Handbook of Large Hydro Generators: Operation and Maintenance: 66 (IEEE Press Series on Power and Energy Systems) [1 ed.] 0470947578, 9780470947579

Citation preview

HANDBOOK OF LARGE HYDRO GENERATORS Operation and Maintenance

IEEE Press 445 Hoes Lane Piscataway, NJ 08854 IEEE Press Editorial Board Ekram Hossain, Editor in Chief Jón Atli Benediktsson Xiaoou Li Saeid Nahavandi Sarah Spurgeon

David Alan Grier Peter Lian Jeffrey Reed Ahmet Murat Tekalp

Elya B. Joffe Andreas Molisch Diomidis Spinellis

HANDBOOK OF LARGE HYDRO GENERATORS Operation and Maintenance GLENN MOTTERSHEAD STEFANO BOMBEN ISIDOR KERSZENBAUM GEOFF KLEMPNER

© 2021 by The Institute of Electrical and Electronics Engineers, Inc. All rights reserved. Published by John Wiley & Sons, Inc., Hoboken, New Jersey. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data: Names: Mottershead, Glenn, author. | Bomben, Stefano, author. | Kerszenbaum, Isidor, author. | Klempner, Geoff, author. Title: Handbook of large hydro generators : operation and maintenance / Glenn Mottershead, Stefano Bomben, Isidor Kerszenbaum, Geoff Klempner. Other titles: Operation and maintenance of large turbo generators Description: Eleventh edition. | Hoboken, New Jersey : Wiley, [2020] | Series: IEEE press series on power engineering | Original edition published under title: Operation and maintenance of large turbo generators / Geoff Klempner, Isidor Kerszenbaum. | Includes bibliographical references and index. Identifiers: LCCN 2020035414 (print) | LCCN 2020035415 (ebook) | ISBN 9780470947579 (hardback) | ISBN 9781119524182 (adobe pdf) | ISBN 9781119524168 (epub) Subjects: LCSH: Turbogenerators. | Turbogenerators–Maintenance and repair. Classification: LCC TK2765 .K58 2020 (print) | LCC TK2765 (ebook) | DDC 621.31/3–dc23 LC record available at https://lccn.loc.gov/2020035414 LC ebook record available at https://lccn.loc.gov/2020035415 Set in 10/12pt Times by SPi Global, Pondicherry, India 10 9 8 7 6 5 4 3 2 1

To our families Joanne and sons Stephen, Brian, Jeffrey Mottershead and their families Victoria, Kristina, and Kayla Bomben Susan Klempner Jackie, Livnat, and Yigal Kerszenbaum, and To the operators, technicians, and engineers in the power stations around the world who keep the lights on and the power flowing around the clock through adverse conditions for the benefit of everyone.

CONTENTS PREFACE

xi

ABOUT THE AUTHORS

xv

ACKNOWLEDGMENTS

xvii

CHAPTER 1

1.1 1.2 1.3 1.4 1.5 1.6 1.7

Introduction to Basic Notions on Electric Power Electrical–Mechanical Equivalence 6 Alternating Current (AC) 6 Three-Phase Circuits 13 Basic Principles of Machine Operation 14 The Synchronous Machine 18 Synchronous Machine: Basic Operation 23

CHAPTER 2

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11 2.12 2.13 2.14 2.15 2.16 2.17 2.18 2.19 2.20 2.21

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES 1

GENERATOR DESIGN AND CONSTRUCTION

Stator Core 36 Stator Frame 50 Electromagnetics 54 Core-End Heating 62 Flux and Armature Reaction 62 Stator Core and Frame Forces 64 Stator Windings 65 Stator Winding Wedges 79 Endwinding Support Systems 85 Stator Winding Configurations 86 Stator Terminal Connections 88 Rotor Rim 91 Rotor Spider/Drum 103 Rotor Pole Body 106 Rotor Winding and Insulation 110 Amortisseur Winding 116 Slip/Collector Rings and Brush Gear 119 Cooling Air 122 Rotor Fans/Blower 124 Rotor Inertia, Torque, and Torsional Stress Thrust and Guide Bearings 128

1

35

125

vii

viii

CONTENTS

CHAPTER 3

3.1 3.2 3.3 3.4 3.5 3.6

177

235 237

MONITORING AND DIAGNOSTICS

241

GENERATOR PROTECTION

Basic Protection Philosophy 291 IEEE Device Number 295 Brief Description of Protective Functions Tripping and Alarming Methods 307

CHAPTER 7

7.1 7.2 7.3 7.4 7.5 7.6

174

Generator Monitoring Philosophies 242 Simple Monitoring with Static High-Level Alarm Limits Dynamic Monitoring with Load Varying Alarm Limits Artificial Intelligence (AI) Diagnostic Systems 247 Monitored Parameters 250 Radio Frequency Monitoring 273 Capacitive Coupling 274 Stator Slot Coupler 276 Rotor 278 Excitation System 286

CHAPTER 6

6.1 6.2 6.3 6.4

167

OPERATION AND CONTROL

Basic Operating Parameters 177 Operating Modes 188 Machine Curves 190 Special Operating Conditions 200 Basic Operation Concepts 208 System Considerations 225 Grid-Induced Torsional Vibrations Excitation and Voltage Regulation

CHAPTER 5

5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10

157

Oil Systems 157 Stator Surface Air Cooling System 161 Bearing Cooling Coils and Water Supply 165 Stator Winding Direct Cooling Water System Excitation Systems 171 Excitation System Performance Characteristics

CHAPTER 4

4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8

GENERATOR AUXILIARY SYSTEMS

291

296

INSPECTION PRACTICES AND METHODOLOGY

Site Preparation 311 Experience and Training 314 Inspection Frequency 317 Generator Accessibility 318 Inspection Tools 319 Inspection Forms 321

243 244

311

CONTENTS CHAPTER 8

8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 8.10

419

AUXILLIARIES INSPECTION

MAINTENANCE AND TESTING

Stator Core Mechanical 513 Stator Core Electrical Tests 518 Stator Winding Mechanical Tests 531 Stator Winding Electrical Tests 534 Rotor Mechanical Testing 568 Rotor Electrical Testing 583 Bearings 590 Heat-Run Testing 590

411

417

Excitation: Field Breaker 465 Excitation: Static Exciter Components 470 Brushless Exciter 470 Static Exciter Transformer 472 Excitation: Rotating Exciters 473 Excitation: Sliprings, Commutator, and Brushes Surface Air Coolers 499 Fire Protection 502 General Items 504 Thrust and Guide Bearing 507 Miscellaneous Auxiliaries 510

CHAPTER 11

11.1 11.2 11.3 11.4 11.5 11.6 11.7 11.8

ROTOR INSPECTION

Rotor Spider with Shrunk Laminated Rims Rotor Rim 430 Rotor Poles 436 Rotor Brakes 458

CHAPTER 10

10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9 10.10 10.11

337

Stator Frame Soleplates 338 Stator Frame: General 349 Stator Core Air Ducts 354 Stator Core Laminations 356 Stator Core Clamping System 378 Stator Coils/Bars 389 Flow Restriction in Water Cooled Stator Windings 396 Stator Wedging System 398 Stator Endwinding 405 Main and Neutral End Leads, Cables, VTs, CTs, and Insulators

CHAPTER 9

9.1 9.2 9.3 9.4

STATOR INSPECTION

ix

465

481

513

x

CONTENTS

CHAPTER 12

12.1 12.2 12.3 12.4 12.5 12.6 12.7 12.8 12.9 12.10 12.11 INDEX

MAINTENANCE PHILOSOPHIES, UPGRADES, AND UPRATES

General Maintenance Philosophies 595 Operational and Maintenance History 597 Maintenance Intervals/Frequency 598 Planned Outages 599 Rehabilitation, Uprating/Upgrading and Life Extension 601 Excitation System Upgrades 611 Workforce 627 Spare Parts 628 Effect of Uprating on Generator Life 629 Required Information, Tests and Inspection Prior to Uprating/Upgrading Maintenance Schedule After Uprating 632

595

631

633

PREFACE Hydro generators in different plants are rarely identical, and it is not uncommon for small, medium, or large utilities to have a significant variety of unit sizes, origins, and vintage in their fleet of generators. Among these generators, there might be units 60 or more years old with all original components including stator windings due to the robust nature of this class of machinery. Some might still have a pilot and main rotating exciter or a static pilot with a main rotating exciter, or have full static excitation. Additionally, there may be units operating over a wide range of rotating speeds in 50 and 60 Hz power grids with a few still operating at 25 Hz or other frequencies. All are designed and built by a long list of manufacturers from around the globe using a variety of materials and methods governed by different standards. These generators are still owned by traditional utilities and also owned by new deregulated independent power producers (IPPs) that acquire traditional utilities from all over the world. There are new large hydro plants still being built for traditional utilities and IPPs to the most modern design standards and manufacturing methods. The owners of all types are called upon to operate and maintain an incredibly wide variety of machines. The reasons why one may find so many “old” units still in operation is not difficult to determine. First of all, historically generators have been designed and manufactured with the intent to be robust enough to last typically 50 years or more. Second, replacing operating units is very capital-intensive and done only when a catastrophic failure has occurred or some significant economic benefit is possible only with complete replacement. Third, although typically designed to last many years, large hydro generators are known to be capable of having their lives extended far beyond 50 years if well maintained and operated. There are some generators in operation today that were placed in service in 1896, an example is the Dominion Power and Transmission Company’s units in Decew Falls, Ontario, Canada, now operated by Ontario Power Generation. To continue to operate reliably older generators require replacement of at least some major components, such as the armature winding, rotor winding insulation, or replacing the entire stator frame and core or rotor spider. Managing the scope and timing of major maintenance is always a challenge. There are copious amounts of information about the operation, maintenance, and troubleshooting of large hydro generators in many publications and online communities. All vendors at one stage or another have produced and published interesting literature about the operation of their generators. Institutions such as EPRI, CIGRE, IEC, IEEE, CEATI, and other organizations cover various aspects of the operation and maintenance of generators in general, but often have difficulty xi

xii

PREFACE

providing specific information that may help troubleshoot a particular generator design or operating problem. It is no wonder then that with so many dissimilar units in operation having different operating conditions, we are often forced to call the “experts,” who tend to be folks almost as old as the oldest units in operation. These are individuals who have crawled around, inspected, tested, and maintained many diverse generators over the years. In doing so, they have retained knowledge about the different designs, materials, and manufacturing characteristics, typical problems, and workable solutions. This type of expertise cannot easily be learned in a classroom. Unfortunately, not every company retains an individual with the breadth and depth of expertise required for troubleshooting the generators. In fact, with the advent of deregulation, many small nonutility (third-party) power producers operate small fleets of generators without the benefit of in-house expertise. In lieu of that, they depend heavily on OEMs and independent consultants. Large utilities in many places have also seen their expertise dissipate, not to a small extent because of a refocusing of management priorities. All these developments are occurring at the same time that these aging units are called to operate in a deregulated or semideregulated world which typically results in an increase in load-cycling. Some effort has been made over the years to capture the experts’ knowledge and make it readily available to any operator as a computer-based expert system. However, difficulty with adaptation of the associated computer programs to the many different types of generators and related equipment in existence has proved to be the Achilles heel of this technology. There is just no substitute for someone who understands machine design and has the required experience to recognize the significance of visual indications while crawling through a machine on a regular basis. This book is designed to partially fill the gap by offering a comprehensive view of many issues related to the operation, inspection, maintenance, and troubleshooting of large hydro generators. All of the information in the book is the result of many years of combined hands-on experience of the authors, which at the time of this writing, amounts to 157 years. It was written with the machine’s operator and inspector in mind, as well as providing a guide to uprating and life enhancement of large hydro generators. Although not designed to provide a step-by-step guide for the troubleshooting of large hydro generators, it serves as a valuable source of information that may prove to be useful during troubleshooting activities. The topics covered are also cross-referenced to other sources. Many such references are included to facilitate those readers interested in enlarging their knowledge of a specific issue under discussion. For the most part, theoretical equations have been left out, as there are several exceptionally good books on the theory of operation of synchronous machines. Those readers who so desire can readily access those books, several references are cited. This book, however, is about the practical aspects that characterize the design, operation, and maintenance of large hydro generators, and a number of practical calculations used commonly in maintenance and testing situations have been added.

PREFACE

xiii

Chapter 1 (Principles of Operation of Synchronous Machines) provides a basis of theory for electricity and electromagnetism upon which the machines covered in this book are based. As well, the fundamentals of synchronous machine construction and operation are also discussed. This is for the benefit of generator operators who have a mechanics background and are inclined to attain a modicum of proficiency in understanding the basic principles of operation of the generator. It also comes in handy for those professors who would like to adopt this book as a reference for a course on large rotating electric machinery. Chapters 2 and 3 (Generator Design and Construction and Generator Auxiliary Systems) contain a very detailed and informative description of all the components found in a typical generator and its associated auxiliary systems. Described therein are the functions that the components perform, as well as all relevant design and operational constrains. Some additional insight into design methods and calculations are also provided. Chapter 4 (Operation and Control) introduces the layperson to the many operational variables that describe a generator. Most generator–grid interaction issues and their effect on machine components and operation are covered in great detail. Chapter 5 (Monitoring and Diagnostics) and Chapter 6 (Generator Protection) serve to introduce all aspects related to the online and offline monitoring and protection of a large hydro generator. Although not intended to serve as a guideline for designing and setting up the protection systems of a generator, they provide a wealth of background information and pointers to additional literature. Chapters 7 (Inspection Practices and Methodology), leads off the second part of the book with a look at preparing for a hands-on inspection of large hydro generators. The chapter discusses the issues of concern for both safety of personnel and the equipment as well as the types of tools and approaches used in inspecting large hydro generators. This chapter also contains a collection of inspection forms that can be used for inspecting large hydro generators. These forms are very useful and can be readily adapted to any machine and plant. Chapter 8 (Stator Inspection), Chapter 9 (Rotor Inspection), and Chapter 10 (Auxilliaries Inspection) constitute the core of this book. They describe all components presented in Chapters 2 and 3, but within the context of their behavior under real operational constraints, modes of failure, and typical troubleshooting activities. These chapters provide detailed information on what to look for, and how to recognize problems in the machine during inspection. Chapters 8 and 9 also contain hundreds of pictures to assist in the inspection process in a methodical step-bystep crawl through of the machine. Chapter 11 (Maintenance and Testing) contains a comprehensive summary of the many techniques used to test the many components and systems comprising a generator. The purpose of the descriptions is not to serve as a guide to performing the tests as there are well established guides and standards for this purpose. Rather, they are intended to illustrate the palette of possible tests to choose from. Provided as well is a succinct explanation of the character of each test and explanations of how they are carried out.

xiv

PREFACE

Chapter 12 (Maintenance Philosophies, Upgrades and Uprates) is included to provide some perspective to the reader on the many choices and approaches that can be taken in generator and auxiliary systems maintenance, as well as upgrading equipment and uprating of the machine. Often, there are difficult decisions on how far to take maintenance. In some cases, only basic maintenance may be required, and on other occasions, it may be appropriate to carry out extensive rehabilitation of existing equipment or even replacement of components that can yield a higher efficiency or higher rating for the machine. This chapter discusses some of the issues that need to be considered when deciding on what, how much, and where to do it. We hope that this book will be not only useful to the operator in the power plant but also to the design engineer and the generator operations engineer. We have provided a wealth of information obtained in the field about the behavior of such machines, including typical problems and conditions of operation. The book should also be useful to the student of electrical rotating machines as a complementary reference to the books on machine theory. When read in its entirety, this book will assist the user in performing a complete machine inspection and understand with reasonable clarity, what they are observing, if there is a problem, and how to go about finding a solution to fix it. Although we have tried our best to cover each topic as comprehensively as possible, the book should not be seen as a guide to troubleshooting. In each case in which a real problem is approached, a whole number of very specific issues only relevant to that very unique machine come into play. These can never be anticipated or known and thus described in a book. Thus, we recommend the use of this book as a general reference source, but that the reader should always obtain adequate on-the-spot expertise when approaching a particular problem.

