8. The Contrast Of The Income.

Elastricity Analysis To analyze how and to what extent each variable influences the decision-making of policy makers; we conduct the elasticity analysis to find out how the mode choice ratio changes when only the studied variable changes while other variables remain constant. So in this paper we choose access distance, bus access time, and bus access fee as the variables to analyze.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

408

Huang Shan, Guan Hongzhi and Yan Hai

Access Distance

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Based on model1, figure9 shows the elasticity analysis of access distance variable. The conclusion is that when access distance increases, the proportion of the walking mode decreases sharply, while the proportion of bike and bus modes surge. (Since the proportion of taxi is zero until the access distance exceeds 4 km, figure9 excluded the taxi mode.) It is interesting to note that when access distance reaches 1.5 km, the proportion of the walking mode is equal to that of the survey. It can be confirmed that the convenient walking access distance is 1.5 km, slightly longer than the common belief. The reason may be that the density of railway stations and the railway network in Beijing are not large enough, leaving the rail station far away from the origins and destinations. So the conclusion is that the best access distance between two rail stations is less than 1.5 km.

Figure 9.Analysis Of Sensitiveness On Distance.

Bus Access Time Based on model 1, figure10 shows the sensitive analysis of bus access time variable. Due to the negligible influence of bus access time on walking mode proportion, figure10 has not shown the walk mode. In figure10, the conclusion is that with the increase of bus access time, the proportion of bus mode decreases while those of the other three modes increase. When the bus access time is 10 minutes, the proportion reaches 25%, nearly the same as the actual proportion revealed in the survey. So it can be concluded that when the bus access time is less than 10 minutes, the proportions of the other modes will be transferred to the bus mode; when the bus access time is between 10 minutes and 20 minutes, the bus mode proportion will shift to the walk mode; when the bus access time increases from 30 minutes to 50 minutes, the bus mode proportion will mainly move to the bike mode, and when the bus access time is more than 50 minutes it will mainly transfer to the taxi mode. So the convenient access time is less than 20 minutes for the bus mode.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Study on Railway Accessibility Mode Choice

409

Figure 10. Analysis Of Sensitiveness On Bus Time.

Based on model 1, figure11 shows the sensitive analysis of bus access fee variable. For the same reason with figure10, the walk mode proportion is excluded in figure11. The proportion of bus mode decreases with the increase of access bus fee. But there is an intriguing result in figure8 of the bus mode proportion curve, which is different from figure9 and figure10. In the previous two figures, the main mode decreases quickly at the beginning as the variable increases and the pace slows down when the variable becomes large enough. Obviously, figure11 shows the reverse transformation ratio. In figure11, the ratio of bus mode proportion decreases sharply as the variable increases, especially when the bus access fee exceeds 4 RMB. So we draw the conclusion that bus access mode is more sensitive to the bus access fee, especially when the latter exceeds a certain value. 0.2 0.18 choice proportion

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Bus Access Fee

0.16 0.14 0.12 0.1

bike bus taxi

0.08 0.06 0.04 0.02 0 0

2 4 6 bus access fee(yuan)

8

Figure 11. Analysis Of Sensitiveness On Bus Fee.

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

410

Huang Shan, Guan Hongzhi and Yan Hai

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Conclusion By analyzing the railway lines in Beijing, this paper, in order to improve the integration of railway lines design and construction into the overall transportation system, evaluates the connections between railways and regular public transportation systems. To this end, we designed the survey and obtained the basic data on travel mode and relevant personal attributes. From statistical analyses, we found that the proportion of the access mode differed little between arrival and departure mode. Walking is the main mode followed by the bus mode. This survey also implicated that 72% of the access fee was zero, and around 20% was less than 2 RMB. Both access distance and bus access time differed a lot between arrival and departure mode. The arrival mode had fairly small values in both variables. In addition, the ratio of the access time to the entire trip time generally ranged from 0.2 to 0.6. We developed the NL model to combine the arrival and departure data together. This model was superior to the ML model, which used both arrival and departure data without scale parameters. Based on the NL model, we found that the three variables in the model, i.e. access distance, access time, and access fee, were all negative for the mode choice. Further, the time value took value as much as around 80 RMB per hour, which is relatively high. According to elasticity analyses of the variables, we found that access distance had significant impact on the walking mode. As access distance increased, the proportion of the walking mode decreased sharply. Hence the appropriate access distance should be less than 2 km. The appropriate access time for the bus mode was around 20 min; if it exceeded 20 minutes, the choice would shift to the bike and taxi modes. The bus access mode was more sensitive to the bus access fee. In effect, given the low cost of bus fee in Beijing at the present time, it needs to concentrate on the improvement of bus services and comfort to maintain the increase of constant patronage. The extension improvement of this paper is to analyze the influence of the access mode on all the travel modes. Therefore, the structure of the model can be extended to 3 levels with the upper level having 5 travel modes. Moreover, this will lead to more variables in the corresponding survey during the following research, which requires a more in-depth and more accurate study on the policy and personality .

References [1] Korf J. L, M.J.Demetskv. Analysis of Rapid Transit Access Mode Choice. TRR. 817, 1981 [2] BatesE.G.. A study of Passenger Transfer Facilities. TRR. 662. 1978 [3] Ian S.J. Dlckins. Park and Ride Facilities on Light Rail Transit systems. [J]Transportation 18:23~36, 1991 [4] Wang Qiuping, Feng L. On the coordination of switching non-rail transport modes to rail transport modes inurban passenger traffic system. Journal of Xi'an University of Architecture &Technology (Natural Science Edition), 2003, 35(2): 136-139 in Chinese [5] Du Caijun Jiang Yukun. Connection of urban rail transit with other public transportations. Urban Rapid Rail Transit, 2005, 18(3): 45-49. in Chinese

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Study on Railway Accessibility Mode Choice

411

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

[6] Wu Youmei, Zhang Xiuyuan. The transfer with bus system and counter measurements of urban mass transit. Urban Rail Transit, 2005, 25(8): 19-21. [7] Hongzhi Guan, Yuanfei Yin. Urban Railway Accessibility. [J]TsingHua Science And Technology volume 12, number2 April 2007 [8] Jeffrey M.Wooldridge Economic analysis of cross section and panel data the MIT Press [9] Ben-akiva. M, Morikawa T. Estimation of travel demand models from multiple data sources [J]. Transportation and Traffic theory, 1990(b) 461-476. [10] Hongzhi Guan, Kazuo Nishii Atsushi Tananka and Takeshi Morikawa. A Model of Car andP&BR Choice in Commuting Trips Combining the Experiment Day Data with the StatedPreference Data [J]. Infrastructure Planning Review. No. 16, 1999, 955-961 (in Japanese)

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved. Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

