Table of contents : Preface Acknowledgements Contents Contributors Acronyms Design and Modelling Modelling of the Dynamic Train-Track-Substructure Interaction for Track Monitoring 1 Introduction 2 Vehicle-Track-Substructure Model 2.1 Vehicle-Track Dynamic Model 2.2 Track-Substructure Model 2.3 Simulation of Wheel-Rail Force Transfer 2.4 Rail Points for Analysis 3 Rail Acceleration of Ballasted Track 3.1 Distribution of Acceleration 3.2 Effect of Vehicle Speed 3.3 Effect of Track Irregularity 3.4 Effect of Offset of Wheel-Rail Patch 4 Conclusions References Statistical Approach to Model Track Dynamics Towards the Monitoring of Railway Turnouts 1 Introduction 1.1 Literature Review 1.2 Main Contribution 2 Track Side Measurement Set-Up 3 Robust Estimation of the Track Resonance Frequencies 3.1 Empirical Mode Decomposition 3.2 N4SID Subspace Identification Method 3.3 Frequency Estimation 4 Method Implementation 4.1 Model Validation and Robustness 5 Statistical Modelling of the Track Resonance Frequencies 5.1 Towards Condition Monitoring of Railway Turnout 5.2 Discussion 6 Conclusions References Improving Switches and Crossings Performance and Reliability 1 Introduction 1.1 Background 1.2 Methodology 2 Case Studies 2.1 Repeat Shearing of Screws at a Crossing 2.2 Componentry Optimisation at a Sharp Angled Double Junction 3 Conclusions Reference Metallurgical Characterization Damage and Microstructure Evolution in Cast Hadfield Steels Used in Railway Crossings 1 Introduction 2 Material and Experimental Methods 3 Results and Discussion 3.1 Evolution of Deformation Twinning Under Impact and Rolling Contact Fatigue Loading 3.2 Solidification Porosity and Non-Metallic Inclusions and Their Role in Damage 3.3 Strain-Induced Austenite-to-Martensite Transformation Under the Impact and RCF Loading and Its Effect on the Damage in Hadfield Steel 3.4 Effect of the Grain and Twin Boundaries on Fatigue Crack Growth in an Undeformed Cast Hadfield Steel 4 Conclusions References Metallurgical Investigation of Crossing Noses 1 Introduction 2 Experimental 3 Results and Discussion 3.1 Head Hardened Pearlitic Steel Nose (Welded)—Microscopy 3.2 Head Hardened Pearlitic Steel Nose (Welded)—Tomography 3.3 Austenitic Manganese Steel Nose—Microscopy 3.4 Austenitic Manganese Steel Nose—Tomography 3.5 Austenitic Manganese Steel Nose—Residual Stresses 4 Conclusions References Analysis of Sleeper Screw Failures 1 Introduction 2 Experiment 3 Results and Discussion 3.1 Microstructure 3.2 Three-Dimensional Mapping of Screw Shapes and Cracks 3.3 Fracture Surface Analysis 3.4 Sub-Surface Deformation, Cracks and Corrosion 4 Conclusion References Condition Monitoring and Asset Management Data Sources and Research Models for Turnouts 1 Introduction 2 Current Inspection Method 3 Analysis of the Current Inspection Data 4 Methods of Inspecting Turnouts Automatically 5 Automated Turnout Inspection 6 The Standard Element Approach 7 Lifetime Limiting Components 7.1 Renewal Project List 7.2 Analysis with Standard Elements 7.3 Questionnaires and Expert Interviews 8 Summary 9 Future Work References CoMPAcT-Data Based Condition Monitoring and Prediction Analytics for Turnouts 1 Introduction 2 Literature Review 3 Proposed Methodology 4 Data Source, Collection and Cleaning 4.1 Suitability for Turnout Description 4.2 Post-Positioning Algorithm 5 Ballast Condition Monitoring 6 Results and Discussion 6.1 Critical Area Identification 6.2 Description of