Glenn Mottershead Stefano Bomben Isidor Kerszenbaum Geoff Klempner Oakville, Ontario, Canada Mississauga, Ontario, Canada Toronto, Ontario, Canada Irvine, California

ABOUT THE AUTHORS Mr. Mottershead has worked or consulted on rotating apparatus for over 45 years with 33 of these years as an engineer at Westinghouse, where he was mentored by a select group of electrical and mechanical generator design and manufacturing engineers. These mentors had lineage that reached back to Nikola Tesla, George Westinghouse, and other key pioneers of the early power generation industry. His objective in writing this book with the other expert authors is to pass on lessons he was fortunate to receive to those working at all levels of hydro power generation. Mr. Mottershead is an IEEE Life Member and a Principle Consultant at HDR. Mr. Bomben is a Sr. Engineer at Ontario Power Generation (formerly Ontario Hydro) with over 29 years of experience inspecting large hydro generators, providing oversight on rewinds, overhauls, new machines, failure investigations, repairs and testing, and writing technical specifications. He is a senior member of the IEEE with many contributions to the development of generator operation, maintenance, and insulation standards. The inspiration for this book was to produce a comprehensive written knowledge base for use by any power engineer interested in large hydro generators, informed by theory, operational history and physical inspection. Mr. Klempner is an IEEE Fellow and large rotating electrical machines specialist in the power industry for over 43 years. He provides electrical machine technical services on a global basis, regarding large generators and motors. This includes inspection, testing, design evaluation, failure analysis, life assessment, preparation of technical specifications, and test procedures. Previously, he worked with Ontario Hydro (now Ontario Power Generation) for over 25 years, and then NSS (Nuclear Safety Solutions) and Kinectrics Inc. He has authored or coauthored 65 papers and articles, and 4 textbooks. He also has an extensive background of professional activities, with IEEE, EPRI, and CIGRE. Dr. Kerszenbaum has been involved in design, manufacturing, maintenance, and operation of large electrical machines for about 40 years. He also has been a contributor in writing IEEE standards for this type of equipment. He authored and coauthored several books on the subject, and educated hundreds of engineers over the years.

xv

ACKNOWLEDGMENTS The contents of this book are in part the result of personal experience accumulated over years of working with large hydro generators. It is also the result of the important long-term contribution of coworkers and associates. Each author was motivated by an important individual at an early stage of his career, and by many outstanding individuals in the profession over subsequent years. Two engineers, Frank Barnard and John F. Lyles, need to be recognized here as they had significant hydro generator mentoring roles for Mottershead and Bomben, respectively. Coauthors Geoff Klempner and Izzy Kerszenbaum are also important mentors as they pioneered the writing of the book Operation and Maintenance of Large Turbo Generators, which was the model for this book. The authors would like to give special recognition to Sungsoo Kim for writing Chapter 6 (Generator Protection); his patience and contribution has produced a magnificent compilation of his expertise. The authors would also like to thank Tim Maricic and Wayne Martin for their gracious contributions to Chapter 2. The authors are privileged to have had two very patient technical reviewers, John Linn and Richard Huber, who painstakingly went through the manuscript and contributed useful ideas. The authors are also very grateful to the individuals who kindly supplied the many pictures and information that make up this handbook. The authors wish to thank Ontario Power Generation for the incredibly large volume of pictures that form part of this book, without this support, this book would not have been possible. Unless otherwise indicated, all pictures in the book are courtesy of Ontario Power Generation. Special thanks to Victoria Bomben and Paolo Bomben for their assistance with the design of the front cover, and to Voith for the picture. The authors are most indebted to the IEEE Press for supporting its publication. The authors also would like to thank the members of the editorial departments of the IEEE Press and Wiley, the reviewers, and all others involved in the publication of this book for their support in making its publication possible. Finally, but certainly most intensely, the authors wish to thank their immediate families for their continuous support and encouragement while we played with big machines around the world.

xvii

CHAPTER

1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

The synchronous generator belongs to the family of electric rotating machines. Other members of the family are the direct-current (DC) motor or generator, the induction motor or generator, and a number of derivatives of these three. What is common to all the members of this family is the basic physical process involved in their operation, which is the conversion of electromagnetic energy to mechanical energy, and vice versa. Therefore, to gain an understanding of the physical principles governing the operation of electric rotating machines, one has to understand some rudiments of electrical and mechanical engineering. Chapter 1 is for those who are involved in operating, maintaining, and trouble-shooting electrical generators. Specifically, those who want to acquire a better understanding of the principles governing the machines’ design and operation, but lack an electrical engineering background. The chapter starts by introducing the rudiments of electricity and magnetism, quickly building up to a description of the basic laws of physics governing the operation of the synchronous electric machine, which is the type of machine to which all salient pole hydro generators belong.

1.1 INTRODUCTION TO BASIC NOTIONS ON ELECTRIC POWER 1.1.1

Magnetism and Electromagnetism

Certain materials found in nature exhibit a characteristic to attract or repel each other. These materials, called magnets, are also called ferromagnetic because they include the element iron as one of their constituent elements. Magnets always have two poles: one called north, the other called south. Two north poles will repel each

Handbook of Large Hydro Generators: Operation and Maintenance, First Edition. Glenn Mottershead, Stefano Bomben, Isidor Kerszenbaum, and Geoff Klempner. © 2021 The Institute of Electrical and Electronics Engineers, Inc. Published 2021 by John Wiley & Sons, Inc.

1

2

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

N

S

Lines of force

Figure 1.1-1 Representation of two magnetic poles of opposite polarity, with the magnetic field between them shown as “lines of force.”

other, as will two south poles. However, north and south poles will attract each other. A magnetic field is defined as a physical field established between two poles. Its intensity and direction determine the forces of attraction or repulsion existing between the two magnets. Figures 1.1-1 and 1.1-2 are typical representations of two interacting magnetic poles and the magnetic field established between them. Magnets are found in nature in all sorts of shapes and chemical constitution. Magnets used in industry are artificially made. Magnets that sustain their magnetism for long periods of time are denominated “permanent magnets.” The magnetic field produced by the north and the south pole of a permanent magnet is directional from north to south as shown in Figure 1.1-3. These are widely used in several types of electric rotating machines, including synchronous machines. However, due to mechanical as well as operational reasons, permanent magnets in synchronous machines are restricted to those with ratings much lower than large hydraulic (“hydro”) turbine-driven generators, which is the subject of this book. Hydro generators take advantage of the fact that magnetic fields can be created by the flow of electric currents in conductors, see Figure 1.1-4. The direction of the lines of force is given by the “law of the screwdriver”: mentally follow the movement of a screw as it is screwed in the same direction as

N

N

Lines of force

Figure 1.1-2 Representation of two north poles and the magnetic field between them. South poles will create similar field patterns, but the lines of force will point toward the poles.

1.1 INTRODUCTION TO BASIC NOTIONS ON ELECTRIC POWER

Figure 1.1-3 Representation of a “permanent magnet” showing the north and south poles and the magnetic field between them flowing from north to south outside the magnet.

N

3

S

Conductor

Electric current

Figure 1.1-4 Representation of a magnetic field created by the flow of current in a conductor.

Lines of force

that of the current; the lines of force will then follow the circular direction of the head of the screw. The magnetic lines of force are perpendicular to the direction of current. A very useful phenomenon is that forming the conductor into the shape of a coil can augment the intensity of the magnetic field created by the flow of current through the conductor. In this manner, as more turns are added to the coil, the same current produces larger and larger magnetic fields. For practical reasons, all magnetic fields created by current in a machine are generated in coils as shown in Figure 1.1-5. The use of coils to amplify the magnetic field intensity requires them to be constructed in a very specific manner so that the resulting flux is produced in an effective way. When the coil is operating in air, the magnetic field direction, shape, and intensity depends on the number of turns in the coil, the size of the coil, and the direction of electric current flow in the coil winding. The flux produced is basically divided into two types. One is the effective flux that links the entire coil and does the useful work, and the other is the leakage flux which is a more localized effect and does no useful work. In fact, the leakage flux creates additional losses that make the coil less efficient, electromagnetically speaking (see Figure 1.1-6). The

4

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

Lines of force

Current flow

Figure 1.1-5 Representation of a magnetic field produced by the flow of electric current in a coil-shaped conductor.

Leakage flux

Effective flux

Current out

Current in

Figure 1.1-6 Representation of a magnetic field produced by the flow of electric current in a coil-shaped conductor operating in air, showing the effective and leakage flux components of the magnetic field produced.

principles illustrated here become very important later on as we discuss the magnetic field in the generator and stray losses. To use the flux produced in a coil as effectively as possible, highly permeable ferromagnetic materials are used to capture and direct the flux so that the amount of leakage flux is minimized. This allows the coil to do more useful work and keeps losses to a minimum. Iron in various derivatives is by far the most widely used material because it has all the magnetic characteristics required. It is structurally suitable, and cost-effective. When an “iron” core is used within the coil, and current is flowing, the magnetic field produced is shaped effectively, and the iron core essentially becomes a north–south magnet in the process (see Figure 1.1-7). This is why stator cores and rotor poles of generators are made of steel, containing iron and a few small quantities of additional elements. The iron allows the principles discussed above to become a reality and is one of the reasons generators can be built to at least 97.5% efficiency.

1.1 INTRODUCTION TO BASIC NOTIONS ON ELECTRIC POWER

5

Effective flux

N

S

Current out

Current in

Figure 1.1-7 Representation of a magnetic field produced by the flow of electric current in a coil-shaped conductor with an “iron” core. The majority of the field produced is effective flux and the leakage field is reduced to a minimum.

1.1.2

Electricity

Electricity is the flow of positive or negative charges. Electricity can flow in electrically conducting elements (called conductors), or it can flow as clouds of ions in space or within gases. As will be shown in later chapters, both types of electrical conduction are found in hydro generators (see Figure 1.1-8).

(I)

Positive charge

Positive ionic cloud

Negative charge

Negative ionic cloud

(II) Electron Current = Flow of free electrons

Figure 1.1-8 Electricity. (I) Ionic clouds of positive and negative currents. The positive clouds are normally atoms that lost one or more electrons; the negative clouds are normally free electrons. This effect can be found inside the generator as partial discharge in the stator winding. (II) The flow of electrons inside a conductor material, for example, the copper windings of the rotor and stator.

6

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

Electrical

Mechanical

Battery Heat loss

Pump

V

I

Heat loss

H

Q

R ΔH

ΔV V = Voltage I = Current R = Resistance ΔV = Voltage drop ΔV = I × R Power = I × V

H = Pressure head Q = Flow rate R = Resistance ΔH = Pressure drop ΔH = Q × R Power = Q × H

Energy storage

Electrostatic storage

Spring

Q (charge)

K 1

E=2

CV 2

V

Δx

Magnetic storage

E = 2 K(Δx)2 1

I

E = ½ m V2 1

E=2

LI2

(inertia)

Figure 1.2-1 Electrical– mechanical equivalence.

1.2 ELECTRICAL–MECHANICAL EQUIVALENCE There is an interesting equivalence between the various parameters describing electrical and mechanical forms of energy. People with either electrical or mechanical backgrounds find this equivalence useful to the understanding of the physical process in either form of energy. Figure 1.2-1 describes the various forms of electrical–mechanical equivalence.

1.3 ALTERNATING CURRENT (AC) Synchronous generators operate with both alternating-current (AC) and directcurrent (DC) electric power. The DC can be considered a particular case of the general AC, with frequency equal to zero. The frequency of an alternating circuit is a measure of the number of times the currents and/or voltages change direction (polarity) in a unit of time. The hertz (Hz) is the universally accepted unit of frequency, and measures cycles per second. One Hz equals one cycle per second. Alternating currents and voltages encountered in the world of industrial electric power are for all practical purposes of constant frequency. This is important because periodic systems, namely systems that have constant frequency and sinusoidal signals, allow the currents and voltages to be represented by phasors.

1.3 ALTERNATING CURRENT (AC)

7

Voltage (e) E (phasor)

ω

α

α

Figure 1.3-1 A phasor E that can represent the voltage impressed on a circuit.

A phasor is a rotating vector. The benefit of using phasors in electrical engineering analysis is that it greatly simplifies the calculations required to solve circuit problems. Figure 1.3-1 depicts a phasor of magnitude E, and its corresponding sinusoidal trace representing the instantaneous value of the voltage quantity e. The magnitude E represents the maximum value of voltage (e). The phasor is made of a vector with magnitude proportional to the magnitude of E, rotating at a constant rotational speed ω. The convention is that phasors rotate counterclockwise. The vertical projection of the phasor results in a sinusoid representing the instantaneous voltage (e) existing at any time. In Figure 1.3-1, α = ω × t, where t is the time elapsed from its zero crossing. When a sinusoidal voltage is applied to a closed circuit, a current will flow in it. After a while, the current will have a sinusoidal shape (this is called the steadystate current component) and the same frequency as the voltage. An interesting phenomenon in periodic circuits is that the resulting angle between the applied voltage and the current depends on certain characteristics of the circuit. These characteristics combine into one representative parameter, impedance and are broken down into resistive, capacitive, and inductive. The angle between the voltage and the current in the circuit is called the power factor angle and is defined as φ. The cosine of the same angle is called the power factor of the circuit or, for short, the PF. In the case of a circuit having only resistances, the voltages and currents are in phase, meaning that the angle between them equals zero. Figure 1.3-2 shows the various parameters encountered in a resistive circuit. This is a representation of a sinusoidal voltage of magnitude “E” applied on a circuit with a resistive load “R.” The schematics show the resultant current (i) in phase with the voltage (v). It also shows the phasor representation of the voltage and current. It is important to note that resistances have the property of generating heat when a current flows through them. The heat generated equals the square of the current times the value of the resistance. When the current is measured in amperes and the resistance

8

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

Loss = I2R

I

E

R

Δυ

Resistive circuit Phasor (vector) diagram υ

E ω

i

α

α Power (p) = υ • i (active power) In a resistive circuit, the voltage and the current wave forms are in phase, i.e. the power factor of the circuit equals 1 ω

E I

I and E in phase

X

Figure 1.3-2 Alternating circuits (resistive).

in ohms, the resulting power dissipated as heat is given in watts. In electrical machines, this heat represents a loss of energy. One of the fundamental requirements in designing an electric machine is the efficient removal of the energy resulting from these resistive losses, with the purpose of limiting the temperature rise of the internal components of the machine. In resistive circuits, the instantaneous power delivered by the source to the load equals the product of the instantaneous values of the voltage and the current. When the same sinusoidal voltage is applied across the terminals of a circuit with capacitive or inductive characteristics, the steady-state current will exhibit an angular (or time) displacement in relation to the driving voltage. The magnitude of the angle (or power factor) depends on how capacitive or inductive the load is. In a purely capacitive circuit, the current will lead the voltage by 90 , whereas in a purely inductive one, the current will lag the voltage by 90 (see Figure 1.3-3). Here, the sinusoidal voltage E is applied to a circuit comprised of resistive, capacitive, and inductive elements. The resulting angle between the current and the voltage depends on the value of the resistance, capacitance, and inductance of the load. A circuit that has capacitive or inductive characteristics is referred to as being a reactive circuit. In such a circuit, the following parameters are defined:

1.3 ALTERNATING CURRENT (AC)

R

C

Resistance

L

Capacitance

Inductance

υC

υR

9

υL I

E

Reactive circuit

υ

i

τ

φ S = υ•i (VA) is apparent power p = υ•i•cos φ (W) is active power q = υ•i•sin φ (volt-amperes-reactive [VAR's] is reactive power E

I

E

φ φ I Inductive

Capacitive

Figure 1.3-3 Alternating circuits (resistive-inductive-capacitive).