INDEX

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

A abatement, ix, 141, 161 access charges, 27, 32, 41, 45, 46, 61 accessibility, xiii, 56, 295, 390, 399, 400 accidents, 9, 27, 40, 49, 50, 52, 90, 280, 288 accountability, 21, 29, 32 accounting, 18, 21, 26, 39, 40, 41, 47, 52, 67, 360, 390 accuracy, 100, 145, 182, 184, 191, 212, 283, 294, 313, 315, 316, 335, 377 achievement, 25, 53, 56, 372 activation, 288 actuation, 278 actuators, 270, 272, 273, 274, 275, 338, 341 adaptability, 29, 36 adaptation, vii, xi, 66, 142, 303 ADC, 363 adjustment, viii, xii, 27, 63, 68, 70, 71, 72, 75, 82, 293, 355, 356, 364, 366, 367, 368, 369, 370, 376, 377 administrators, 71, 72, 76, 83 Africa, xii, 379, 380, 381, 383, 384 age, 6, 89, 90, 92, 93, 156, 169, 184, 203, 204, 211, 234, 400 aggregates, 41, 350 aggregation, 390 aging, x, 217, 218, 221, 222, 225, 228, 230, 233 air-traffic, 168 Albania, 30, 35, 37 algorithm, ix, 117, 163, 164, 165, 166, 167, 168, 169, 171, 172, 173, 174, 175, 177, 178, 188, 270, 274, 278, 294, 295, 296, 328 allocative efficiency, 45 alternative, 19, 50, 66, 143, 148, 151, 157, 167, 266, 280, 309, 407 alternatives, 143, 230 alters, 139 ambiguity, 386

amortization, 21 amplitude, viii, ix, 87, 97, 102, 104, 106, 107, 108, 111, 112, 114, 116, 117, 120, 124, 125, 138, 181, 185, 188, 190, 192, 194, 196, 199, 214, 215, 262, 273, 290, 293, 296, 298, 329, 342, 345 animals, 51, 225 annual rate, viii, 63, 83 appropriate technology, 57 Argentina, 279 argument, 39, 143, 150, 158 arrhythmia, 89, 126 Asia, 8, 38, 59 assessment, ix, 16, 35, 40, 52, 55, 56, 70, 116, 137, 181, 182, 183, 187, 191, 205, 209, 214, 215, 226, 228, 232, 277, 333, 392 assessment procedures, 277 assets, 20, 26, 54, 58, 67, 388, 390, 396 assignment, 23, 169, 171 assumptions, 46, 73, 149, 150, 155, 157, 222, 396, 406 asymmetry, 369 atherosclerosis, 89 atmospheric pressure, 314 ATP, 281 attacker, 390 attacks, 137, 222, 386, 389, 396, 397 attractiveness, 53, 70, 280 attribution, 227, 228 Australia, 350, 354 Austria, 5, 78, 81, 398 automation, xi, 167, 168, 178, 179, 279, 280 autonomy, 69 availability, 72, 168, 227, 232 average costs, 47, 67 average variable cost, 41 averaging, 242, 362 aversion, 80, 83 awareness, 156, 168

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

414

Index

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

B balance sheet, 21 Balkans, 33 bandwidth, 99 barriers, 29, 30, 31, 33, 34, 35, 37, 63, 280, 282, 390 Bayesian estimation, 390 beams, 227, 230 behavior, 41, 49, 57, 68, 69, 72, 76, 83, 106, 109, 112, 121, 123, 139, 148, 166, 187, 193, 200, 202, 205, 207, 209, 212, 361, 365, 375, 376, 400, 404, 407 Beijing, xiii, 353, 399, 400, 407, 408, 410 Belgium, 5, 6, 21, 23, 45, 73, 78, 81, 138, 277, 301 benchmarking, 30 bending, 190, 196, 197, 200, 202, 215, 270, 326, 328, 339, 340, 341 bias, 75, 81 biological activity, 139 biological responses, 126 biological systems, 88, 116 birds, 114 birth, 242 blame, 164 blocks, 196, 244, 252, 253, 254, 255, 266, 270 blood, 91, 114, 139, 155, 156, 158 blood flow, 114 blood pressure, 91, 155, 156, 158 BMA, 350 bonding, 207 border crossing, 32, 53 Bosnia, 35, 37 bounds, 394 brain, 114, 136 brain tumor, 136 brainstorming, 388, 389 Brazil, 380 breast cancer, 89, 137 breathing, 129 Britain, 44, 46, 67 Brno, 7 broadband, 242, 243 Bulgaria, 1, 5, 26, 28, 30, 35, 37, 44 bureaucracy, 226 business management, 31, 32

C C++, 116 cables, 94, 285, 370 calibration, xii, 100, 293, 355, 360, 361, 363, 364, 374, 375 Canada, 301

cancer, 89, 136 candidates, 7 capital intensive, 73 capsule, 100 carbon, 57 cardiovascular disease, 89, 90, 91 cardiovascular morbidity, 90 cardiovascular risk, 90, 126, 127 cardiovascular system, 126 carrier, 24 case study, 158 cash crops, 383 cast, 199, 270 catalyst, 26, 63 catastrophes, 126 Caucasus, 8 CEC, 38, 40, 42, 48, 59, 138 cell, 335 Central Europe, 34 central planning, 164 certificate, 19, 25 certification, 19, 156 channels, 227 chaos, 280 children, 380 Chile, 70 China, 339 chronic lymphocytic leukemia, 89 circulation, 4, 284, 292, 300 classes, 14, 183 classification, 183, 184, 193, 209, 294 clean air, 147 cleaning, 47, 226, 230, 334 clients, 12, 31, 37, 54 clusters, 114 CO2, 17 coaches, viii, ix, 5, 9, 30, 35, 87, 101, 121, 163 codes, 5, 63, 296, 298 cognition, 179 cognitive load, 168, 177 cognitive models, 166 cognitive perspective, 177 cohort, 89, 91 collaboration, 28, 89, 275 collisions, 221, 222 colonisation, 380 combined effect, 308 commerce, 383 commodity, 7 common rule, 21 communication, x, 32, 94, 217, 226, 283, 285 community, 20, 54, 88, 145, 166, 381, 386 compatibility, 25, 301

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

415

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Index compensation, 25, 32, 50, 90 competition, 11, 19, 21, 22, 26, 27, 39, 42, 43, 44, 45, 48, 53, 67, 158 competitive advantage, 280 competitive conditions, vii, 1, 48 competitiveness, 1, 11, 12, 18, 24, 28, 29, 31, 32, 34, 35, 36, 38, 48 complexity, 35, 45, 164, 168, 169, 170, 175, 177, 178 compliance, 15, 41, 49 components, xii, 6, 7, 55, 56, 88, 93, 98, 99, 101, 102, 104, 106, 107, 108, 109, 111, 112, 116, 117, 118, 119, 120, 121, 123, 124, 125, 143, 182, 183, 184, 191, 192, 199, 214, 227, 230, 239, 241, 252, 253, 284, 303, 304, 305, 309, 312, 326, 334, 405 composition, 70, 187, 262 comprehension, xi, 39, 188, 238, 244, 255 computation, 106, 116, 117, 120, 393 computer systems, 385 computing, 120, 313 concentrates, 291 concentration, x, 49, 182, 183, 184, 206, 212, 213 conception, 4, 47 concrete, 41, 199, 218, 222, 234, 262, 268, 295, 331 conditioning, 285, 286, 295 conductivity, 114 conductor, 284 confidence, 34, 57, 92, 117, 407 confidence interval, 92, 117 confidentiality, 46 configuration, 93, 259, 260, 261, 268, 273, 311, 312 conflict, 33, 143, 167, 172 conformity, 14, 70 Congress, 278, 302, 397 consciousness, 142 consensus, 88 consolidation, 320 constant load, 328 construction, xi, xiii, 17, 22, 23, 24, 45, 47, 74, 92, 106, 182, 205, 210, 218, 220, 221, 223, 228, 230, 231, 238, 255, 259, 303, 320, 323, 331, 384, 399, 410 consulting, 21 consumers, 24, 42 consumption, 25, 48, 56, 143, 144, 149 continuity, 207 control, vii, x, xi, 11, 21, 23, 24, 25, 32, 36, 37, 45, 47, 56, 69, 75, 82, 116, 148, 168, 223, 231, 237, 270, 273, 274, 277, 278, 279, 280, 281, 282, 283, 291, 300, 340, 356, 358, 363, 369, 370, 385, 386, 390, 392, 397 cooling, 207 coronary heart disease, 89, 90, 91, 92, 116