Tamping Actions 6.3 Description of the Ballast Condition 7 Conclusion References Monitoring of Switches and Crossings/Tracks Using Smart Sensors 1 Introduction 2 System Architecture 2.1 Wireless Sensor Network Hardware Platform 2.2 Power Consumption of Wireless Sensor 3 System Evaluation and Validation 3.1 Feasibility Test of Concept 3.2 Antenna Test 3.3 Rail Coverage Test 3.4 Field Tests 4 Conclusions References Labelling the State of Railway Turnouts Based on Repair Records 1 Introduction 2 Data 2.1 Repair Records: Combined Repair and Tamping/Grinding Data 2.2 Recording Car Measurements 3 Methods 3.1 Principal Component Analysis 3.2 Clustering 3.3 Local Regression (Loess) 4 Results 4.1 Labelling Through Historical Repair Records 4.2 Validation of the Labelling Through the Track Recording Car Data 5 Conclusions References Automatic Detection of Rail Defects from Images 1 Introduction 2 The Data 2.1 The Rail Image 2.2 Image Groups 0 and 1 3 Method 3.1 Detecting Rails in Images 3.2 Detecting Defects on Rails 3.3 Comparing the Two Groups of Rail Images Using Three Defect Detection Methods 4 Results 4.1 Detecting Rails in Images 4.2 Detecting Defects on Rails 5 Conclusions References Deep Learning for Automatic Railway Maintenance 1 Introduction 2 Data 2.1 Images from the Recording Car 2.2 Defects and Isolation Joints Classified on the Images 3 Method 3.1 Labelling and Data Augmentation 3.2 Performance Metrics 4 Results 4.1 Training and General Performance Results 4.2 Performance Results for Individual Classes 4.3 Application Interpretation 5 Discussion and Conclusion 6 Funding 7 Appendix 1: You Only Look Once Object Detection 7.1 YOLO Object Detecion Architecture 7.2 The Loss Function 7.3 Back-Propagation in Yolov3 7.4 Pre-processing 8 Appendix 2: Results 8.1 Training and General Performance Results References The Effects of Renewal and Tamping on Ballast and Track Geometry in Turnouts 1 Introduction 2 Data Cleaning and Alignment 3 Methodology 3.1 Geometric Quality Indicators 3.2 The Applied Fractal Analysis Method 3.3 Track Geometry Degradation 3.4 Renewal/Tamping Effect 4 Results 4.1 Recovery Effects 4.2 Effects on Ballast and Vertical Geometry Degradation Rates 4.3 Comparison with Open Track 5 Conclusions References The Organic Growth of an Asset Management System at ProRail 1 Introduction 2 Development of ProRail Vision on Asset Management 2.1 Role of an Asset Manager 2.2 The Essence of Asset Management 3 Outsourcing of Maintenance 3.1 Performance Specification and Control 3.2 The Development of Outsourcing 4 Risk Management 4.1 One Risk Instrument, Multiple Applications 4.2 Sector Knowledge Platform 5 Organic Growth Model of an Asset Management System 5.1 Lessons Learned by ProRail 6 Conclusions References Ways to Improve the Performance (And Cost) of Switches 1 Introduction 2 Differentiation in Maintenance Concept Depending on Usage 3 Focus on the Most Vulnerable Switch Types 4 Focus on Switches with Repetitive Failures 5 Improve the Design of (High Speed) Switches 6 Increase Network Capabilities by Reducing the Amount of Switches 7 Use of Risk Management Technique/Instruments 8 Introduce Risk-Based Performance Management 9 Uniform Measuring of Switch Condition Degradation 10 Fixed Tele-Monitoring Systems to Measure Quality Degradation 11 Explore the Possibilities of Flexible, Cost Efficient LoRa Sensors 12 A Condition Dashboard for Switches: Maintenance Performance Indicator 13 Conclusions