S: The apparent power P: The active power Q: The reactive power

S = E × I, given in units of volt-amperes or VA. P = E × I × cos φ, where φ is the angle between the voltage and the current. P is given in units of watts. Q = E × I × sin φ, given in units of volt-amperesreactive or VAR.

The active power P of a circuit indicates a real energy flow. This is power that may be dissipated on a resistance as heat, or may be transformed into mechanical energy. However, the use of the word “power” in the definition of S and Q has been an unfortunate choice that has resulted in confounding most individuals without an electrical engineering background for many years. The fact is that apparent power and reactive power do not represent any measure of real energy. They do represent the reactive characteristic of a given load or circuit, and the resulting angle (power factor) between the current and voltage. This angle between voltage and current significantly affects the operation of an electric machine. For the time being, let us define another element of AC circuit analysis: the power triangle. From the relationships shown above among S, P, Q, E, I, and φ, it can be readily shown that S, P, and Q form a triangle. By convention, Q is shown as positive (above the horizontal), when the circuit is inductive, and vice versa when capacitive (see Figure 1.3-4).

10

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

υ

E I

i

τ

φ

Q (inductive) S φ P

Q (capacitive)

Q - generates losses in the system, and voltage drop on lines and cables

Figure 1.3-4 Definition of the “power triangle” in a reactive circuit.

To demonstrate the use of the power triangle within the context of large generators and their interaction with the power system, we need to consider a one-line schematic that includes the generator, transmission system, and the connected load at the end (see Figure 1.3-5). The voltage required at the load, so that it will operate correctly, is given as 1000 V. The transmission line resistance and reactance are provided and the line impedance calculated as shown, using the power triangle approach. If we now consider an actual load for the simple system of Figure 1.3-5, we can calculate the current drawn by the load and the voltage required from the generator source to compensate for all the line losses and voltage drop across the line. Two cases are provided to illustrate the effect of a purely resistive load versus a load with

Line losses = I2Rline

Load

G VS

Vload

Simple power system showing a generator, bus, line, and load Vload = 1000 V Line resistance (R) = 10 Ω (resistive) Line reactance (X) = 10 Ω (inductive) Line impedance (Z) = √(102 + 102) = 14.14 Ω

Figure 1.3-5 Schematic of a simple system in oneline form.

1.3 ALTERNATING CURRENT (AC)

11

Case 1– Load = 100 kW (Unity power factor {i.e. cos φ =1} ) P = √3* I * V * cos φ I = 100 000 W/1000 V/1.73/1 = 57.8 A Losses in the line (I2 Rline) = 57.82 A * 10 Ω = 33.4 kW Voltage drop along the line = I*Z = (I * R)2 + (I * X)2 = 817 V The required delivery voltage at the source (Vs) is: Vs = (Vload + IR)2 + (IX)2

Figure 1.3-6 Case 1. The load is purely resistive in this example, and the system is operating at the “unity” power factor.

Vs = (1000 + 578)2 + (578)2 = 1 680 V Vs Vload = 1000 V

I

I*Z

I*X

I*R

Case 2– Load = 100 kW and 50 kVAR-inductive (lagging power factor) S = 3* I * V = 100 0002 + 50 0002 = 111 803 VA I = 111803 VA /1000 V/1.73 = 64.55 A Losses in the line (I2 Rline) = 64.552 A * 10 Ω = 41.6 kW Voltage drop along the line = I*Z = (I * R)2 + (I * X)2 = 913 V The required delivery voltage at the source (Vs) is: Vs = (Vload + IR(sin φ + cos φ)]2 + [IX(cos φ – sin φ)]2 Vs = (1 000 + 645.5 (sin φ + cos φ)]2 + [645.5(cos φ – sin φ)]2 = 1 892 V Vs φ

I

Vload = 1000 V

I*Z I*R

I*X

Figure 1.3-7 Case 2. The load is resistive and inductive in this example, and the system is operating in the “lagging” power factor range.

a reactive component include (see Figures 1.3-6 and 1.3-7). Working out the required voltage from the generator for the two different loads by the power triangle method shows how reactive loads greatly affect the power system operation and the generation requirements. Reactive power compensation is a large part of synchronous generator operation and affects generator design in a significant way, as will be discussed later on in Chapter 2. There is a delicate balance between generation and load that is clearly shown by the two cases presented and the comparison of operational results (see Table 1.3-1). Although the “real” power consumed is the same, the addition of the reactive component in Case 2 has caused an increase in current drawn from the generator, an increase in line losses, a higher volt drop across the line, and, therefore, a higher voltage required from the generator source.

12

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

TABLE 1.3-1 A comparison of Case 1 and Case 2

Load

100 kW

100 kW and 50 kVAR

Power consumed by the load (kW) Current (A) Line losses (kW) Voltage drop along line (V) Required delivery voltage at generating end (V)

100 57.8 33.4 817 1680

100 64.6 41.6 913 1892

The above examples show that there is a considerable demand placed on the generator to operate the various loads on a system. In reality, the generator terminal voltage Vs is constant, plus or minus 5% by design. As the load increases or decreases, the current from the generator changes significantly and the voltage drop on the system Vload requires compensation (Figure 1.3-8). Therefore, the second major function of the generator, after production of “real” power, is to produce “reactive” power to help control the voltage on the grid, which will also be discussed later in Chapter 4.

Vload φ

I

Vs = 1000 V

I*Z

I*X

I*R

Lagging power factor

Vload

I

I*Z

I*X

I*R

Vs = 1000 V Unity power factor

Vload

I*X I*Z

I

Vs = 1000 V Leading power factor

I*R

Figure 1.3-8 The effect on the voltage drop as the circuit goes from lagging through unity to leading power factor operation.

1.4 THREE-PHASE CIRCUITS

13

1.4 THREE-PHASE CIRCUITS The two-wire AC circuits discussed above (called single-phase circuits or systems), are commonly used in residential, commercial, low voltage, and low power industrial applications. However, all electric power systems to which industrial generators are connected are three-phase systems. Therefore, any discussion in this book about the “power system” will refer to a three-phase system. Moreover, in industrial applications, the voltage supplies are, for all practical reasons, balanced, meaning that all three-phase voltages are equal in magnitude and apart by 120 electrical degrees. In those rare events in which the voltages are unbalanced, the implications for the operation of the generator will be discussed in other chapters of this book. Three-phase electric systems may have a fourth wire, called “neutral.” The “neutral” wire of a three-phase system will conduct electricity if the source and/or the load are unbalanced. In three-phase systems, two sets of voltages and currents can be identified. These are the phase and line voltages and currents. Figure 1.4-1 shows the main elements of a three-phase circuit. Three-phase circuits can have their sources and/or loads connected in wye (also known as “star”) or in delta forms. (See Figure 1.4-2 for a wye-connected source feeding a delta-connected load.)

IA Load A

EA Neutral EC

EB

Load B

IB

Load C

IC

EC ECA IC

φC

φB

IB

Figure 1.4-1 Three-phase systems. Schematic depiction of a three-phase circuit and the vector (phasor) diagram representing the currents, voltages, and angles between them.

ω

EBC

IA

φA

EAB EB EA, EB, EC – Phase voltages ECA, EBC, EAB, – Line voltages

EA

14

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

Y-Connected source Δ-Connected load

Figure 1.4-2 A “wye-connected” source feeding a “delta-connected” load.

Almost without exception, hydro generators have their windings connected in wye (star) form. Therefore, in this book, the source (or generator) will be shown wye-connected. There are a number of important reasons why hydro generators are wye-connected. They have to do with considerations about its effective protection as well as design (insulation, grounding, etc.). These will be discussed in the chapters covering stator construction and operations. On the other hand, loads can be found connected in wye, delta, or a combination of the two. This book is not about circuit solutions; therefore, the type of load connection will not be brought up herein.

1.5 BASIC PRINCIPLES OF MACHINE OPERATION In Section 1.1, basic principles were presented showing how a current flowing in a conductor produces a magnetic field. In this section, three important laws of electromagnetism will be presented. These laws, together with the law of energy conservation, constitute the basic theoretical bricks on which the operation of an electrical machine is based.

1.5.1

Faraday’s Law of Electromagnetic Induction

This basic law of Electromagnetic Induction, derived by the genius of the great English chemist and physicist Michael Faraday (1791–1867), presents itself in two different forms: 1. A moving conductor cutting the lines of force (flux) of a constant magnetic field has a voltage induced in it. 2. A changing magnetic flux inside a loop made from a conductor material will induce a voltage in the loop. In both instances, the rate of change of the magnetic field is the critical determinant of the resulting voltage potential. Figure 1.5-1 illustrates both cases of

1.5 BASIC PRINCIPLES OF MACHINE OPERATION

1. Changing flux

15

If φ changes in time:

e e

–

dφ dt

φ v

e

Conductor 2. Moving conductor B (into palm)

v

Generator

N ℓ

S

Rule of the “right hand” • Thumb always the driving • Flux into palm

B

Length of wire in the field

• For generator: Thumb → Direction of movement Fingers → Voltage induced • For motor: Thumb → Current Fingers → Force direction

e =B×ℓ×v Volts Tesla Meters m/s

Figure 1.5-1 Both forms of Faraday’s basic law of electromagnetic induction. A simple rule (the “right-hand” rule) is used to determine the direction of the induced voltage in a conductor moving across a magnetic field at a given velocity.

electromagnetic induction and also provides the basic relationship between the changing flux and the voltage induced in the loop. The first case shows the relationship between the induced voltage in a wire moving across a constant field. The second case shows one of the simple rules that can be used to determine the direction of the induced voltage in the moving conductor.

1.5.2

Ampere–Biot–Savart’s Law

This basic law is attributed to the French physicists Andre Marie Ampere (1775–1836), Jean Baptiste Biot (1774–1862), and Victor Savart (1803–1862). In its simplest form, this law can be seen as the “reverse” of Faraday’s law. Whereas Faraday’s law predicts a voltage induced in a conductor moving across a magnetic field, the Ampere–Biot–Savart law establishes that a force is generated on a current-carrying conductor located in a magnetic field.

16

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

Length of wire in the field

F =B ×ℓ ×I Newtons Tesla Meters Amperes I B (into plam)

Motoring F I

ℓ N

S F B

Figure 1.5-2 The Ampere–Biot–Savart law of electromagnetic-induced forces as it applies to electric rotating machines. Basic numerical relationships and a simple rule are used to determine the direction of the induced force.

Figure 1.5-2 presents the basic elements of Ampere–Biot–Savart’s law as applicable to electric machines. The figure also shows the existing numerical relationships, and a simple hand rule to determine the direction of the resultant force.

1.5.3

Lenz’s Law of Action and Reaction

Both Faraday’s law and Ampere–Biot–Savart’s law neatly come together in Lenz’s law, written in 1835 by the Estonian-born physicist Heinrich Lenz (1804–1865). Lenz’s law states that electromagnetic-induced currents and forces will try to cancel the originating cause. For example, if a conductor is forced to move, cutting lines of magnetic force, a voltage is induced in it (Faraday’s law). Now, if the conductors’ ends are closed together so that a current can flow, this induced current will produce (according to Ampere–Biot–Savart’s law) a force acting upon the conductor. What Lenz’s law states is that this force will act to oppose the movement of the conductor in its original direction. Here, in a nutshell, is the explanation for the generating and motoring modes of operation of an electric rotating machine. This law explains why, when the load in a generator is increased (i.e. more current flows in its windings, cutting the magnetic field in the gap between rotor and stator), more force is required from the

1.5 BASIC PRINCIPLES OF MACHINE OPERATION

V

Example

e i

i

17

e R

B

F

R

1. The upward moving conductor in a magnetic field induces a voltage (Faraday) 2. Closing the circuit generates a current 3. The current creates a force opposing the movement (Ampere and Lenz) Hint: Use the rule of the palm to show the direction of “F ” This phenomenon explains the torque applied by the generator on the turbine, when the unit is loaded Induced currents and forces will try to cancel the originating cause

Figure 1.5-3 Lenz’s law as it applies to electric rotating machines. Basic numerical relationships and a simple rule are used to determine the direction of the induced forces and currents.

turbine to counteract the increase in induced larger forces and keep supplying the larger load. Similarly, Lenz’s law explains the increase in the supply current of a motor as its load increases. Figure 1.5-3 neatly captures the main elements of Lenz’s law as it applies to electric rotating machines.

1.5.4

Electromechanical Energy Conversion

The fourth and final physical law that captures, together with the previous three, all the physical processes occurring inside an electric machine, is the “principle of energy conversion.” Within the domain of the electromechanical world of an electric rotating machine, this principle states that All the electrical and mechanical energy flowing into the machine, less all the electrical and mechanical energy flowing out of the machine and stored in the machine, equals the energy dissipated from the machine as heat. It is important to recognize that although mechanical and electrical energy can go into or out of the machine, the heat generated within the machine always has a negative sign, that is, heat generated in the machine is always released during the conversion process. A plus sign indicates energy going in; a minus indicates

18

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

“Principle of energy conversion” in electromechanical systems: WM

WE Input/Output of electric energy

+

WS

Input/Output of mechanical energy

+

WH +

Change in stored energy

=0 Heat dissipated

Electrical = 1

I2L +

2

1

V2C

2

Mechanical = rotational energy WH

WM WE WS

WH is always negative (i.e. heat is always released during the conversion process) WE, WM , and WS can have “+” or “–” signs WE and WM with a “+” means input to the machine “–” means output from the machine WS with a “+” means increase of stored energy “–” means decrease of stored energy

Figure 1.5-4 Principle of energy conversion as applicable to electric rotating machines.

energy going out. In the case of the stored energy (electrical and mechanical), a plus sign indicates an increase of stored energy, whereas a minus sign indicates a reduction in stored energy. The balance between the various forms of energy in the machine will determine its efficiency and cooling requirements, as well as its critical performance and construction parameters. Figure 1.5-4 depicts the principle of energy conversion as applicable to electric rotating machines.

1.6 THE SYNCHRONOUS MACHINE The rudiments of electromagnetism have been presented along with the four basic laws of physics describing the physical processes existing in any electrical machine. Therefore, it is the right time to introduce the basic synchronous machine, which, is the type of electric machine used for all large hydro turbine driven generators.