correlation, 81, 296, 297, 298, 343, 361, 390, 406 correlation analysis, 406 correlation function, 296 corrosion, x, 196, 199, 212, 217, 221, 222, 225, 230, 231, 233 cost benefit analysis, ix, 141 cost saving, 34 cost-benefit analysis, 54, 56 costs, viii, xi, 21, 22, 23, 26, 29, 30, 32, 33, 34, 35, 36, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 58, 59, 60, 61, 63, 66, 67, 68, 70, 71, 72, 75, 82, 83, 142, 143, 149, 150, 151, 154, 158, 162, 165, 227, 228, 231, 279, 281, 282, 285, 286, 294, 307, 308, 309, 393, 396 Council of Europe, 19 Council of Ministers, 18 couples, 363 coupling, 70, 242, 243, 247, 250, 254, 256, 257, 258, 259, 261, 264, 265, 298, 371 covering, 26, 42, 43 CPB, 85 crack, 183, 185, 190, 213, 215 creditors, 25 creep, 207, 243 criminality, 388, 394 critical infrastructure, 397 critical state, 211 criticism, 168 Croatia, 28, 35, 60 crystals, 114 Cuba, 380 cues, 168 cultural differences, 384 current limit, 273 customers, 17, 29, 32, 34, 35, 57 CVD, 90, 91 cycles, 183, 184, 185, 188, 190, 191, 209, 313, 321, 322, 369 Cyprus, 5 Czech Republic, 5, 6, 344

D damping, 100, 244, 246, 247, 252, 253, 254, 255, 259, 265, 266, 267, 268, 269, 306 data analysis, viii, 88, 89, 101, 389 data collection, 91 data processing, 290 data set, 106, 109, 112, 118, 406 database, 47, 91, 102, 124, 225, 226, 229, 232, 233, 389, 393, 394 dating, 73, 145 deaths, 89, 91, 92, 156 debt, 11, 21, 24, 54

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

416

Index

decay, 93, 267 decision makers, 166 decisions, xi, 24, 31, 36, 44, 54, 71, 76, 155, 156, 168, 170, 182, 231, 303, 304, 309 deconvolution, 347 deduction, 231 defects, x, 217, 218, 221, 222, 226, 227, 228, 229, 230, 231, 233, 234, 244 defense, 12 deficiency, 56, 227 definition, 11, 17, 26, 104, 116, 188, 226, 228, 386 deformation, 183, 192, 221, 284, 294, 304, 320, 321, 322, 350, 352, 354, 356 degradation, xii, 167, 285, 300, 303, 304, 346 delivery, 15, 23, 32 demand curve, 50 Denmark, 5, 23, 45, 73, 77, 81, 161, 162 density, 4, 5, 6, 93, 106, 111, 120, 200, 210, 211, 241, 280, 281, 287, 408 depreciation, 148 designers, 231, 388 detection, xi, 215, 279, 282, 283, 284, 285, 286, 287, 288, 289, 290, 291, 292, 293, 294, 295, 296, 297, 298, 299, 300, 301, 302, 358, 397 developed countries, 231 deviation, 76, 343 differentiation, 44, 49, 52, 156, 158 direct costs, 52 direct measure, 90 directives, 19, 20, 23, 32, 37, 67, 280 discharges, 286 disclosure, 31 discomfort, 67, 238 discontinuity, 184 discrimination, 21, 42, 43, 48, 290, 302 discs, 239, 246, 371 dispersion, 76, 291 displacement, 100, 114, 200, 201, 222, 284, 306, 325, 335, 338, 339, 340, 358 dissatisfaction, 142 distortions, 42, 44, 48, 97 distribution, 4, 25, 73, 81, 95, 96, 97, 98, 102, 109, 120, 121, 123, 147, 153, 166, 207, 291, 312, 318, 326, 331, 357, 378, 402 distribution function, 153 diversity, 42, 226 division, 24 doppler, 340 dose-response relationship, 99 double counting, 156 drainage, 322, 333, 334 DSL, 169 ductility, 194

duration, 51, 89, 97, 120, 164, 183, 184, 196, 210, 211, 281, 292 duties, 225 dynamic factors, 187, 202 dynamic loads, 100, 311, 312, 334, 335, 348

E East Asia, 234 Eastern Europe, 7, 14, 18, 27, 28, 29, 30, 33, 35, 36, 37, 44, 59, 61 economic boom, 219 economic development, 55, 63 economic efficiency, 144 economic growth, 56, 57, 280 economic incentives, 57 economic performance, 33 economic resources, 25 economic welfare, 17 economics, 4, 38, 43, 59, 150 elaboration, xii, 18, 28, 53, 355, 361, 377 elastic deformation, 359 elasticity, 43, 47, 75, 83, 314, 326, 327, 400, 407, 408, 410 electric current, 93, 114 electric field, 94, 114, 139 electrical conductivity, 94 electricity, 284 electromagnetic, 88, 93, 99, 114, 136, 137, 138, 286, 292, 301 electromagnetic fields, 88, 99, 136, 138 electronic systems, 291 elongation, 294 emission, vii, ix, x, 17, 141, 142, 154, 155, 156, 158, 237, 238, 239, 241, 242, 243, 244, 245, 246, 247, 248, 249, 251, 252, 254, 256, 258, 259, 261, 262, 263, 264, 265, 266, 268, 270, 274, 275, 285, 296, 298 emitters, 238, 243 employees, 31, 89, 90, 91, 116, 136, 137, 381, 384 employment, 56 EMU, 90, 91, 101, 102, 104, 107, 108, 109, 111, 116, 117, 119, 120, 121, 122, 123, 124, 125, 126 endorsements, 38 energy, viii, 17, 55, 69, 88, 93, 154, 239, 258, 263, 265, 267, 307, 363 energy consumption, 17, 55, 307 energy supply, 363 enlargement, 16, 17, 47 enthusiasm, 380 entrepreneurship, 25 entropy, 117 environment, viii, xiii, 9, 10, 12, 28, 32, 47, 49, 53, 54, 55, 56, 57, 58, 66, 87, 88, 89, 93, 94, 105,