1.6 THE SYNCHRONOUS MACHINE

1.6.1

19

Background

The birth of commercial alternating current (AC) hydro generation dates back to June 1891 with the delivery of AC power in Colorado USA from the Ames Hydro Power Station to the Gold King Mine 4.2 km away. The generator for the power plant and the motor for the mill were identical Westinghouse synchronous singlephase machines rated 73.5 kW, 3,000 volts, 133 Hz. Later in 1891 an Oerlikon synchronous 3-phase hydro generator at 180 kW, 55 volts, 40 Hz, with transformer extended power transmission 160 km from Lauffen to Frankfurt Germany during an international electrical exhibition in Frankfurt, see Figure 1.6-1. These pioneering concepts happened more than 125 years before the writing of this book and the basic principles of the hydro electrical power system are the same today. These very early power generation experiments were instrumental in the adoption by New York’s Niagara Falls Power Company to use this technology at their Niagara Falls Adams Hydro Station. This pioneering power plant started a rapid and continuing increase of unit ratings at these falls by operating their first of ten Westinghouse 3,677 kW, 2,000 volt, 25 Hz hydro generators in August of 1895. For all practical purposes, the great DC (Edison) versus AC (Westinghouse) duel was over. It is interesting to note that although tremendous development in generator structural, magnetic and insulating materials, and manufacturing methods has occurred over the years, the basic design elements of these electric machines have remained practically unchanged. The very earliest concept was that a synchronous generator is used to drive a synchronous motor but Tesla’s induction motor quickly replaced the synchronous motor for the vast majority of electric

Figure 1.6-1 The hydroelectric generator from Lauffen, now in the Deutches Museum, Munich. Source: Reprinted with permission from Neidhofer [3].

20

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

Rating (MVA*) 1200 1100

Water-cooled

Air-cooled * Mega Volt-ampere

Wu Dong De (China)

1000 Three Gorges (China)

900 Itaipu (Brazil/Paraguay)

800

Guri II (Venezuela) Xi Luo Du (China)

700 600 500

Bath county (USA) Helms (USA) Samrangjin (South Korea)

300

EI Chocon (Argentina) Fumas (Brazil)

200 100

Paulo Afonso (Brazil)

Raccoon Mountain (USA)

400

Necaxa (Mexico)

Wehr (Germany) Rottau/Malta (Austria)

Herdecke (Germany)

Vianden (Luxembourg) Suiho (China)

0 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020

Year

Figure 1.6-2 “Growth” graph, depicting the overall increase in size over the last century, of hydro generators. Source: Courtesy of Voith.

motor applications. However the synchronous generator remained the universal machine of choice for the large-scale generation of hydroelectric power. The world today is divided almost exclusively between utility systems generating their power at either 50 Hz or 60 Hz and these synchronous generators have continuously grown in size over the years. Today, it is now possible to see these hydro machines with terminal voltages over 20,000 volts and ratings over 900 MVA as shown in Figure 1.6-2.

1.6.2

Principles of Construction

Synchronous machines come in all sizes and shapes, from the miniature permanent magnet synchronous motor in wall clocks, to the largest hydro generators of up to about 944.5 MVA. Synchronous machines are one of two types: stationary field or rotating DC magnetic field. The stationary field synchronous machine has salient poles mounted on the stator, the stationary member. The poles are magnetized either by permanent magnets or by a DC current. The armature, normally containing a three-phase winding, is mounted on the shaft. The armature winding is fed through three sliprings (collectors) and a set of brushes sliding on them. This arrangement can be found in machines up to about 5 kVA in rating. For larger machines covered in this book the typical arrangement used is the rotating magnetic field. The rotating magnetic field (also known as “revolving field”) synchronous machine has the field winding wound on the rotating member (the rotor) and the armature wound on the stationary member (the stator). A DC current, creating a magnetic field that must be rotated at synchronous speed, energizes the rotating

1.6 THE SYNCHRONOUS MACHINE

21

field winding. The rotating field winding can be energized through a set of sliprings and brushes (external excitation) or from a diode bridge mounted on the rotor. The rectifier bridge is fed from a shaft-mounted synchronous generator, which is itself excited by the pilot exciter. In externally supplied fields, the source can be a shaft-driven DC generator, a separately excited DC generator, or a solid-state rectifier. Several variations or combinations of these variations are used. The stator core is made of insulated silicon-steel laminations. The thickness of the laminations and the type of steel are chosen to minimize eddy current and hysteresis losses, while maintaining required effective core length and minimizing costs. The core is mounted directly onto the frame. The core is slotted (the slots are normally open), and the coils making up the winding are placed in the slots. The most popular arrangements are lap and wave windings of various types. Modern large machines typically are wound with double-layer lap windings. A rotor field of salient pole construction is as shown in Figure 1.6-3. In a large generator, the rotor magnetic field is generated by a coil wrapped around it with current passing through the coil. For simplicity, this figure shows a

Stator (armature)

A 120° Rotor (field)

Salient pole machine

S

N

RPM B

C

Stator winding

Voltage waveforms for phases A, B, C e 120° 120° 120° c 0

b

a

180

t # of poles f(HZ) =

p × RPM 120

Figure 1.6-3 Synchronous machine construction.

22

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

two-pole rotor. Salient pole rotors normally have many more than two poles. When designed as a generator, large salient pole machines are driven by water turbines. The bottom part of this figure shows the three-phase voltages obtained at the terminals of the generator, and the equation relates the speed of the machine, the number of poles, and the frequency of the resulting voltage. This figure includes all synchronous hydro generators, almost every synchronous condenser, and the overwhelming majority of synchronous motors. Large salient pole rotors are typically made of laminated poles retaining the winding under the pole head. The poles are keyed or bolted onto the shaft (spiderand-wheel structure). The majority of salient pole machines have an additional winding on the rotating member. This winding, made mostly of copper bars short-circuited at both ends, is embedded in the head of the pole, close to the face of the pole at its airgap. In synchronous generators this winding serves to dampen the oscillations of the rotor around synchronous speed and is, therefore, named the damping or damper winding (also known as amortisseur). In synchronous motor applications, this winding can be used to start the motor or condenser as an induction motor, and take it to almost synchronous speed, when the rotor is “pulled in” by the synchronous torque.

1.6.3

Rotor Windings

In hydro generators, the winding producing the magnetic field is made of a number of coils connected in a series circuit, energized with DC power typically supplied through the shaft from slip or collector rings on the shaft. In self-excited generators, the shaft-mounted exciter and rectifier (diodes/commutators) generate the required field current. The shaft mounted exciter is itself excited from a stationary winding. Alternately, static excitation eliminates the need for shaft driven exciters. This method is commonly applied to modern generators or as an upgrade for older machines where maintenance and performance are issues. Generators typically have field supplies of 125 or 250 V DC, and in some machines even higher depending on the design. A more elaborate discussion of rotor winding design and construction can be found in Chapter 2.

1.6.4

Stator Windings

The magnitude of the voltage induced in the stator winding is function of the magnetic field intensity, the rotating speed of the rotor, and the number of turns in the stator winding. An actual description of individual coil design and construction, as well as how the completed winding is distributed around the stator, is meticulously described in Chapter 2. In this section, a very elementary description of the winding arrangement is presented to facilitate the understanding of the basic operation of the machine. As stated above, coils are distributed in the stator in a number of forms, each with its own advantages and disadvantages. The basic goal is to obtain three balanced and sinusoidal voltages having very little harmonic content (harmonic

1.7 SYNCHRONOUS MACHINE: BASIC OPERATION

1

N

2

S

3

N

4

S

1

23

N

Stator slots

Figure 1.6-4 “Developed” view showing four-poles, slots, and a section of the winding.

voltages and currents are detrimental to the machine and other equipment in a number of ways). To achieve a desired voltage and MVA rating, the designer may vary the number of slots and the manner in which individual coils are connected, producing different winding patterns. The most common winding arrangement is the lap winding, shown in Figure 1.6-4, for salient pole machines. The section shown is a portion of one of the three phases. It can be readily seen that the winding runs clockwise under a north pole, and counterclockwise under a south pole. This pattern repeats itself until the winding covers the number of pole groups in a parallel. A similar pattern is followed by the other two phases, but located at 120 electrical degrees apart. A connection scheme that allows great freedom of choice in designing the windings to accommodate a given terminal voltage is one that allows connecting sections of the winding in parallel, series, and/or a combination of the two as shown in Figure 1.6-5.

1.7 SYNCHRONOUS MACHINE: BASIC OPERATION For a more in depth discussion of the operation and control of hydro generators, the reader is referred to Chapter 4. In this chapter, the most elementary principles of operation of synchronous machines will be presented. As mentioned above, hydro generators are almost always three-phase machines. Thus, the best place to start describing the operation of a three-phase synchronous machine is a description of its magnetic field. Earlier, we described how a current flowing through a conductor produces a magnetic field.

24

CHAPTER 1

Six circuits in series Phase

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

Six circuits in parallel Phase

Two circuits in parallel Three circuits in parallel and three in series and two in series Phase Phase

Neutral Representation of 1 circuit = group of turns

Neutral Neutral

Neutral

Figure 1.6-5 Typical winding configurations. Source: Courtesy of Voith.

It was also shown that by coiling the conductor, a larger field is obtained for the same current magnitude. Recall, that if the three phases of the winding are distributed at 120 electrical degrees apart, three balanced voltages are generated, creating a three-phase system. Now, a new element can be brought into the picture. By a simple mathematical analysis, it can be shown that three balanced currents (of equal magnitudes and 120 electrical degrees apart) flow in a balanced three-phase winding when a magnetic field of constant magnitude is produced in the airgap of the machine. This magnetic field revolves around the machine at a frequency equal to the frequency of the currents flowing through the winding (see Figure 1.7-1). As shown, a constant magnitude and constant rotational speed magnetic flux is created when three-phase balanced currents flow through a three-phase symmetrical winding. The sketch is for a four pole winding, however, similar result applies for any number of pairs of poles. The importance of a three-phase system creating a constant field cannot be stressed enough. The constant magnitude flux allows power, megawatts, to be transformed inside an electric machine from electrical to mechanical power, and vice versa. It is important to remember that a constant-magnitude flux produces a constant-magnitude torque.

1.7.1

Magnetic Representation

To describe the fundamental principles describing the operation of a synchronous machine, it is convenient to use the constructs of an ideal salient pole rotor machine connected to an infinite bus. The infinite bus represents a bus which can deliver or absorb active and reactive power without any limitations and whose voltage and frequency are essentially constant. The ideal machine has zero resistance and leakage reactance, infinite permeability, and no saturation, as well as zero reluctance torque.

25

1.7 SYNCHRONOUS MACHINE: BASIC OPERATION

τp

Φ

B 1

B

1 Stator yoke 2 Stator teeth 3 Stator winding in slots 4 Salient pole rotor 5 Excitation winding 6 Magnetic field lines

τp = pole pitch = D π/ 2p B = induction in air gap B1 = amplitude of fundamental wave D = stator bore diameter 2p = number of poles

Three-phase voltage at generator terminals 1.5

Voltage (p.u)

1 0.5 0 0

–0.5

0.01

0.02

0.03

0.04

Phase A Phase B Phase C

–1 –1.5 Time (seconds)

• Three-phase system: voltage or current waves 120° • Phase A voltage: U = Umax·cos(w.t) • Phase B voltage: U = Umax·cos(w.t – 120°) • Phase C voltage: U = Umax·cos(w.t + 120°)

Figure 1.7-1 Production of stator rotating field. Source: Courtesy of Voith.

The production of torque in the synchronous machine results from the natural tendency of two magnetic fields to align themselves. The magnetic field produced by the stationary armature is denoted as Φs. The magnetic field produced by the rotating field is Φf. The resultant magnetic field is Φr = Φf + Φs The flux, Φr, is established in the airgap of the machine. (Bold symbols indicate vector quantities.) When the torque applied to the shaft equals zero, the magnetic fields of the rotor and the stator become perfectly aligned. The instant torque is introduced to the shaft, either in a generating or in a motoring mode, a small angle is created between the stator and rotor fields. This angle (λ) is called the torque angle of the machine, the angle (ψ) is called the internal power factor angle, and (β) is

26

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

d-Axis

Φf Φr β

Φaq λ

Ef

Iq

ψ

q-Axis

Φad Id

Φs Ia

Figure 1.7-2 Phasor diagram of an unsaturated salient pole generator (lagging pf ).

the space angle between the fundamentals of the Φf and Φr waves (see Figure 1.7-2) [1]. Due to saliency, the reactance measured at the terminals of the generator is a function of rotor positon and thus the two reactance theory can be applied [1]. The armature current Ia can be resolved into two components, namely, Id and Iq representing direct and quadrature axis currents. Id is in time quadrature with the internal excitation voltage Ef, where Iq is in time phase with Ef. The direct axis component of the armature Id produces an armature reaction flux Φad along the axis of the field poles. Iq however produces an armature reaction flux Φaq in space quadrature with the field poles. The magnetic effect of Φad is centered on the axis of the field pole where the magnetic effect of Φaq is centered on the inter polar axis. The armature reaction flux Φs is the space phasor sum of the components Φad and Φaq. It follows that the resultant flux Φr is a result of the space phasor sum of the main field flux Φf and armature reaction flux Φs.

1.7.2

Generator Mode: Steady State Using Vectors

This section describes various practical use diagrams of a salient pole machine that is operating under steady-state conditions [2]. Most generators in operation are in the mode of supplying real power and either supplying or absorbing reactive power. A machine is said to have a lagging power factor when supplying reactive power and a machine is said to have a

1.7 SYNCHRONOUS MACHINE: BASIC OPERATION

27

y Eo

G IqXq H

IdXd O

φ

δ I

Et

x

A

Id Iq Two-axis voltage vector diagram Salient pole generator Power factor – lagging

Figure 1.7-3 Vector diagram of generator with a lagging power factor.

leading power factor when absorbing reactive power. It is also common to have the machine supply just real power and the machine is said to be at unity power factor. Figure 1.7-3 shows a diagram of a generator suppling power to the system with a lagging power factor and Figure 1.7-4 shows a diagram of the same generator with a leading power factor. When looking at these diagrams a few assumptions need to be kept in mind. First, the machine is in steady-state operation, second, the machine is connected to an “infinite bus,” and lastly, magnetic saturation is neglected. The following explanation does not discuss how to develop these curves in detail and is left to the reader by referencing [2]. Referring to Figure 1.7-3, the line OA represents the terminal voltage Et, OG represents the internal generated voltage Eo, and I is the armature current lagging behind the terminal voltage by the angle φ which is the power factor angle. The armature current I can be represented by the direct and quadrature currents Id y

G Eo IqXq

O Id

H IdXd

Iδ φ Iq

Et

A

x

Two-axis voltag vector diagram Salient pole generator power factor– leading

Figure 1.7-4 Vector diagram of generator with a leading power factor.

28

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

and Iq, respectively. The vectors IdXd (parallel to Eo) and represented by AH and IqXq (perpendicular to Eo) and represented by HG are also mutually perpendicular. The angle (δ) is the torque or power angle. The diagram demonstrates the internal generated voltage is higher than the terminal voltage during lagging power factor operation. This makes sense as the generator must push out the reactive power and requires the higher internal potential voltage of Eo in order to accomplish this. Referring to Figure 1.7-4, there are some key differences worth noting while operating at a leading power factor. The armature current I is now leading the terminal voltage Et by the angle φ. The quadrature axis components Iq and IqXq are now larger in magnitude than when the machine was lagging. The takeaway from this diagram is that the terminal voltage is higher than the internal generated voltage. This also makes sense as the generator is now absorbing reactive power so the terminal voltage must be higher to push the power back into the machine. It is important to recognize that as soon as the internal generated voltage is higher or lower than the terminal voltage, even by the smallest amount, reactive power will be supplied or absorbed by the machine accordingly.

1.7.3

System Support: Reactive Power

For all intents and purposes, during reactive power support operation, there is no water flowing through the turbine, thus no MW are being supplied to the system. However, by adjusting the excitation higher (over excitation), the machine can deliver reactive power and by decreasing the excitation (under excited), the machine can absorb reactive power. This mode of operation for a conventional generator (that normally supplies MW as well) is only during times where there may be an abnormal system condition and the generator will need to provide this reactive support or there is system emergency and extreme reactive support is required. On one end of extreme operation, there is almost no current supplied to the DC field winding of the rotor and large amounts of reactive power (MVARS) are being absorbed by the generator. The capability curve, which is discussed later in the book, will have limits set as to how many MVARS the generator may absorb on a continuous basis without damaging effects. This mode of operation is not far from operating with a loss of excitation. Protections are put in place to prevent migration into this potentially damaging area of operation. Depending on the design of the generator and excitation system, rotating or static, the minimum excitation value can approach 10 A or less before the excitation system bottoms out and/or protections operate. On the other end of extreme operation, the generator will need to supply large amounts of reactive power (MVARS) to the system, again, for abnormal system events or emergencies. Again, the capability curve from the manufacturer will outline what the maximum capability is and over excitation limits can be put in place to protect the generator from damaging effects from rotor overheating.

1.7 SYNCHRONOUS MACHINE: BASIC OPERATION

29

+P

Motor

Resultant power P Power due to field excitation

–π/2

–δ

π/2

–π

π

+δ

Power due to saliency

Generator

–P

Figure 1.7-5 Steady-state power angle characteristic of a salient pole synchronous machine (with negligible armature resistance).

It is now appropriate to expand on the effects of “saliency” on generator operation. Power developed in a salient pole machine can be written as Equation (1.1) [1]: P = E t E o X d sin δ +

1 2 1 1 E sin 2δ − 2 t Xq Xd

(1.1)

The term on the left-hand side of the “+” sign is the power developed due to the field excitation. The term on the right-hand side of the “+” sign is the power developed due to saliency. A graphic representation of the resultant power is shown in Figure 1.7-5 [1]. Even if no current is supplied to the DC field winding, there is a torque generated known as reluctance torque, due to the saliency effect. It is this torque which keeps the rotor in synchronism with the system, and prevents the machine from slipping poles.

1.7.4

Motor Operation

The electric machine can be specifically designed to operate as a pump motor so it performs useful physical work. In this particular arrangement, the motor pumps water through the penstock for irrigation or as storage when power is abundant to be used later to generate when power is in short supply. Synchronous pump motors are designed to operate at a fixed speed without any slip between the armature frequency and the fields rotating frequency. To bring the motor to synchronous speed, it must be accelerated by a method consistent with its design. A separate

30

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

induction starting motor dedicated to accelerating the synchronous rotor is a common starting method. Other methods incorporating variable frequency sources or across-the-line inductive starting by the rotor damper windings are also used. In the situation, where across-the-line starting is used, the damper windings will get very hot quickly and the machine must be synchronized within the recommended time by the manufacturer. Failed synchronization attempts must allow for cooling of the damper circuit before an attempt to start the machine is made again. This cooling off period will also be specified by the manufacturer. Failure to follow these set procedures can result in overheated and even melted damper assemblies which in turn will affect starting performance and may require repair depending on the severity. If a known excursion from these procedures has occurred, it is highly recommended that a visual inspection of the damper assembly by a knowledgeable person be performed.

1.7.5

Equivalent Circuit

When dealing with three-phase balanced circuits, electrical engineers use the one-line or single-line representation. This simplification is allowed because in three-phase balanced circuits, all currents and voltages, as well as circuit elements are symmetrical. Thus, by “showing” only one phase, it is possible to represent the three-phase system, as long as care is taken in using the proper factors. For instance, the three-phase balanced system of Figure 1.4-1 or Figure 1.4-2 can be represented as shown in Figure 1.7-5. Hereinafter, when describing a threephase generator by an electrical diagram, the one-line method will be used. The most convenient way to determine the performance characteristics of synchronous machines is by means of equivalent circuits. These equivalent circuits can become very elaborate when saturation, armature reaction, harmonic reactance, and other nonlinear effects are introduced. However, the simplified circuit in Figure 1.7-6 is conducive to obtaining the basic performance characteristics of the machine under steady-state conditions. In Figure 1.7-7, the reactance Xa, represents the magnetizing or demagnetizing effect of the stator windings on the rotor. It is also called the magnetizing reactance or armature reaction reactance. Ra represents the effective resistance of the stator. The reactance Xl represents the stator leakage reactance. The sum of Xa and Xl is used to represent the total reactance of the machine and is called the synchronous reactance (Xs). Zs is the synchronous impedance of the machine. The internal generated voltage is represented by Eo, the airgap voltage by Er, and finally the terminal voltage of the machine by Et. It is important to remember that the

Line Load Generator

Figure 1.7-6 One-line representation of the circuit shown in Figures 1.4-1 and 1.4-2.

1.7 SYNCHRONOUS MACHINE: BASIC OPERATION

31

Xs Xa

XI

Ra

Ia

Zs

Eo

Er

Figure 1.7-7 Steadystate equivalent circuit of a synchronous machine.

Et

Machine terminals

equivalent circuit described in Figure 1.7-7 represents the machine only under steady-state conditions. The simple equivalent circuits of Figure 1.7-8 suffices to determine the steady-state performance parameters of the synchronous machine connected to a power grid. These parameters include voltages, currents, power factor, and power angle. The regulation of the machine can be easily found from the equivalent circuit for different load conditions by using the regulation formula: Percent regulation = 100 V no load − V load V load For a detailed review of the performance characteristics of the synchronous machine, in particular the hydro generator, the reader is referred to Chapter 4. Note: Regulation in a generator indicates how the terminal voltage of the machine varies with changes in load. When the generator is connected to an infinite bus (i.e. a bus that does not allow the terminal voltage to change), a change in load will affect the machine’s output in a number of ways. (See Chapter 4 for a discussion of this topic.)

1.7.6

Machine Losses [4]

Current flows through the machines conductors generating heat (a loss). However, there are a number of other sources within a working synchronous generator that produce heat and, thus, losses. The following is a list of some of those sources of losses: • I2R losses of stator winding • I2R losses of rotor winding • Core loss • Stray load loss • Excitation system loss

32

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

(a) Z Eo

I Et

Eo

Eo IZ

δ ϕ

ϕ

IZ I

Et

Et

I Lagging power factor (overexcited)

Leading power factor (underexcited)

(b)

Z I Eo

Et

I ϕ

Et

Et

δ

ϕ

δ

IZ

IZ

Eo

Eo I

Leading power factor (overexcited) AQ5

Lagging power factor (underexcited)

Figure 1.7-8 Steady-state equivalent circuit and vector diagram. (a) Generator operation and (b) motor operation.

• Windage and Friction loss • Ventilation and cooling loss In determining I2R losses, the resistance of windings should be corrected to the reference temperature. During the commissioning tests for a generator, the OEM may use a common reference temperature of 75 C. When calculating the actual losses at a specific temperature, then the resistance at 75 C must be converted to the resistance at the desired temperature.

1.7 SYNCHRONOUS MACHINE: BASIC OPERATION

33

1.7.6.1 Stator Winding I2R Loss The stator winding I2R loss is the sum of the I2R losses in all of the stator winding current paths. The I2R loss in each current path shall be the product of its resistance in ohms as measured with direct current and the square of its current in amperes. 1.7.6.2 Rotor I2R Loss The rotor winding I2R loss shall be the product of the measured resistance in ohms of the rotor winding and the square of rotor current in amperes. 1.7.6.3 Core Loss The core loss shall be taken as the difference in power required to drive the machine at normal speed when separately excited to produce a voltage at the terminals corresponding to the rated voltage at open circuit, and the power required to drive the unexcited machine at the same speed. 1.7.6.4 Stray Load Loss The stray load loss is determined by subtracting the stator winding I2R loss at a specific value of stator current from the short circuit loss at the same value of stator current. The short circuit loss shall be taken as the difference in power required to drive the machine at normal speed, when separately excited to circulate current in the stator winding with its terminals shorted, and the power required to drive the unexcited machine at the same speed. The stator winding I2R loss shall be calculated for the temperature of the winding during the short circuit test. 1.7.6.5 Excitation System Losses These losses are the total of electrical and mechanical losses in the equipment supplying excitation. They shall include the exciter, voltage regulator, and associated devices comprising the excitation system of the synchronous machine. Where common equipment is provided in the excitation system for two or more machines, the common equipment loss shall not be included in the evaluation of the synchronous machine efficiency. Motor loss shall be included if a unit motor-generator exciter set is used; if a unit rectifier is used, the rectifier and rectifier transformer losses shall be included. Include collector electrical brush contact voltage drop loss if applicable. 1.7.6.6 Friction and Windage Loss The friction and windage loss, including brush mechanical friction, is the power required to drive the unexcited machine at rated speed with the brushes in contact, deducting that portion of the loss which results from a. forcing the air through any part of the ventilating system that is external to the machine and cooler (if used); b. the driving of direct-connected flywheels or other direct-connected apparatus. c. In the case of machines furnished with a complete set of bearings, only that portion of the friction and windage loss produced by the bearing load due to

34

CHAPTER 1

PRINCIPLES OF OPERATION OF SYNCHRONOUS MACHINES

the generator or generator/motor itself shall be included. In the case of machines not furnished with a complete set of bearings, only that portion of the friction and windage loss associated with the equipment supplied with the generator or generator/motor shall be included. Where losses are apportioned between equipment or between manufacturers, the method of allocation of the losses shall be subject to agreement between the manufacturer and purchaser. 1.7.6.7 Ventilating and Cooling Loss This loss includes any power required to circulate the cooling medium through the machine and cooler (if used) by fans or pumps that are driven by external means (such as a separate motor), so that their power requirements are not included in the friction and windage loss. The power consumption of a separate blower system used for a specific unit and necessary for continuous operation of the unit shall be included. In the following chapters, these losses and their origin, control, and consequences to the machine’s design and operation will be covered in detail.

1.8 REFERENCES 1. Sarma, M. S. (1986). Electric Machines: Steady-State Theory and Dynamic Performance, St. Paul, MN, West Publishing Company. 2. Walker, J. H. (1952). Operating characteristics of salient-pole machines. Proceedings of the IEE – Part II: Power Engineering 100(73), 13–24. 3. Neidhofer, G. (1992). The evolution of the synchronous machine. Engineering Science and Education Journal 1(5), 239. Asea Brown Boveri. 4. IEEE (2006). IEEE Std C50.12-2005: IEEE Standard for Salient-Pole 50 Hz and 60 Hz Synchronous Generators and Generator/Motors for Hydraulic Turbine Applications Rated 5 MVA and Above. New York, IEEE.

1.9 FURTHER READING Say, M. G. (1978). Alternating Current Machines, Pitman Publishing.

CHAPTER

2

GENERATOR DESIGN AND CONSTRUCTION

The focus of this chapter is on the design and construction of the generator and its major individual components. Although not a design book, this chapter will go into enough detail on how the components are designed and fabricated, to assist the reader in maintaining them. In addition, issues that significantly influence the design of the various generator components are discussed. If the reader wishes to learn more in depth about a specific component, it is recommended that the original equipment manufacturer (OEM) be consulted. The class of generators under consideration is water-driven generators commonly called hydro or waterwheel generators. They range from relatively small machines of a few megawatts (MW) to very large generators with ratings up to 944.5 MW. These generators typically have speed ratings of 72–900 RPM and are installed all over the world. The basic function of the generator is to convert mechanical power, delivered from the shaft of the turbine, into electrical power. There are many different types of turbines such as Francis, Kaplan, Pelton, and Deriaz which will impact the generator design. The discussion of each type of turbine is out of the scope of this book, and the reader is referred to Ref. [1] for a comprehensive discussion on turbines. The mechanical energy from the turbine is converted by means of a rotating magnetic field produced by direct current in the copper winding of the rotor or field, into three-phase alternating currents and voltages in the copper winding of the stator (armature). The stator winding is connected to terminals, which are in turn connected to the power system for delivery of the output power to the system. Generators are made up of two basic members, the stator and the rotor, but the stator and rotor are themselves constructed from numerous parts. Rotors are the rotating member of the two, and they undergo severe dynamic mechanical loads as well as the magnetic and thermal loads. The stator is stationary, as the term

Handbook of Large Hydro Generators: Operation and Maintenance, First Edition. Glenn Mottershead, Stefano Bomben, Isidor Kerszenbaum, and Geoff Klempner. © 2021 The Institute of Electrical and Electronics Engineers, Inc. Published 2021 by John Wiley & Sons, Inc.

35

36

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

suggests, but it also experiences significant dynamic forces such as vibration, torsional, and radial loads, as well as the magnetic, thermal, and electrical loading. From the previous discussion, it becomes obvious that there are many issues to consider in generator design and each of these influences the performance of the overall machine. Design issues of high-voltage insulation, electrical currents (AC and DC), magnetic flux, heat production and cooling, mechanical forces, and vibrations all must be accounted for and made to work together for proper operation of a large generator. As stated in Chapter 1, some hydro generators also perform as large synchronous motors to allow the turbine to pump water from the tailrace back to the reservoir during hours of low consumption of electric power. The same water is later used to drive the machine as a generator during hours of high electric power consumption. These so-called “pump-hydro” or “pumped-storage” machines have specific components, such as “pony motors” or squirrel cage windings, that allow them to start and pick up speed until they synchronize to the grid.

2.1 STATOR CORE 2.1.1

Laminations

The stator core in large AC machines are constructed from thin, sheets of electrical grade silicon steel typically 0.35 mm (0.014 ) to 0.50 mm (0.019 ) thick. These thin sheets are called laminations (or punchings or core plate). The laminations are cut from the silicon steel sheet using a very large die cutter or a laser cutter. The die cutter is commonly referred to as punch. There are slots punched or laser-cut into the laminations to accommodate the stator winding, and core attachment keybars and/or the core compression bolts, see Figures 2.1-4 and 2.1-10. It is very important that the manufactured lamination is within the tolerances as outlined in the manufacturing drawing. Dimensions such as the outer extremities, keybar fitments, slot dimension, and core bolt cut-out just to name few are critical for proper core stacking and long-term performance. The manufacturing facility for the laminations is an excellent place to visit for an inspection and dimensional verification during manufacturing of the core. Each lamination is insulated on one or both sides (normally both) to prevent circulating currents between sheets. Numerous types of organic and in-organic insulation material can be applied. Segmented laminations are laid next to each other to form a completed 360 ring, then the next layer is laid on top, circumferentially offset so that the joints do not coincide. If the machine is stacked (piled) at site, then the core will be a continuous stack or pile all the way around the machine, thus, there is no “split” in the core. If a machine is piled in the factory, it may be piled in circumferential sections and then shipped to site for assembly. The frame and stator core sections are joined at site to form the circular stator of the machine. The joints are commonly called splits, and the number of splits is related to transportation size limits. It is typically cheaper for any generator to have the core piled in the factory, however the core may have a structural weakness at each split that could reduce the service life of the core or stator winding.

2.1 STATOR CORE

37

The core is built up from these thin laminations to limit eddy current losses in the core. Some manufacturers bond 10 or 20 laminations together at one time and install as a group to increase core stiffness. Each lamination is insulated from other laminations; laminations when installed may be grounded at the back of the core via the keybar, the insulation ensures that circulating currents between laminations as a result of the induced magnetic flux cannot occur. If the insulation between the laminations is damaged, and the laminations are grounded at the back of the core, localized current flow and overheating of the laminations would result. Figure 2.1-1 shows an exaggerated yoke for a hydro machine for the purposes of illustrating the circulating current concept. Normally, the yoke on a hydro machine versus the slot depth would be more in line with Figure 2.1-4. The stator core is designed to carry the electromagnetic fluxes and must be capable of handling the magnetic flux density in the stator teeth and core back areas which can on average be between 1.7 and 1.45 T, respectively, depending on the design and manufacturer. These flux levels can vary from these stated values, but they are in the ballpark of what can be expected. Figure 2.1-2 demonstrates, in a simplified manner, how breaches in the interlaminar insulation result in larger than normal eddy current and losses. Larger than normal means currents and losses substantially higher than those found when the machine operates within its designed parameters. Figure 2.1-3 carries the analysis in Figure 2.1-2 further. It can be seen therein that with 100 or so divisions, the eddy current losses are reduced to about 2% of the original losses. Further divisions make this type of loss very small when compared to that of an equivalent single

Insulation damage

Conductor bars

Groundwall insulation

Flux

Keybars

Currents induced through insulation damage

Figure 2.1-1 Cross section of core showing inter-laminar damage and the eddy current flowing as a result. Source: Courtesy of Qualitrol-Iris Power.