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

417

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Index 120, 138, 139, 160, 188, 228, 231, 232, 287, 295, 296, 309, 370, 377, 385, 386 environmental conditions, 277, 295 environmental effects, 56 environmental factors, 51 environmental impact, 56, 58 environmental policy, 18 Environmental Protection Agency, 159, 161 environmental sustainability, 54 equilibrium, 2, 21, 50, 144 ergonomics, 167 erosion, 222 estimating, 31, 39, 41, 46, 47, 50, 51, 69, 74, 146, 148, 149, 150, 154 Estonia, 5, 9, 26 EU enlargement, 16, 17 Europe, vii, 1, 4, 8, 11, 12, 14, 15, 17, 20, 22, 36, 42, 43, 45, 57, 58, 59, 60, 61, 63, 67, 281, 282, 285, 400 European Commission, 12, 19, 30, 38, 64, 65, 66, 68, 69, 85, 89, 136, 142, 156, 158, 160, 302 European integration, 27 European Parliament, 19, 27, 85, 160, 301 European Union (EU), vii, ix, 1, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 30, 37, 38, 39, 40, 42, 43, 45, 46, 50, 52, 53, 54, 57, 58, 59, 60, 63, 64, 66, 68, 69, 85, 141, 142, 145, 146, 147, 150, 160, 280, 281, 282, 302 EUROSTAT, 5, 6, 8, 9, 10, 13 evening, 72, 146, 147 evolution, xii, 68, 70, 71, 75, 80, 102, 111, 131, 195, 300, 303 examinations, 187, 200 excitation, 239, 242, 256, 276, 278, 305, 306, 325, 342, 343 execution, 22, 172, 173, 174, 176, 228 exercise, 19, 24, 26, 30, 164, 166 expenditures, 2, 29, 31, 40, 53, 54, 232 experimental design, 164, 168 expertise, 15 exploitation, 4, 22, 36, 37, 38, 383, 387 exposure, viii, ix, 51, 87, 88, 89, 90, 91, 92, 114, 115, 116, 125, 126, 136, 137, 139, 141, 143, 144, 148, 149, 150, 155, 295 external costs, vii, 1, 39, 42, 49, 50, 52 external environment, xii, 385, 386, 387 externalities, ix, 25, 48, 49, 52, 54, 141, 142, 143, 154, 158 extraction, 99

failure, x, 56, 71, 164, 168, 182, 184, 185, 188, 190, 191, 209, 210, 211, 217, 221, 225, 233, 234, 285, 290, 304, 313, 320, 321, 322, 326, 386, 391 fairness, 42, 50 family, 380 fatigue, ix, x, xii, 181, 182, 183, 184, 185, 186, 187, 188, 190, 191, 192, 193, 194, 195, 196, 198, 199, 200, 203, 205, 206, 207, 209, 210, 211, 212, 213, 214, 215, 217, 221, 225, 229, 233, 303, 304, 326 fault tolerance, xi, 279 feedback, 100 FEM, 203, 207, 291 fertilizers, 55 FFT, 116 fiber optics, 284, 286 fibers, 99 fidelity, viii, 63, 72, 76, 83, 84 field trials, 322 filters, 294 filtration, 116, 117 finance, 2, 21, 29, 58, 224 financial performance, 20, 21, 66 financial resources, 50 financing, 11, 17, 21, 22, 37, 45, 50, 142, 143 Finland, 5, 6, 23, 25, 41, 45, 73, 78, 81, 101 fire fighting, 166 firms, 26, 58, 71, 72, 381 fixed costs, 40, 43, 44, 50 flexibility, 9, 32, 46, 54, 55, 58, 67, 99, 306, 338, 339 flight, 358, 359, 360, 361, 362, 363, 366, 367, 371, 372, 373, 375, 376, 377 floating, 262 flooding, 320, 390 flora, 381 fluctuations, 70, 79, 93, 97, 182 fluid, 114, 239 focusing, xi, 56, 88, 225, 237, 325 forecasting, 72 fractures, 182 France, 5, 8, 10, 21, 23, 44, 45, 58, 73, 78, 81, 138, 218, 230, 231, 345, 381, 384, 385 franchise, 26, 44 freedom, 11, 25, 31, 32, 81, 278 friction, 207, 270, 277, 278, 320 fuel, 23, 52, 56, 58 funding, 16, 24, 32, 38, 43, 52, 55, 58, 393 funds, 2, 4, 16, 18, 22, 24, 25, 26, 28, 34, 35, 38, 58, 150 fusion, 399

F factor market, 2

G GDP, 52, 54, 80

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

418

Index

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

GDP per capita, 80 gender, 400 generation, xi, 115, 132, 145, 166, 214, 237, 242, 243, 276, 277, 295, 297 genre, 381, 382, 384 geography, 384 Germany, 5, 8, 9, 10, 14, 24, 26, 36, 44, 45, 60, 67, 73, 77, 81, 98, 160, 218, 275, 278 global economy, 55 global trade, 55 globalization, 27 goals, vii, x, 1, 6, 27, 28, 31, 43, 50, 55, 57, 164, 172, 217, 234, 377 goods and services, 25 government, 2, 11, 12, 20, 23, 24, 27, 31, 32, 34, 35, 36, 37, 38, 39, 41, 44, 45, 49, 54, 56, 58, 63, 143, 223, 224, 226, 232, 384 GPS, 282, 302 grants, 24, 43 graph, 178, 251 Great Britain, 15, 23, 44, 162, 379 Greece, 5, 6, 21, 35, 37, 77, 81, 136, 137 greenhouse gases, 51 groups, 12, 26, 35, 40, 90, 92, 142, 147, 150, 221, 298, 384, 400 growth, vii, 1, 8, 10, 12, 14, 16, 17, 27, 30, 35, 52, 57, 63, 73, 75, 76, 243, 261, 273, 280, 325, 384 growth rate, 16, 76 guidance, 99 guidelines, 21, 385, 397 guiding principles, 58 Guinea, 381

H harm, 56 harmonization, 28, 42 hazards, 88 health, viii, 51, 57, 87, 88, 115, 116, 121, 126, 137, 138, 139, 142, 149, 150, 155, 158, 182, 214, 215, 225, 227, 228, 229 health effects, viii, 87, 115, 126, 139 health problems, 88 health status, 150, 155 heart attack, 120 heart disease, 89, 91, 92, 136 heart rate, 89, 126, 139 heat, 284 heating, 23, 356 height, 98, 99, 335 higher quality, 45 histogram, 108, 118, 120, 121, 123, 183, 185, 204 holding company, 21, 45 hot spots, 243

House, 61, 381, 384 housing, 157, 158 human brain, 114, 138, 139 human capital, 52 human resources, 21, 56 human subjects, 116, 139 human will, 167 humidity, 222 Hungary, 5, 44, 73 hypertension, 90, 91, 92 hypothesis, 188, 209, 259, 359, 364, 367 hysteresis, 306, 366