38

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

L

L

I

L

L/4

Steel lamination material

Due to skin-effects, most current flow in the periphery of the conducting material. Therefore, the following applies: -

Induced volts ∝ area Loop resistance ∝ perimeter Current ∝ area / perimeter Losses = I2R

For the full block: -

Induced volts ∝ L2 Loop resistance ∝ 4L Current ∝ L/4 Losses ∝ L3/4

For the same block divided into four insulated “laminations”: -

Induced volts in one section ∝ L2/4 Loop resistance of one section ∝ 2.5L Current in one section ∝ L/10 Losses in one section ∝ L3/40 Total loss for all four sections ∝ L3/10

This means that by dividing the original metal sheet into four insulated laminations, the total eddy-current loss was reduced 2.5 times

Figure 2.1-2 Inter-laminar insulation reduces eddy current losses in the steel.

body core. From the content of this paragraph, it becomes obvious how by damaging the insulation between the laminations in a few spots, the eddy current losses can significantly increase, with consequential temperature rises and additional damage to the insulation, further increasing the amount of short circuited core. This has the potential to becoming a runaway situation, leading to melted core material and a catastrophic failure. The issue of short-circuited laminations and consequential core damage due to increased eddy current loss is broached in other places in this book when discussing foreign metallic objects left inadvertently on the core

2.1 STATOR CORE

39

Loss in % of unlaminated block

120

100

80

60

40

20

0

1

10

100

1000

Number of laminations

Figure 2.1-3 Laminated-core eddy-current loss as percentage of full-block loss. This graph shows the cumulative effect of decreasing eddy current losses by laminating the core steel.

laminations in the bore, or metal parts (e.g. bolts and washers) becoming loose during operation and landing on the core in such a way that laminations are shorted.

2.1.2

Lamination: Slot and Yoke Section

Referencing Figure 2.1-4, the stator core lamination is slotted on its inner diameter forming stator teeth to accommodate the stator winding that are form-wound with multi-turn diamond coils, or single-turn bars. In both of these winding types, the

Stator yolk Stator slots

Wedge grooves

Stator teeth

Inner diameter of core

Figure 2.1-4 Stator core segment described.

40

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

sections that fit into the slots are rectangular in shaped and consequently the core slots for them are also rectangular in shape. The slots for this type of winding have grooves (one on each side of the slot) near the inner diameter of the lamination to retain the nonmetallic wedges that hold the winding coils or bars tightly in the stator core slots. Magnetic wedges are not common but are usually used with small airgaps. There are also machines out there, although rare that have a semi-closed slot instead of an open rectangular slot. These machines have “hair pin or U-shaped” windings pressed in one end of the machine and then joined at the other. Early days of asphalt insulation had these types of windings, but modern insulation is also used in the present day when replicating this type of winding. Also, this has been done with larger machines where the coil cross section is more than one square inch and connections are then installed at either end, see Figure 2.1-5. The portion of the core behind the bottom of the winding slots is called the yoke or core-back area and has lower flux density than the tooth area. For example (these are average values), if the teeth are loaded at 1.5 T, then the core yolk area would be loaded at 1.2 T for a typical hydro machine. The size of the yolk is due to several factors such as the amount of heat dissipation required, reactance of the machine, core stiffness, and whether it is a high or low flux machine.

Figure 2.1-5 Showing the U shaped or “hairpin” coil. Source: Courtesy of Motion Electric & Delom Services, members of Groupe Delom.

2.1 STATOR CORE

2.1.3

41

Core Piling (Stacking) and Clamping

A stator core can contain 100 000s of laminations stacked onto keybar assemblies (this includes double dovetail keybars) that are placed circumferentially inside the frame, see Figure 2.1-6 for a conventional keybar arrangement and Figure 2.1-8 for the double dovetail arrangement which can help prevent buckling of the core. The stator keybar placement during construction of the stator frame assembly is very critical for proper core piling and compression. If during assembly, the tolerances

Core bolt

Keybar Gusset

Piling pin

Vent lamination

Figure 2.1-6 Lamination stacking in the stator, also showing the conventional keybars (rectangular) welded to gussets at the back of the core as well as the round core clamping studs. Piling pins are used in the slots during construction to align laminations. Vent laminations have a red colored protective coating applied from the factory, Figure 2.1-7.

Radial duct Radial duct curled to direct air flow into core from the air gap

Figure 2.1-7 Heavy lamination segment with I-shaped blocks (radial duct) – red protective coating applied.

42

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Section A-B

A

B

Stator frame 5.5

46.5

1.5

48

36

Clamping strap

0.5

Double dovetail bar Stator core 43

Figure 2.1-8 Shows double dovetail design that prevents core buckling. Source: Courtesy of Voith.

(which are in the thousandths of an inch) with respect to the keybar placement are not maintained, the core will not pile correctly and will not compress correctly. Thus, ensure that the OEM is following their internal or industry-accepted standard for keybar circularity, concentricity, axial trueness, keybar flatness, and keybar stepping and core compressing (normally every 304–457 mm (12–18 ) of core height). In fact, when placing laminations onto the keybars, the laminations will not fit smoothly and may have to be forced into position if some of the previously mentioned tolerances are not in check. One method of checking the keybar placement is to take a few laminations, stack them together on a piece of plywood, screw the yolk to the plywood so that the keybar interface portion is hanging out in the air, and run the laminations up and down the keybar assembly all the way around the machine. This should be able to be done with ease without any shaking or jiggling of the lamination stack. It will be evident very quickly whether or not the keybars are in the right position or not. When doing a core restack, it is highly advisable that the keybar arrangement be verified against the original construction tolerances. If the keybars are out of tolerance, resetting of the keybars is paramount. The core can be stacked in sections onto the frame that are shipped to site (if convenient for shipping) or continuously piled onto the frame which is done at site. In either case, the core must be stacked as level as possible from front to back and as well as circumferentially (minimal wave at the top of the core when measuring core height). As a result of the lamination punching, the tooth assembly at the front of the core will tend to pile higher than the back of the core due to distortion of the

2.1 STATOR CORE

43

Lamination shim

Tangential clearance Core to frame clearance (some designs)

Radial clearance

Figure 2.1-9 Core-to-keybar mounting arrangement in the stator frame. Shim located at outer core diameter to level the stacking.

lamination when punched. Shims are added to the back of the core to level the stack from front to back (see Figure 2.1-9). The distortion problem is largely eliminated with laser-cutting of the laminations; however, it is not economical to laser cut large amounts. Figure 2.1-6 shows the stator core bolts attached to the frame assembly and not going through the core laminations and separate keybars for the core attachment. Another variation of the core attachment and bolting arrangement is shown in Figures 2.1-10–2.1-12. In this design, the core bolts go through the core itself in order to provide the clamping pressure, and the keybars are again separate for core attachment. The clamping is achieved by using a threaded core bolt at both ends with an insulating sleeve. A bolt with a washer arrangement as shown in Figure 2.1-10 can be used to secure the clamping plate. The insulating sleeve prevents the core bolt from contacting the core causing circulating currents as shown in Figure 2.1-10. The sleeve then passes through an insulated bushing that is housed inside the vent lamination structure as shown in Figure 2.1-11. In this figure, for demonstration purposes, the bushing is placed inside the lamination hole where the core bolt goes through. The core bolts may have a powder coating applied as an additional barrier of protection. This bushing will prevent the core bolt and sleeve from rattling inside the core during service. Again, for demonstration purposes, the bushing is removed from Figure 2.1-11, the red colored ventilation layer is put in place as shown in Figure 2.1-12 and then bushings would be placed in every hole in this layer. This layer accommodates this insulating bushing along with the I-shaped blocks used for airflow. The I-shaped blocks will be

44

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Keybar welded to frame

Clamping plate

Corebolt insulating sleeve

Figure 2.1-10

Shows another arrangement for the core bolt.

Insulating bushing used for the core through bolts

Figure 2.1-11 For demonstration purposes this figure shows an insulating bushing for core through bolt. The bushing does not sit on this lamination but on top of the red vent laminations as shown in Figure 2.1-12.

discussed shortly. Some OEMs do not paint the ventilation layer with insulating varnish, thus this layer will look like the rest of the core in terms of color. OEMs have different nut and washer arrangements that can include a compression washer assembly as shown in Figure 2.1-13.

2.1 STATOR CORE

45

Insulating bushings will be here I-shaped blocks

Figure 2.1-12 Showing the vent lamination with the hole to accommodate the bushing in Figure 2.1-11.

Figure 2.1-13 Shows compression washer assembly. Source: Courtesy of Voith.

Yet another variation of core clamping and attachment is shown in Figure 2.1-14, where the keybar and core clamping bolt are one piece with threaded ends. One additional design feature employed on some hydro generators in the core-ends to reduce the higher losses in the stator teeth due to fringing flux is to

46

CHAPTER 2

Figure 2.1-14

GENERATOR DESIGN AND CONSTRUCTION

Shows keybar and core clamping bolt as one piece with threaded end.

split the teeth into smaller sections in the radial direction (also called “slitting”), or use step-back punchings, see Figures 2.1-15 and 2.1-16 respectively. These special laminations are punched or laser cut at the factory along with the rest of the regular laminations. This reduces the eddy current effect in the teeth and hence the losses and core-end heating effect. In order to cool the core, nonmagnetic radial ventilation ducts, “space blocks or I-shaped spacers” are spot welded and secured to a thicker core lamination (typically called the “heavy” and is typically 0.711 mm (0.028 ) thick installed at strategic and set locations in the core stack to form a radial air path from the stator bore

Clamping finger not welded to clamping plate

Slit in tooth

Core bolt nut and washer assembly

Clamping finger welded to heavy lamination

Figure 2.1-15 76 MVA newly piled stator core showing the clamping finger assembly at the top of the core, I-spacer vent assembly, first two packets with a slit tooth to reduce eddy currents due to fringing fluxes.

2.1 STATOR CORE

Clamping plate

47

Clamping finger is welded to clamping plate along underside of plate

Clamping finger

Figure 2.1-16 86 MVA newly piled stator core showing the first packet with step back punchings to reduce eddy currents and piling pin shown at far left – fingers not in final position on the core yet.

Piling pin

Step back punchings instead of slit in tooth

to the back of core area, and vice versa, see Figure 2.1-7 for the heavy lamination itself and Figure 2.1-6 for the locations along the stack height. Once the stator core is properly stacked, it must be held together tightly under pressure in the axial direction to ensure long-term performance and a fretting free environment. In order to achieve this performance, tightness of the core should be checked periodically and may need retightening after decades of operation. If the core is properly stacked and clamped, and no buckling or waves develop, issues of widespread or local area core looseness should not develop over time. For larger hydro machines, the clamping mechanism typically consists of heavy steel plate (length varies by OEM and design) that has heavy finger assemblies welded to it. This assembly is then placed on top of the core clamping bolts and torque into their final position. Another method is to have the heavy fingers as part of the core assembly (welded to a heavy lamination just like the I-shaped assemblies are) and then use the heavy plate on top to secure the entire assembly. Each OEM has their own proprietary method of using these finger plate assemblies, some are more effective than others over time. See Figures 2.1-15 and 2.1-16 for a couple of variations of this assembly. The assembly in Figure 2.1-15 has the fingers welded to the heavy lamination in the core assembly and the plate sitting on top separately, whereas Figure 2.1-16 has the plate and fingers welded together as one piece separate from the laminations. There are various combinations on how fingers and plates fit together at the top and bottom of the stator core. The general concept in most designs is that the frame is used as a fulcrum and the fingers have lever action onto the stator core. In a typical design, the outer edge of the clamping plate

48

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Adjustable fulcrum points

Figure 2.1-17 Shows finger plate installation with adjustable fulcrum points.

sits on the frame and the bolts when tightened exert typically more axial force on the teeth than they do on the back iron. Again, providing more force on the ends of the teeth than the back of the core (see Figure 2.1-17). For smaller hydro machines, the piling and clamping mechanisms may be the same as for a large hydro or would be done as described in the following. In cores with single-piece laminations, the core can either (i) be built into the stator frame and clamped by tooth support fingers and steel rings (rings that are one piece equal to the circumference of the stator core at each end and then welded to the core support bars), or (ii) built as a separate assembly and then fitted into a stator frame with support bars that have been machined with a profile and dimensions that provide a tight fit between the two assemblies. In either case vent ducts are installed during the core building process and the core has to be placed under a high axial pressure before the end support structures, consisting of end fingers and clamping rings, are fixed in place. Segmented cores use either (i) through bolts that are installed through holes punched in the core laminations, or (ii) keybars or dove tails at the core back to which the laminations are assembled to. In either case, an even axial force is applied over the surface of the laminations by the use of ring flanges or plates. The core design, and thereby the tightness, must be able to accommodate the steady load machine torque as well as the transient torques experienced during fault conditions. Such torques are transmitted through the laminations to the stator frame, via the keybars mounted on the stator frame. The gussets that are welded onto the keybar must be capable of taking these tangential loads without failing, particularly during phase-to-phase faults and faulty synchronization (see Figure 2.1-6). Ensure the welds on these gussets are not cracked, if they are, they must be repaired. Overtightening of the stator core can result in damage to the stator frame, laminations, and ventilation duct “I-shaped blocks or beams” (“I,” “Square,” or “U” shaped assemblies), resulting in a weakening or even cracking of the beams and thin steel laminations, which in time will result in slackening of the core. Adversely, not tightening the core sufficiently will result in lamination vibration, producing a low-pitch hum and resulting in fretting of the lamination insulation followed by potential burning, cracking, and breaking of the lamination steel and also a high potential for damage to the stator bar insulation installed in these

2.1 STATOR CORE

49

slots. Also, a loose stator core will not be able to withstand the additional forces during fault conditions and may result in premature failure of the stator. Furthermore, a low clamping force will decrease the core capacity to resist the buckling phenomenon. Core pressures once the core bolts have been tightened on modern new cores are typically in the range of 150–200 PSI. Once a core buckle or wave develops, it is nearly impossible to remove it. Buckling can occur for a variety of reasons such as • overheating of the core • frame not accommodating thermal expansion of the core • tolerances used up when the core was assembled due to stator frame manufacturing • lamination being out of tolerance when manufactured The core is designed to thermally expand by a certain amount radially, circumferentially, and axially when at rated load with rated ventilation and cooling medium flowing. There are tolerances between the keybar and laminations to accommodate thermal expansion of the laminations with respect to the stator keybar and frame assembly. When the machine heats up, the stator frame expands radially and circumferentially as does the core. If a stator core is overheated, the allowed clearance between the laminations and the keybars may be exhausted and the core will then push up against the stator frame assembly, radially and tangentially, and begin to buckle. Refer to Figure 2.1-9 and pay particular attention to the clearance or interface gap (tangential and radial) between the laminations and the keybar and the lamination and the stator frame. If the stator frame is not able to expand freely on the soleplates for whatever reason, this may cause the core in that section to expand hard up against the stator frame and core buckling may result. There are many stator designs that have no initial clearance between the lamination and the stator frame. The core in these designs will thermally expand in operation more than the frame and is normally able to safely elastically expand the frame to accommodate the expansion difference. When these generators operate at excessive temperatures, have lost core clamping compression, or in some designs the frames are prevented from radially expanding, the cores can buckle from excessive compressive stresses. Cooling of the core is accomplished in large generators with the use of air to water heat exchangers and modern-day designers use Computational Fluid Dynamics software to optimize airflow and analyze temperature rises for different components of the machine. Radial “ducts” which are formed by the vent laminations allow airflow in the core for this purpose. The losses generated in the core are dissipated to the cooling air at the surface of the radial ducts. The width of the ducts and the thickness of the core packages are chosen as required by the ventilation needs of the machine and the temperature permitted in the core. The ventilation scheme of the generator, that is, the amount of airflow and where the air is flowing should never be adjusted without consulting the manufacturer first. In particular, the rotor fan arrangements should be left as designed unless the manufacturer has sanctioned a change.

50

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

2.2 STATOR FRAME The basic purpose of the stator frame is to provide support for the stator core. It also is segregated internally to create a ventilation circuit within the generator. The stator frame includes an outer shell, commonly called a wrapper plate (see Figure 2.2-1), to which circumferential shelves and the keybars are attached (see Figure 2.1-6). Figure 2.1-6 is the inside of what is shown in Figure 2.2-1. Also, the surface air coolers are supported by the outer wrapper via cutouts. On the bottom of the stator frame, there is a welded structure of steel footings to secure the generator to the foundation otherwise known as the soleplate assembly, see Figure 2.2-2 for a typical arrangement. Soleplate assemblies are very diverse from

Stator frame

Surface air cooler cutout

Figure 2.2-1 Shows outside part of stator frame showing the wrapper plate and frame rings for the two shelves. The surface air cooler opening cutout is covered by cardboard for protection during construction.

Wrapper plate

Stator hold down bolt

Stator frame

Threaded bolt and nut not in final torque position yet

Leveling key

Stator soleplate

“J or T” hook

Key to prevent tangential rotation May provide radial expansion

Leveling bolts

Figure 2.2-2 Typical soleplate positioned in the foundation awaiting final setup and encasement in grout.

2.2 STATOR FRAME

51

one manufacturer to another and some assemblies are very complex. The soleplates are required to carry the weight of the generator, rotational torque and transient tangential and axial loads due to system disturbances, and if required, provide means of radial expansion. Referencing Figure 2.2-2 as an example, this particular soleplate assembly consists of four major components as outlined beneath.

2.2.1

The Steel Box

The steel box is of welded construction and houses the soleplate and “J” hook and is set into a pocket in the concrete. This box is also adjustable for level by using steel bolts threaded into each corner to allow for infinite adjustment before final grouting is done. The steel boxes, which vary from as few as 4 for a smaller machine, to upwards of 12–16 for a larger machine, carry the entire weight of the stator frame, plus all upper bracket weight and operational loads.

2.2.2

The Soleplate

The soleplate is a level and precision machined surface which allows the stator frame to slide radially during thermal expansion guided by rectangular or cylindrical keys, yet keep the stator tangentially locked so it cannot rotate see Figure 2.2-2 for a typical arrangement. Some machine designs only use the soleplate for a surface for the stator frame to sit on, there is no free thermal expansion in the radial direction allowed as the stator frame is locked into position using nut and bolt assemblies also known as stator hold-down bolts. These bolts are torqued down to ensure no stator frame movement is possible in the axial and radial direction. Further, the hole machined in the stator frame is very close in diameter to the hold-down bolt size thus limiting any thermal expansion. In other designs, the hold-down bolts to stator frame clearances are such that radial movement is possible. Further, the hold-down bolts themselves will have a finite clearance between the bottom of the bolt head and the stator frame itself, thus allowing for thermal radial expansion. One way of achieving this is by placing a small machined cylinder inside the hole of the stator frame (where the bolt passes through) so that the cylinder is slightly proud of the stator frame when in the final installed position. The bolt is then torqued down onto this cylinder preventing axial movement but allowing for radial expansion. The radial expansion is possible because the hold-down bolt and machined cylinder have clearance between them so the frame can expand. The hold-down bolt head keeps the stator frame in position during faults since the hold-down bolt head diameter is larger than the machined cylinder diameter. This latter arrangement is easy to identify during an inspection as small feeler gauge will pass under the hold-down bolt head to stator frame interface. This soleplate assembly is a very critical component of the generator that is often forgotten once the machine has been placed in service. Finally, some machine designs have the stator frame sitting on the keys and not the soleplate. It is very important that the soleplate surface be kept in pristine condition throughout the life of the machine. This opportunity (during a major overhaul

52

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

for example) is only possible if the stator frame can be lifted in the air exposing the soleplate surface. The surface should be checked for corrosion and uniformity of contact between the mating surfaces. The radial keys or dowel fit can also be checked and all surfaces can be cleaned and lubricated before the stator is lowered.

2.2.3

The “J or T” Hook

The “J” hook is a thick steel rod formed in the shape of the letter “J.” The top end (the top of the “J”), where the soleplate is fastened to, has a threaded stud and a nut. The other end of the “J” is encased 457 mm (18 ) or more into the solid concrete foundation (the hook part of the letter “J” grabs the concrete). The rest of the “J” passes through the steel box as shown in Figure 2.2-2. Once the hook is set into the concrete and the pocket where the steel box resides is encased in grout, and the nut torqued to design specifications, the soleplate is solidly connected to the foundation and is not going to move. There are variations to this “J” hook installation in that the “J” may be an upside down “T” and there may be more than one holding down the soleplate. For example, instead of one “J” or upside down “T” in the middle there may be one on each side of the box instead with a larger soleplate face to accommodate as shown in Figure 2.2-3.

2.2.4

The Grout

After the steel box has been set at the correct height, diameter from pit center, and leveled, a wooden mold is placed in front of the grout pocket. The grout is then

“J or T” hook

Leveling bolts

Figure 2.2-3 Shows another style of soleplate with double “J or T” hooks.

2.2 STATOR FRAME

53

poured to the correct height with respect to the soleplate elevation. The grout serves as a permanent encasement of the soleplate assembly. The frame structure must also be capable of withstanding abnormal events from the power system and generator faults, which cause high transient stresses in the frame. Since the frame provides the basic support for the stator core, it must also be able to move with the core expansion and contraction from heating and the magnetic pulsating forces associated with the rotating flux patterns in the core. To accommodate all this, the core-to-frame mechanical coupling is usually done with some flexibility installed. This is typically done by providing keybar (frame) to core clearance when assembled, as well as using the soleplates which allow the frame to expand and contract radially. Frame stiffness and natural frequencies of vibration are important parameters due to the (120 or 100 Hz) mechanical and electromagnetic forces developed in the generators in conjunction with the stimulus from the power system frequency. Therefore, great care is taken to ensure that the natural frequencies of the core and the frame together are not near 120 (100) Hz or any multiple thereof. It is suggested that these natural frequencies differ at least 20% from all multiples and modes of 120 Hz to allow for safe operation of the machine. To provide stiffness for the outer shell of the frame or casing, there are frame rings or shelves welded to the wrapper at spaced axial intervals over the height of the stator as shown in Figure 2.2-1. These are designed to give the stator frame the strength it needs for its intended purpose of supporting the core. The entire frame structure is dimensioned to ensure the correct strength and to avoid the natural frequencies of the once-and twice-per-revolution characteristics of the generator. The type of material used in the frame is generally mild steel which is easy to weld with good strength and low-temperature ductility. A ventilation path must be provided to direct cooling air from the exit of the core (stator hot air) to the surface air coolers. The stator hot air is sent through stator frame ventilation path to the surface air coolers to become stator cold air after passing through the coolers. The air is then sent back through the generator to the various components such as the rotor field poles, stator core, circuit rings, and main and neutral leads to remove heat and become stator hot air again. Of course, the sizing of the cutouts and cooling passages is determined by the amount of cooling required in each part of the generator. Stator frames are also designed with lifting and handling in mind. Once a machine is built, it must be delivered to a site, and to do this, requires transportation by any number of means such as a large truck for smaller machines and by rail and ship for larger machines. The method of lifting is generally by craning and to achieve this, lifting beams are designed to get the stator lifted and set into the pit area. The machine may be erected in sections or as one piece depending on the size. It is, in fact, the transportation mode that governs the maximum size that a component can be manufactured. There is no point in building a machine so big that it cannot be transported to the generation site. Therefore, such things as overall weight of the stator and the transport system must be accounted for, as well as the overall dimensions. Some of the things to consider are time of year, rural road

54

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

conditions, clearance to railroad bridges, tunnels, station platforms, and other obstructions along the route. There is also another issue with large generator design that seems to be minimally considered during the design phase, undoubtedly due to size and cost considerations. This is the issue of accessibility to the core back and other generally inaccessible areas of the stator. Regardless of the discussion above on size and transportation, the inescapable reality is that all large generators require maintenance and need to be accessible for inspection and to carry out any repairs or modifications that may be required in future. More often than one would like, problems such as core and frame vibrations occur, resulting in the need to inspect for damage and make repairs or modifications, and the core and keybar interface area has limited accessibility. Attention to some “designed in” accessibility should be considered to accommodate future maintenance and inspections, although it is recognized that such accessibility would affect machine size and cost.

2.3 ELECTROMAGNETICS For simplicity, cross-sectional view presented in Figure 2.3-1 shows an airgap separating the slotted outer surfaces of both the rotor and the stator. The major elements of the magnetic circuit, as shown, are the rotor (including the rotor winding, pole bodies and the rim), the airgap (which constitutes the principal reluctance in the circuit), and the laminated steel stator core (including the stator teeth/slots and stator yoke below the slots). The airgap is the annular region between the rotor body and the stator core and probably has the largest influence on the electromagnetic design of the generator. Although the airgap is large to accommodate insertion of the rotor, it is small in relative terms to the rest of the magnetic circuit of the generator. It has a major influence with regard to the reluctance of the total magnetic circuit and, hence, the overall stability of the generator. The airgap greatly affects the steady-state stability of the generator when connected to the power system by simple variation of the length of the space between the stator and rotor outer surfaces. The length of this

Φ

1 Stator yoke 2 Stator teeth 3 Stator winding in slots 4 Salient pole rotor 5 Excitation winding 6 Magnetic field lines

Figure 2.3-1 Four pole generator flux pattern. Source: Courtesy of Voith.

2.3 ELECTROMAGNETICS

55

airgap is used to determine the short circuit ratio (SCR), which is calculated as described elsewhere in this book. In practical terms, this means that the longer the airgap, the higher is the magnetic circuit reluctance, and, therefore, the higher the short circuit ratio. Furthermore, the generator will tend to be more stable, producing higher ampere-turns (A-T) to achieve the required level of magnetic flux across the airgap. In real terms, this means more field current is required. A reasonable rule of thumb for the ampere-turns of the generator as a whole is that the airgap generally accounts for up to 90% of the total ampere-turns produced by the rotor. The remainder of the iron in the total magnetic circuit uses the other 10% or more and yet accounts for the majority of the electromagnetic flux path. This is because of the high permeability of the iron and high reluctance of air in the airgap. Therefore, a larger generator is required for higher apparent power output if the SCR ratio is to remain constant. This is because a larger rotor is required to handle the extra field current for the higher output and the airgap would be required to be about the same size to maintain the constant SCR. The airgap always needs to be large enough to permit insertion of the rotor through the stator bore with sufficient clearance for safe handling. This and stability requirements limit the minimum possible SCR in generators. Electromagnetic finite element analysis (FEA) is the preferred method to determine the actual magnetic field and its distribution in the machine and serves as a good visual representation to better understand what is happening within the machine at various loads. An example of a salient pole generator analysis on open circuit is shown in Figure 2.3-2, at full load in Figure 2.3-3, and during a sudden short circuit in Figure 2.3-4. In the open circuit example of Figure 2.3-2, the flux pattern is completely symmetrical about the pole axis of the rotor. Although the flux path includes the stator, the stator winding is on open circuit, and no current is flowing. Therefore, there is no back electromotive force (EMF) from the stator winding, and no electromagnetic torque coupling between the stator and rotor windings. In the case where the generator is connected to the system, there is current in the stator winding that is leading or lagging the voltage and significant torque is developed (see Figure 2.3-3). This shows the increase in pole flux density with increased load to compensate for the stator back EMF. As the turbine drives the rotor (counter clockwise direction in this example), the electromagnetic coupling between the stator and rotor windings tries to pull the rotor back in line with the axis of the stator poles. This difference in position of the stator and rotor pole axis creates a load angle that can be varied by changing the power output from the turbine and the field current for magnetic coupling between the stator and rotor. Increased field current pulls the rotor back toward the direct axis in the clockwise direction. In the case of the short circuit, the flux pattern symmetry is lost on the stator side and is now different on the rotor side. A few more interesting examples of FEA when the generator is in the overexcited and under excited modes of operation are shown in Figure 2.3-5–2.3-8. Again, these examples are presented to give the reader a visual idea of what is occurring magnetically inside the generator during these modes of operation.

56

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

A(Wb/m) 1. 2240e–001 1. 0490e–001 8. 7394e–002 6. 9891e–002 5. 2389e–002 3. 4886e–002 1. 7384e–002 –1. 1900e–004 –1. 7622e–002 –3. 5124e–002 –5. 2627e–002 –7. 0129e–002 –8. 7632e–002 –1. 0513e–001 –1. 2264e–001

B(T) 2. 4000e+000 2. 2286e+000 2. 0571e+000 1. 8857e+000 1. 7143e+000 1. 5429e+000 1. 3714e+000 1. 2000e+000 1. 0286e+000 8.5714e–001 6.8571e–001 5.1429e–001 3.4286e–001 1.7143e–001 0.0000e+000

Y

X Z

Time = 0.633 506 486 209 975 s Speed = 94.7368 RPM Position = 360.098°

Figure 2.3-2 Flux distribution and flux density at no load obtained by FE simulation (65 MVA, 76 poles, 13.8 kV, 396 slots). Source: Courtesy of Dr. Arezki Merkhouf.

A(Wb/m) 1. 5913e–001 1. 3282e–001 1. 0650e–001 8. 0191e–002 5. 3879e–002 2. 7566e–002 1. 2540e–003 –2.5058e–002 –5.1371e–002 –7. 7683e–002 –1. 0400e–001 –1. 3031e–001 –1. 5662e–001

B(T)

Y X Time = 0.633 506 486 209 975 s Speed = 94.736 800 RPM Position = 360.098264°

Z

2. 4000e+000 2. 2286e+000 2. 0571e+000 1. 8857e+000 1. 7143e+000 1. 5429e+000 1. 3714e+000 1. 2000e+000 1. 0286e+000 8. 5714e–001 6. 8571e–001 5. 1429e–001 Y 3. 4286e–001 1. 7143e–001 0. 0000e–000X

Z

Figure 2.3-3 Flux distribution and flux density at rated load obtained by FE simulation (65 MVA, 76 poles, 13.8 kV, 396 slots). Source: Courtesy of Dr. Arezki Merkhouf.