I identification, viii, 47, 88, 102, 126, 244 idiosyncratic, 68, 81, 83 imagery, 383 images, 383 imbalances, 72 immunity, 285 implementation, xi, 19, 20, 25, 32, 36, 37, 38, 43, 46, 154, 158, 226, 279, 281, 298, 300 impulsive, 326 incentives, ix, 22, 27, 43, 44, 48, 141, 142 incidence, 89, 90, 136 inclusion, 276 income, 2, 4, 26, 45, 400, 407 independence, 1, 9, 22 independent variable, 81, 168 India, 348 indication, 45, 333 indicators, 20, 143, 172, 173, 175 indices, 4, 30, 41, 49, 50, 89, 387, 391, 394 indigenous, 381 individual character, 49 individual characteristics, 49 individual differences, 173, 174, 177 indivisibilities, 68 induction, 114, 301 industrialized countries, 67 industry, viii, 9, 12, 20, 26, 33, 40, 54, 63, 66, 67, 83, 92, 178, 283, 285, 292, 295, 296 inefficiency, 69, 72, 73, 74 inertia, 100, 326 infarction, 89, 92 infrastructure, vii, ix, xi, xiii, 1, 2, 3, 4, 6, 7, 10, 11, 12, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 31, 32, 33, 35, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 57, 58, 59, 60, 67, 69, 72, 141, 142, 143, 156, 157, 158, 223, 234, 279, 285, 295, 302, 338, 385, 386, 392, 394 initiation, 183, 185, 191, 215 inspections, 227, 228, 229, 230, 295

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

419

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

Index instability, 243, 356 institutional change, 56 institutional reforms, 35, 36, 37, 38 institutions, 28, 37, 231, 285 instruction, 228 instruments, xii, 48, 55, 56, 99, 101, 355, 370, 388 insulation, 323 insurance, 49 integration, vii, 1, 11, 32, 35, 38, 357, 410 integrity, 215, 385, 390 intentions, 28 interaction, viii, x, xii, 83, 87, 89, 99, 114, 128, 132, 135, 149, 161, 164, 166, 167, 168, 178, 237, 238, 239, 242, 243, 244, 262, 276, 303, 304, 318, 326, 327, 329, 332, 350, 351, 352 interactions, 386 interface, 17, 178, 207, 285, 312, 320, 321, 327, 328, 334, 363 interference, 129, 130, 262, 285, 287, 293 internalization, vii, 1, 50, 51, 52, 53 internalizing, 52 International Classification of Diseases (ICD), 90, 91 international relations, 6 international trade, 6 internet, 381 interoperability, 16, 53, 70, 280, 281, 301 interrelations, 23, 31 interval, 102, 103, 107, 108, 110, 118, 120, 121, 124, 170, 183, 184, 196, 335 intervention, 12, 24, 377 interview, 148 intuition, 227 investment, 7, 12, 16, 17, 21, 24, 27, 33, 35, 40, 42, 56, 58, 70, 71, 75, 234, 300, 393, 396 investors, 34 ionizing radiation, 115 Ireland, 5, 21, 77, 81 iron, 106, 182, 199, 200, 205, 209, 212, 213, 267, 270 isolation, 11, 266, 282, 318 Israel, 379 Italy, xi, 5, 8, 10, 14, 21, 44, 45, 73, 77, 81, 87, 237, 355, 385, 397, 398 Ivory Coast, 381, 382

J Japan, 138, 181, 182, 230, 232, 234, 282, 296 job satisfaction, 167 jobs, 54, 89 joints, 222, 277, 323, 325, 326, 350, 353, 356, 377 judgment, 213, 389 justification, 54, 226, 396

L labor, 11, 20, 32, 33, 34, 35, 55, 69, 90 land, 10, 22, 23, 42, 47, 145, 157, 235, 280 land use, 42 landscapes, 381 laptop, 169 lasers, 341, 344 Latvia, 5, 9 laws, 26, 28 leakage, 94 learning, 169 legislation, 1, 2, 18, 20, 24, 28, 32, 46, 142 leisure, 50, 51, 150 leisure time, 50, 150 leukemia, 89, 136 liberalization, 18, 26, 36, 67 licenses, 19, 23, 24 life cycle, 229, 309 lifespan, 226, 293 lifetime, 321 light rail, 400 likelihood, 16, 405 limitation, 218, 229, 243 limited company, 12 linear dependence, 242 linear function, 312 linear model, 313, 348 links, 14, 28, 53, 73 Lithuania, 5, 9, 26, 44 loans, 21 local authorities, 24, 230 local government, 72 logistics, 12, 14, 54, 163 long distance, 145, 238, 280 lower prices, 36 LTD, 349 luggage, 67 lymphatic system, 137

M Macedonia, 30, 35, 37 machinery, 23 magnet, 99, 288 magnetic field, viii, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 104, 105, 106, 107, 108, 109, 110, 111, 112, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 136, 137, 138, 139, 288, 290, 291, 293, 370 magnetic properties, 114, 288 magnetic sensor, 104, 290, 293, 295, 301 magnetism, 139

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

420

Index

magnetosphere, 93 magnets, 284 maintenance tasks, xi, 279, 300 males, 92 management, vii, x, 1, 4, 12, 17, 18, 20, 21, 22, 23, 24, 25, 26, 29, 30, 31, 32, 34, 35, 36, 37, 38, 39, 42, 56, 66, 67, 69, 171, 178, 179, 217, 218, 222, 223, 224, 225, 226, 231, 232, 233, 234, 281, 282, 283, 284, 353, 358, 370, 393, 394, 396 manufacturing, 54, 55, 179, 283, 357 marginal cost pricing, 42, 143, 158 marginal costs, ix, 40, 41, 42, 44, 46, 47, 49, 58, 67, 141, 142, 143, 144, 151, 154, 156 marginal external cost, 42, 46, 50, 52 marginal social cost, ix, 38, 39, 40, 42, 43, 45, 46, 47, 48, 50, 52, 141, 142, 143 mark up, 142 market, vii, ix, 1, 2, 4, 8, 9, 10, 11, 12, 16, 17, 18, 19, 20, 21, 24, 25, 27, 28, 29, 31, 32, 34, 35, 36, 37, 38, 39, 43, 45, 48, 50, 51, 52, 55, 59, 65, 66, 70, 141, 147, 148, 149, 150, 151, 284 market access, 48 market failure, 25 market segment, 16, 17, 43 market share, 10, 11, 12, 29, 35, 36, 37, 45, 65, 66 marketing, 12, 32, 34, 35, 36 markets, 3, 6, 11, 12, 18, 19, 20, 26, 28, 45, 53, 55, 56, 66, 70, 76 marriage, 384 material degradation, 222 matrix, 31, 328 measurement, viii, 74, 76, 87, 88, 93, 98, 99, 101, 107, 109, 202, 203, 214, 215, 229, 262, 265, 276, 277, 284, 290, 295, 300, 301, 305, 324, 333, 335, 338, 339, 340, 341, 342, 343, 344, 345, 346, 347, 348, 357, 358, 359, 366, 371, 375, 376, 377 measures, vii, viii, 1, 12, 18, 19, 20, 21, 22, 25, 27, 28, 29, 30, 35, 36, 37, 40, 41, 48, 50, 51, 54, 58, 67, 72, 88, 126, 228, 265, 304, 321, 326, 344, 384 mechanical properties, 187, 349 medical care, 150 membership, 7, 28 mental load, ix, 163, 168, 169, 171, 172, 173, 175, 177, 178 merchandise, 380 mercury, 284 metals, 358, 359, 378 Mexico, 70, 73 mice, 126 migration, 333 military, 218, 220 mineral resources, 384 minority, 24

misunderstanding, 29, 164 mobility, 9, 10, 12, 54, 56, 268 modeling, 31, 69, 164, 166, 191, 192, 194, 195, 198, 212, 214, 278, 389, 391, 392 models, vii, ix, xii, 1, 34, 36, 37, 38, 75, 82, 138, 141, 146, 166, 167, 207, 214, 242, 247, 278, 303, 309, 312, 313, 315, 316, 317, 318, 321, 323, 326, 339, 349, 352, 354, 380, 387, 390, 394, 396, 397, 400, 406, 411 modern society, 396 modernization, 6, 35 modules, xi, 232, 279 modulus, 200, 306, 307, 312, 313, 314, 316, 317, 318, 325, 330, 331, 332, 334, 342, 344, 349, 351, 352, 353, 359 moisture, 222, 231 money, 57, 150, 388 monopoly, vii, 1, 11, 22, 25, 27, 38, 67 monopoly power, 11 morbidity, 90, 91, 116 morning, 72 morphology, 390 mortality, 89, 91, 92, 93, 136, 137 mortality rate, 89 motion, 101, 106, 302, 329 motivation, 33, 290 mountains, 220 movement, 2, 54, 106, 114, 127, 128, 129, 131, 132, 139, 239, 242, 320, 357, 364, 367, 369, 370 MTS, 360 multiplication, 151 myocardial infarction, 89, 91, 92, 116, 126, 136 myocardial ischemia, 126

N National Research Council, 214, 349, 353 natural disasters, 222, 387 navigation system, 281 negative externality, 143 negotiation, 25 Netherlands, ix, 5, 23, 26, 67, 68, 73, 78, 85, 163, 169, 177, 179, 340, 348, 349, 354 network, xi, 6, 7, 9, 10, 14, 16, 17, 19, 23, 25, 26, 37, 40, 44, 48, 53, 55, 69, 218, 219, 226, 233, 234, 279, 280, 352, 408 network density, 6, 69 newspapers, 380 Nigeria, 380, 384 nodes, 260, 328 North Africa, 38 North America, 285, 353, 400 Northern Ireland, 23 Norway, 78, 81, 350, 397

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Index null hypothesis, 81 numerical analysis, 184

O objectivity, 32 observations, 81, 182, 212, 213 observed behavior, 148 occupational groups, 116 OECD, 30, 33, 35, 36, 59, 61, 67, 86 oil, 48 omission, ix, 181 operator, 12, 23, 50, 61, 156, 158 opportunity costs, 39 optical fiber, 293 optimization, vii, 1, 6, 27, 230, 268 organism, 88 organizational culture, 164 orientation, 36, 38, 84, 114, 139 oscillation, 117, 118, 273 ownership, 2, 20, 21, 23, 24, 25, 27, 33, 39, 230, 400

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

P Pacific, 59, 234 pain, 52 paints, 230, 231 paradigm shift, 386 parallelism, 386 parameter, 75, 81, 99, 152, 258, 315, 331, 346, 356, 389, 390, 404, 405, 406 parameter estimates, 75, 81 Parliament, ix, 24, 141 particles, 93, 114 partnership, 27, 36, 38 passive, vii, x, 237, 265, 270 pathogenesis, 126 pathways, 126 pattern recognition, 294 performance indicator, 172, 173, 226 periodicity, 293, 294 permeability, xi, 93, 114, 279, 287 permit, 19, 44 personal life, 55 personality, 227, 410 PHARE, 29 phase shifts, 117, 118, 121, 123 phase transformation, 243 photographs, xii, 379, 380, 381, 382 physical mechanisms, 347 physics, 94 physiology, 137 planned investment, 53

421

planning, vii, ix, xiii, 17, 24, 31, 34, 39, 41, 42, 44, 47, 56, 72, 73, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 174, 178, 179, 230, 380, 399, 400 plastic deformation, 207, 304, 320, 321, 326, 351 Poland, 5, 8 polarity, 97 polarization, viii, 87, 89, 99, 116, 117, 120, 121, 123, 125, 126, 358 police, 47 policy instruments, 54 policy makers, 407 politics, 380 pollutants, 17 pollution, 10, 40, 42, 48, 49, 51, 55, 56, 88, 238 poor, 11, 32, 33, 37, 55, 218, 225, 227, 230, 233, 309, 310, 333, 334 population, 4, 10, 33, 63, 69, 89, 149, 153, 154, 157, 232 porosity, 322 ports, 14, 380, 381 Portugal, 5, 23, 77, 81, 303, 351, 352 positive externalities, 28 power, viii, 23, 30, 31, 36, 47, 88, 94, 95, 98, 101, 102, 104, 105, 106, 112, 113, 116, 117, 124, 125, 126, 137, 142, 241, 249, 250, 270, 272, 282, 285, 316, 318, 323 prediction, 80, 184, 185, 199, 211, 212, 214, 215, 232, 276, 353 preference, 50, 66, 83, 116, 155, 175, 227, 306, 400, 403 premiums, 49 pressure, xii, 66, 146, 147, 152, 158, 196, 207, 243, 244, 246, 248, 249, 250, 252, 265, 313, 314, 321, 328, 355, 371, 372, 377 prevention, 282 price index, 80 prices, 2, 23, 41, 42, 49, 51, 54, 79, 80, 147, 148, 149, 150, 151 private enterprises, 27 private investment, 53 private sector, 24, 37, 54 privatization, 26, 27, 32 probability, 71, 108, 116, 117, 124, 146, 184, 209, 215, 280, 310, 312, 389, 404 probability distribution, 116, 117, 215 probe, xii, 295, 355, 361, 363, 372 problem solving, 166, 167 process control, 26, 168 producers, 381 production, 11, 31, 54, 66, 69, 94, 143, 150, 179, 384 productivity, 11, 21, 30, 32, 33, 34, 35, 50, 53, 142, 280, 300 profit, 37, 48, 53

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

422

Index

profitability, 66, 67, 280, 300 program, 22, 55, 73, 118, 165, 170, 202, 276, 328 propagation, 154 prototype, 298, 299, 340 PSD, 117 public enterprises, 24 public interest, 24 public sector, 38, 56 public service, 18, 21, 22, 24, 25, 39, 42, 72, 74 public-private partnerships, 16 pulse, 290, 296, 311, 336 purchasing power, 33

Q qualifications, 25 qualitative controls, 15 quality assurance, 231 quality of life, 54, 55 quality of service, 11, 17, 21, 43, 57 quartz, 99

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

R race, 89 radar, 295 radiation, 137, 138, 239, 241, 252, 254, 258, 266, 268, 276, 278, 284, 334 radio, 284, 301, 370 radius, 242, 243, 255, 258, 266, 274, 292 rainfall, 320 raw materials, 380 reading, 40 real estate, 163 real time, 280 reality, 93, 146, 178, 300, 339 reasoning, 164, 166, 168, 169, 170, 173, 174, 175 reception, 152, 296 reciprocity, 26 recognition, viii, 88 reconstruction, 333, 334 recovery, 32, 39, 43, 45, 46, 47, 49, 50, 143 recreation, 142, 158, 400 redistribution, 211 redundancy, 56 reflection, 50, 293 reforms, 24, 26, 32, 35, 37, 60 refraction index, 293 refractive index, 293 regression, 148, 149, 151, 152, 322 regression analysis, 151 regression method, 148 regulation, vii, 1, 2, 24, 25, 27, 38, 56, 66, 69, 224, 231

regulations, 24, 25, 28, 144, 285, 286, 388 regulators, 31, 63, 72, 83 regulatory bodies, 2, 25, 54 rehabilitation, 32, 33, 35, 37 reinforcement, 326, 333, 348, 352 relationship, 32, 99, 143, 169, 242, 263, 276, 296, 297, 306, 347, 377, 404 relationships, 40, 41, 46, 139, 316, 318, 390, 394 relaxation, 207 relevance, 120 reliability, xi, 24, 57, 58, 100, 214, 279, 280, 285, 295, 300, 370, 372, 377, 385 repair, 4, 22, 23, 234 repetitions, 311, 320, 321 Republican, 384 reputation, 24 reserves, 17 resettlement, 56, 380 resistance, 32, 196, 283, 295, 307, 346 resolution, 339 resources, vii, ix, x, 32, 142, 143, 163, 177, 179, 217, 218, 227, 229, 230, 233, 234, 384 respiratory, 90 restructuring, vii, 1, 11, 18, 19, 21, 22, 30, 32, 33, 34, 36, 37, 54 retention, 19 returns, 53 revenue, 32, 46, 223 rhythm, 139 risk, xiii, 35, 49, 50, 52, 71, 72, 75, 80, 83, 84, 89, 90, 91, 92, 116, 126, 136, 149, 156, 168, 310, 385, 387, 388, 389, 390, 391, 392, 393, 394, 396 risk assessment, 387, 388, 389, 396 risk aversion, 71, 72, 75, 80, 83, 84 risk factors, 90 risk management, xiii, 385, 387, 396 robustness, xi, 279, 286, 293, 300, 331 rolling, x, xii, 6, 9, 10, 15, 20, 21, 25, 32, 34, 35, 37, 67, 69, 76, 83, 84, 145, 164, 170, 174, 176, 237, 239, 241, 242, 243, 244, 247, 252, 259, 260, 262, 265, 266, 276, 277, 278, 285, 303, 305, 335, 338, 339, 340, 380 Romania, 5, 8, 26, 28, 30, 35, 37, 44, 301 rotation axis, 99 rotations, 115 roughness, xii, 161, 242, 243, 260, 262, 263, 270, 276, 277, 355, 372, 373, 377 routing, 12, 38 R-squared, 77, 78, 82 rubber, 244, 252, 253, 254, 255, 259, 266, 267, 270, 272, 335 Russia, 87, 97, 137, 139, 140

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

Index

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

S sabotage, 390, 394 safety, xi, xii, 19, 20, 22, 23, 25, 47, 53, 54, 57, 58, 147, 149, 214, 224, 225, 227, 228, 279, 280, 282, 283, 285, 300, 308, 311, 377, 385, 386, 390 sales, 29 salt, 222, 231 sample, 36, 81, 102, 110, 148, 177, 202, 322, 393 sampling, viii, 87, 100, 101, 102, 110, 124, 200, 363 SAP, 202 Sarajevo, 7 satellite, 178, 282, 388 saturation, 12, 288 savings, 30, 38 scarcity, 40, 42, 44, 48, 49, 143 scatter, 325, 352 scheduling, vii, ix, 163, 166, 167, 168, 177, 178, 179 school, 158, 400 scores, 69 search, 173 searching, 169 security, xii, 28, 29, 32, 35, 57, 225, 385, 386, 387, 389, 390, 393, 394, 396, 397, 398 sensing, 114 sensitivity, 51, 139, 146, 148, 358, 359, 360, 361, 374, 375, 391 sensors, xi, 100, 101, 229, 246, 279, 283, 284, 285, 286, 287, 288, 290, 291, 292, 293, 294, 295, 296, 298, 300, 301, 302, 340 separation, 1, 11, 19, 22, 26, 27, 67, 264 service quality, 21, 56 severity, 226, 234, 304, 326 shape, 2, 202, 205, 209, 213, 239, 240, 241, 247, 255, 258, 262, 266, 268 shareholders, 13 shares, 23, 72 shear, 190, 192, 196, 200, 272, 313, 314, 315, 317, 318, 320, 321, 322, 328, 331, 358 shear deformation, 320 shear strength, 190, 320, 322 short run, ix, 40, 42, 80, 141, 142, 158 Sierra Leone, 380 sign, 288 signalling, xi, 279, 280, 281, 284, 285, 300 signals, xi, 49, 57, 97, 142, 279, 282, 283, 285, 290, 291, 296, 360, 364, 372 signal-to-noise ratio, xi, 279, 286, 296 similarity, 144 simulation, 41, 248, 249, 250, 326 sine wave, 273 Singapore, 234 skills, 32, 49, 56

423

skin diseases, 90 sleep disturbance, 144, 146, 147 Slovakia, 5 smoking, 70, 72, 90, 92 social change, 58, 384 social class, 89 social costs, 38, 42, 43, 54, 150 social marginal cost, 143, 144, 157 social obligations, viii, 32, 63, 73, 76, 83, 84 social relations, 7 social relationships, 7 social welfare, 48 software, viii, 87, 116, 117, 125, 164, 166, 291, 329, 363, 388 soil, 51, 306, 316, 320, 322, 323, 333, 335, 348, 367, 383 soil pollution, 51 South Africa, 338 Soviet Union, 90, 91, 230, 231 Spain, 5, 8, 21, 67, 73, 78, 80, 81, 279, 285, 301, 302, 397 spare capacity, 26 spatial learning, 139 spectrum, 88, 106, 107, 111, 112, 116, 135, 139, 260 Sri Lanka, 181, 199, 212, 213, 215 stability, 100, 338, 346, 371, 376 stages, 20, 28, 51, 231 stakeholders, x, 12, 217, 225, 231, 233 standard deviation, 76, 172, 184, 310, 311, 325 standards, xi, 14, 28, 32, 34, 37, 47, 53, 57, 70, 183, 200, 218, 219, 220, 221, 230, 231, 233, 281, 296, 300, 303, 346, 385 statistics, 81, 117, 387, 390, 406 steel, 94, 106, 194, 195, 196, 197, 198, 213, 214, 215, 218, 220, 221, 222, 224, 225, 226, 227, 229, 230, 231, 233, 234, 244, 246, 252, 257, 284, 342, 371, 378 stock, 6, 9, 10, 20, 21, 23, 25, 32, 34, 35, 37, 54, 67, 71, 72, 73, 76, 83, 84, 164, 170, 172, 174, 176, 277, 285, 326, 380 storage, 102 storms, 93, 97 strain, ix, 181, 182, 183, 185, 186, 187, 200, 214, 215, 295, 312, 315, 318, 320, 321, 322, 323, 335, 353, 358 strategies, viii, x, 32, 34, 40, 63, 84, 161, 217, 218, 226, 229, 230, 232, 233, 234, 237 strength, 46, 93, 98, 99, 114, 115, 129, 153, 191, 194, 196, 200, 215, 322, 323, 333 stress fields, 192 structural defects, 218, 228 structuring, 179 subgroups, 90

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

424

Index

subjectivity, 396 subsidization, 29 subsistence, 55 subsistence farming, 55 suicide, 136 summaries, 319 summer, 80, 220, 365, 380 superimposition, 312 superiority, 406 supervision, 24, 280, 396 suppliers, xii, 67, 72, 385 supply, 2, 4, 12, 17, 20, 27, 47, 49, 53, 54, 72, 74, 76, 101, 111, 280, 285, 298 supply chain, 17 suppression, 278 surface area, 241 surplus, 175 surveillance, 390, 395 survival, 89 sustainability, vii, 2, 42, 48, 54, 55, 56, 67 sustainable development, 55 Sweden, 5, 10, 21, 23, 26, 27, 36, 45, 73, 78, 81, 136, 141, 150, 159, 161, 162, 303, 306, 327, 342, 347, 350, 351, 352, 354 switching, 410 Switzerland, 14, 81, 101, 301, 346 symbols, 100 symmetry, 260, 327 symptoms, 226 synthesis, 352

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

T tactics, 166 targets, 29, 31, 66, 67, 301 tariff, 26, 43, 44, 45 task allocation, 168 task performance, ix, 163, 164, 168, 169, 171, 172, 173, 175, 177 taxation, 49 taxis, 400 taxonomy, 386 technical efficiency, 54 technological change, 58 telecommunications, 38 temperature, xii, 243, 294, 295, 355, 356, 357, 358, 360, 361, 362, 363, 364, 365, 366, 367, 368, 369, 372, 374, 375, 376, 377 temperature dependence, 377 tensile strength, 190, 193, 194 tension, 193, 196, 200, 206, 214 terminals, vii, 2, 12, 14, 16, 17, 31, 33, 41, 403 territory, 4, 6, 7, 381 terrorism, 390, 394

test data, 184 Thailand, 278 thermal expansion, 243, 357, 360, 371 threat, 30, 389, 390, 392, 393, 394, 396 threats, xiii, 29, 31, 385, 386, 387, 388, 389, 390, 392, 394 threshold, 119, 123, 124, 282, 287, 289, 290, 293, 296, 297, 322 threshold level, 119, 123 thresholds, 238, 286, 298, 299 time periods, 229 time pressure, 168 time series, 41, 123 time use, xii, 355 tin, 355 TMC, 153 topology, 283, 286 total costs, 48 total factor productivity, 30 total revenue, 50 tracking, 29, 31 trade, 7, 22, 32, 39, 55, 56, 66, 266, 272, 280, 380, 383 trade union, 32 trade-off, 39, 55, 266, 272 training, x, 32, 217, 234 traits, 93 trajectory, 164 transducer, 289, 335, 371 transfer payments, 2 transformation, 55, 66, 192, 342, 409 transformations, 192 transition, 323, 326, 327, 329, 330, 331, 347 transitions, 304, 347 transmission, 95, 276, 290 transmits, 291, 293 transparency, 11, 18, 21, 26, 32, 42, 46, 47 transport costs, 27, 46, 53, 56 transportation, vii, viii, x, xiii, 7, 24, 25, 37, 42, 48, 50, 52, 65, 66, 68, 72, 74, 76, 87, 219, 235, 237, 238, 280, 281, 300, 302, 396, 399, 406, 407, 410 transshipment, 12, 13, 14, 16 trauma, 90 trees, 387 trust, 168, 175 tumors, 89 Turkey, 70 turnout, 326, 327, 347, 354 turnover, 2, 29

U ultrasound, 295, 358

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest

425

Index uncertainty, viii, 63, 68, 71, 72, 75, 76, 80, 83, 155, 339, 364, 370, 377, 389, 390, 391, 394 uniform, 175, 207, 277, 294, 307 unions, 223 United Kingdom, 5, 21, 23, 25, 26, 27, 59, 81 United Nations, 59 urban areas, 55, 94, 138, 382 urbanisation, 384 urbanization, 142

Copyright © 2008. Nova Science Publishers, Incorporated. All rights reserved.

V Valencia, 397 validation, 187, 203, 276, 278, 287, 354, 397 validity, x, 19, 182, 212, 406 values, ix, 4, 30, 39, 50, 82, 83, 93, 96, 98, 101, 105, 106, 107, 108, 109, 110, 113, 114, 116, 117, 118, 124, 125, 141, 144, 148, 149, 150, 151, 154, 156, 157, 158, 160, 200, 210, 211, 214, 241, 242, 247, 255, 258, 260, 266, 268, 304, 307, 315, 318, 320, 327, 329, 331, 333, 339, 341, 346, 362, 366, 390, 393, 396, 410 vandalism, 286, 288, 290, 388, 390, 394 variability, viii, 87, 89, 109, 116, 125, 126, 139, 334, 349, 370, 373, 377 variable costs, 39, 40, 41, 44, 46 variables, 41, 69, 74, 75, 76, 80, 81, 82, 149, 168, 169, 213, 230, 244, 259, 325, 331, 406, 407, 410 variance, 405 variation, xii, 39, 45, 76, 79, 117, 149, 165, 183, 202, 209, 210, 211, 212, 287, 288, 289, 293, 303, 304, 307, 323, 324, 325, 326, 329, 330, 331, 347, 355, 357, 358, 360, 362, 366, 369, 377, 405 vector, 74, 93, 101, 102, 115, 117, 119, 123, 126, 148, 405 vehicles, viii, xi, 6, 14, 15, 16, 25, 34, 37, 49, 50, 51, 55, 63, 67, 70, 71, 75, 82, 83, 84, 223, 277, 278, 279, 280, 295, 305, 309, 311, 335, 344, 347 vein, 193 velocity, 127, 128, 129, 130, 133, 241, 243, 335, 340, 377 vessels, 222, 230 vibration, 229, 239, 250, 252, 255, 258, 259, 260, 263, 265, 266, 267, 270, 273, 274, 276, 277, 278, 325, 334, 335, 350, 354 victims, 52 Vietnam, x, 217, 218, 219, 220, 221, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235 vision, 294, 295 visualization, 117 voiding, 305 volatility, 76, 80, 81, 83, 84 voting, 397 vulnerability, 389, 390, 392

W wage rate, 50 wages, 30 walking, 130, 400, 402, 404, 408, 410 war, 218 waterways, 20, 55, 65 wavelengths, 242, 260, 262, 263, 293 weakness, 142, 149, 150, 333 wealth, 4, 25, 49 wear, xii, 38, 39, 42, 143, 146, 158, 243, 293, 303, 304, 357, 370 web, 20, 239, 244, 247, 248, 249, 250, 251, 254, 256, 257, 258, 259, 261, 266, 267, 335, 358, 365, 371 web pages, 20 welding, 356 welfare, 39, 53, 67, 144, 150 welfare economics, 39 welfare loss, 144, 150 West Africa, xii, 379, 380, 381, 382, 383, 384 Western Europe, 34 windows, viii, 87, 89, 99, 115 winning, 269 wires, xi, 93, 279, 295, 298 workers, viii, 87, 88, 89, 90, 92, 136, 225 working conditions, 297 working hours, 166 working population, 92 workload, 164, 168, 177 workplace, 90, 102, 104, 107, 110, 111, 158 World Bank, 56, 60, 61, 86 World War I, 379 World Wide Web, 58, 59, 60, 61

Y yield, 72, 177, 192, 194, 200, 210, 323, 368 Yugoslavia, 7

Railway Transportation : Policies, Technology and Perspectives, edited by Nicholas P. Scott, Nova Science Publishers, Incorporated, 2008. ProQuest