2.3 ELECTROMAGNETICS

57

B(T) 2. 4000e+000 2. 2286e+000 2. 0571e+000 1. 8857e+000 1. 7143e+000 1. 5429e+000 1. 3714e+000 1. 2000e+000 1. 0286e+000 8. 5714e–001 6.8571e–001 5.1429e–001 3. 4286e–001 1. 7143e–001 0.0000e+000

A(Wb/m) 5. 4090e–002 4. 5075e–002 3. 6060e–002 2. 7045e–002 1. 8030e–002 9. 0147e–003 –3. 4785e–007 –9. 0154e–003 –1. 8030e–002

Y

–2. 7046e–002 –3. 6061e–002

X Z

Time = 0.633 506 486 209 975 s Speed = 94.736 800 RPM Position = 360.098 264°

–4. 5076e–002 Y –5. 4092e–002

Figure 2.3-4 Flux distribution and flux density during sudden short circuit obtained by FE simulation (65 MVA, 76 poles, 13.8 kV, 396 slots). Source: Courtesy of Dr. Arezki Merkhouf.

Although detailed generator design work usually requires finite element analysis for accuracy and refinement, for some calculations such as the required excitation level for a high energy flux test on the generator stator core, a hand calculation of the total magnetic flux per pole in the generator is all that is needed, and it is determined as shown in Equation (2.2): Machine flux φ =

V LL × k 3 2∗

π 2

(2.1)

∗f kw N p

Simplified: Machine flux φ =

V LL ∗k Wb 7 7∗f ∗kw ∗N ph

where, VLL = line-to-line stator terminal voltage in volts k = number of stator winding parallel paths per phase f = frequency kw = stator winding factor (includes pitch and distribution) Nph = number of stator winding turns-per-phase

(2.2)

58

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

A(Wb/m) 2. 8369E–01 2. 4427E–01 2. 0485E–01 1. 6543E–01 1. 2601E–01 8. 6588E–02 4. 7168E–02 7. 7475E–03 –3. 1673E–02 –7. 1093E–02 –1. 1051E–01 –1. 4993E–01 –1. 8935E–01 –2. 2877E–01 –2. 6819E–01 –3. 0761E–01

Y X Time = 0.505 013 299 449 292 s Speed = 128.600 000 RPM Position = 37.420 362°

Z 0

200

400 (mm)

Figure 2.3-5 Shows the flux distribution in the over-excited mode with field current at 1800 A (310 MVA, 56 Poles, 128.6 RPM). Source: Courtesy of Dr. Arezki Merkhouf.

The winding factor of the machine is largely concerned with reducing harmonic effects and wave shaping. It is comprised of the pitch and distribution factors. The pitch factor is determined from a winding diagram and depends on the number of slots separating the distance (the coil span) between connection from top and bottom coil legs or bars in series, that is, a top leg in slot 1 connected to a bottom leg in slot 7 gives a span of 6 and for 195 slots, 26 poles, gives slots per pole of 7.5. The machine therefore would have a stator winding pitch of 6/7.5 or 0.8. The distribution factor deals with the fact that the EMF induced in different slots are not in phase, therefore, their vector sum must be less than their arithmetic sum. The distribution factor therefore is the ratio of the vector sum divided by the arithmetic sum of the stator coil EMFs for this distribution. To work out the winding factor (kw) from the pitch (kp) and distribution (kd) factors see Equation (2.3): kw = k d ∗kp = sin β 2 η × sin γ 2 × sin ρπ 2

(2.3)

2.3 ELECTROMAGNETICS

59

B (tesla) 3. 1367E+00 2. 9276E+00 2. 7185E+00 2. 5094E+00 2. 3002E+00 2. 0911E+00 1. 8820E+00 1. 6729E+00 1. 4638E+00 1. 2547E+00 1. 0456E+00 8. 3645E–01 6. 2734E–01 4. 1823E–01

Y

2. 0911E–01 6. 3159E–07

Time = 0.505 013 299 449 292 s Speed = 128.600 000 RPM Position = 37.420 362°

Z 0

200

X

400 (mm)

Figure 2.3-6 Shows the flux density in the over-excited mode with field current at 1800 A (310 MVA, 56 Poles, 128.6 RPM). Source: Courtesy of Dr. Arezki Merkhouf.

where, β = π/number of phases = π/3 η = number of slots/number of poles/number of phases γ = π/(number of slots/number of poles) ρ = stator winding pitch (from winding diagram) Equation (2.2) provides the basic level of machine flux required to achieve rated line-to-line terminal voltage in a generator, given a specific winding configuration. This formula will be elaborated on in Chapter 11 and an example provided for determining excitation levels in a flux test. Generators are made with different power factor ratings. The most common are 0.90 and 0.85 lagging. Two machines of the same MVA rating will have different capability design parameters for the two different power factors. The 0.85

60

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

A (Wb/m) 2. 2736E–01 1. 9625E–01 1. 6514E–01 1. 3402E–01 1. 0291E–01 7. 1796E–02 4. 0683E–02 9. 5699E–03 –2. 1543E–02 –5. 2657E–02 –8. 3770E–02 –1. 1488E–01 –1. 4600E–01 –1. 7711E–01 –2. 0822E–01 –2. 3934E–01

Y

X Time = 0.505 013 299 449 292 s Speed = 128.600 000 RPM Position = 37.420 362°

Z 200

400 (mm)

Figure 2.3-7 Shows the flux distribution in the under excited mode with field current at 1190 A (310 MVA, 56 Poles, 128.6 RPM). Source: Courtesy of Dr. Arezki Merkhouf.

power factor machine will require more field current to achieve the same power at the 0.85 power factor. Hence, the machine is somewhat larger to accommodate a rotor that can handle more field current and cooling capacity and is more costly to build. It is easy to see that design optimization to make the best utilization of the magnetic materials is a design priority. The flux density becomes the driving factor for the amount of stator core material that is required. As can be seen from Figures 2.3-2 and 2.3-3, the flux densities are different between open circuit and full load, but only marginally higher on load. However, there is considerable redistribution of the flux when the machine is on load, due to the stator currents. On open circuit, the stator core does not approach the electromagnetic loss limits of the iron, which are typically in the

2.3 ELECTROMAGNETICS

61

B (tesla) 2.9979E+00 2. 7980E+00 2. 5981E+00 2. 3983E+00 2. 1984E+00 1. 9986E+00 1. 7987E+00 1. 5989E+00 1. 3990E+00 1. 1991E+00 9. 9929E–01 7. 9943E–01 5. 9957E–01 3. 9972E–01

Y

1. 9986E–01 0. 0000E+00

X Time = 0.505 013 299 449 292 s Speed = 128.600 000 RPM Position = 37.420 362°

Z 200

400 (mm)

Figure 2.3-8 Shows the flux density in the under excited mode with field current at 1190 A (310 MVA, 56 Poles, 128.6 RPM). Source: Courtesy of Dr. Arezki Merkhouf.

1.7 T range in the stator teeth and under 1.45 T in the stator core back. Lower flux densities will typically be found in the rotor rim, but they are induced by the DC current in the field winding, and so do not cause losses. That is to say, they are unidirectional as far as the rotor is concerned and so there are no eddy current losses in the rotor body due to the main flux. It is the alternating effect in the stator that designers are concerned with, in this instance. Heating of the rotor components is a concern, but more so because of the I2R losses in the field winding as opposed to the effects of magnetic interaction with the stator slots causing heating due to induced stator slot ripple effects (i.e. the variation of the main field due to the slotting of the stator).

62

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

2.4 CORE-END HEATING In addition to the electromagnetics of the main flux distribution across the airgap and in the main body of the stator and rotor, there are end region (also known as core-end heating) effects of the flux produced. The core-end heating effects arise from the endwindings of the stator and rotor, and the core-end fringe effects. This can occur, although rare, particularly when the generator is operating in the leading power factor range and can be exacerbated when the axial alignment of the rotor with respect to the stator is out of tolerance. If the axial alignment is allowed to exceed the OEM tolerances, the end flux from the rotor is perpendicular to the end stator laminations and eddy currents are induced causing possible overheating of the core-end. This can be corrected with the proper axial alignment of the rotor, consult the OEM for the allowed tolerance. Reference [20] addresses the axial misalignment between the top and bottom rotor ends with respect to the stator end and should not exceed 20% of the airgap dimension.

2.5 FLUX AND ARMATURE REACTION The rated apparent power of a generator is proportional to the flux and the armature reaction, in the relationship as shown in Equation (2.4) MVA = KM a ΦPf

(2.4)

where, MVA = rated apparent power K = a proportionality constant Ma = armature reaction Φ = magnetic flux per pole at rated voltage in Webers P = number of poles F = frequency This is really the same as the product of the stator current and the stator terminal voltage. The stator or armature current is proportional to the armature reaction. The stator voltage is proportional to the flux. The field winding ampere-turns or field current at rated load is directly related to the level of armature reaction. Calculation of the flux per pole is described in Section 2.3 and the calculation of armature reaction is as shown in Equation (2.5):

Ma =

Nph × 2P

Nst Nph

k ∗Ia A − T Kp Kd

(2.5)

2.5 FLUX AND ARMATURE REACTION

63

where, Nph = number of phases P = number of poles Nst = number of full turns k = number of parallel paths Kp = winding pitch factor Kd = winding distribution factor Ia = stator current in amperes One other basic relationship that governs the rating of a generator is the output coefficient. Simply put the output of the generator increases with the square of the diameter of the rotor or stator bore, and with the height of the machine, based on the following relationship as shown in Equation (2.6) Output coefficient =

MVA MVA min m3 D2b LS

(2.6)

where, MVA = rated apparent power Db = diameter of the stator bore in meters L = height of the active iron in the stator in meters S = speed of the rotor in RPM Specific generator ratings are accommodated in machine design by trading off the levels of magnetic flux against the level of armature reaction. The actual component dimensions as described above also play a role in optimizing designs of large generators. Therefore, a specific rating can be achieved by a relatively high value of flux and a low level of armature reaction, and vice versa, or some combination in between. Increasing the generator output at a specific combination of flux and armature reaction can also be done by making the machine taller or longer. Using all these factors, one can design a machine to fit any output rating desired. However,when one parameter changes, it affects all the other parameters, some marginally, but some others significantly. Two additional formulas that help to describe the output of the generator in relative terms are specific magnetic loading and specific electric loading. These two formulas as shown in Equations (2.7) and (2.8) in their more basic form can be multiplied together to produce the output coefficient above: Magnetic loading =

Electric loading =

PΦ Wb m2 LDb

(2.7)

I st N c A m Db k

(2.8)

64

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

where, P = number of poles Φ = flux per pole in Webers Db = stator bore diameter in meters L = core iron effective length in meters Ist = stator, single-phase current in amperes Nc = total number of stator conductors k = number of parallel paths in stator circuit Using the above formulas, one can compare basic machine design outputs to determine which is more highly loaded in specific terms. For instance, if a machine is prone to high core-end heating, the specific electric loading of the generator is likely to be high relative to other machines, indicating that high stray losses are present. High stray losses can directly affect core-end flux penetration and, subsequently, the level of core-end heating. Machines with a high level of flux require a relatively large volume of iron to carry the flux and a relatively small amount of copper to carry the stator and field currents. Such machines tend to be larger and more costly to build. Machines with a low level of flux require a relatively small volume of iron to carry the flux but a relatively large volume of copper in their windings. Such machines are termed “copper rich,” and they increase the problem of heat removal from the windings. These machines tend to be smaller and less costly to build. The per-unit transient and subtransient reactances, which play a significant role in the electrical performance of the generator connected to the power system, tend to be low with high-flux levels. The higher-flux generator will, therefore, tend to have a somewhat better inherent transient stability. It will also tend to have higher per-unit transient currents during severe disturbances and, therefore, higher winding forces and torques, than a lower-flux machine. To limit fault currents in the generator and, hence, the forces and torques, minimum values of subtransient reactance are usually specified. The subtransient reactance is a function of the stator leakage reactance and the effects of the rotor amortisseur or damper winding.

2.6 STATOR CORE AND FRAME FORCES As discussed earlier, the principle function of the stator core is to carry electromagnetic flux. The core must handle magnetic field flux densities in the stator teeth and in the core-back or yoke area. The magnetic field is revolving, so it creates an alternating voltage and current effect in the generator components, which is a source of high losses and heating. This alternating effect also causes vibration of the core at the rotational frequency and with harmonics due to the nature of the flux patterns. Because of the inherent vibration and the large mechanical and thermal forces involved, the core must be held solidly together so that there are no natural

2.7 STATOR WINDINGS

65

frequencies near the once and twice per revolution forcing frequencies. Some cores are installed as continuous pile, while others have core splits. Care must be taken with core splits as the packing material between the splits installed in the original day may be dried out, brittle, coming out of the split at the back of the core, and no longer able to perform its designed function. The function of the packing material is to consolidate the split sections of core so that the core behaves as a solid mass when excited by the field current. If this packing is deteriorated or missing, the core may not behave has designed once excited. A sure sign the packing material is missing is fretting at the split of the core. Designers take great care to ensure that the natural frequencies of the core or core and frame are not near 120 (100) Hz or equivalent to other induced forcing frequencies. It is desirable to keep the natural frequencies of the core and frame at least 20% away from forcing frequencies. An example of a forcing frequency above 120 (100) Hz is typically a tooth-pass frequency which is equal to the stator slots per pole rounded to the closest integer multiplied by 120 (100) Hz. There also is a large rotational torque created by the electromagnetic coupling of the rotor and stator across the airgap. This is in the direction of rotor rotation. The torque due to the magnetic field in the stator core iron is transmitted to the core frame via the keybar structure at the core back. Therefore, the stator frame and foundation must be capable of withstanding this torque, as well as large changes in torque when there are transient upsets in the system or the machine. The natural vibration inherent in the core must also be accounted for in the core to frame coupling. Heating and cooling effects in the core and frame materials will also affect this coupling and vibration, due to differences and rates of thermal expansion and contraction in the core and frame components.

2.7 STATOR WINDINGS The stator winding is made up of insulated copper conductor bars or coils that are distributed around the inside diameter of the stator core, commonly called the stator bore, see Figure 2.7-1 for typical multi-turn coil configurations and Figure 2.7-2 for a typical single turn Roebel bar configuration. The winding is installed in equally spaced slots in the core to ensure symmetrical flux linkage with the field produced by the rotor. Each slot contains two Roebel bars or coils, one on top of the other (see Figures 2.7-3 and 2.7-4). These are generally referred to as top and bottom bars or top and bottom legs. Top bars or legs are the ones nearest the slot opening (just under the wedge) and the bottom bars or legs are the ones at the slot bottom. The core area adjacent to the slot is generally called the core teeth as shown in Figure 2.7-5. The stator winding is then divided into three phases, which are almost always wye-connected. Wye connection is done to allow a neutral grounding point and for relay protection of the winding. The three phases are connected to create symmetry between them in the 360 arc of the stator bore. The distribution of the winding is done in such a way as to produce a 120 difference in voltage peaks from one phase to the other, hence the term “three-phase voltage.” Each of the three

66

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Case 1

Case 2

Case 3

2 Turns/coil

2 Turns/coil

2 Turns/coil

6 Strands/turn

10 Strands/turn 5.40 mm × 4.50 mm

20 Strands/turn

5.40 mm × 7.70 mm

5.40 mm × 2.10 mm

Figure 2.7-1 Typical multi-turn coil strand configurations. Source: Courtesy of Dr. Michael Znidarich & Engineers Australia [2].

Transposition epoxy putty

Transposition crossover insulation

Copper strands

Groundwall insulation

Integrally moulded corona portection layer

Figure 2.7-2 Typical single turn Roebel bar configuration. Source: Courtesy of Dr. Michael Znidarich & Engineers Australia [2].

phases may have one or more parallel circuits within the phase. Multiple groups of coils can be connected in series to form the entire phase circuit. The parallels in all of the phases are equal, on average, in their performance in the machine. Therefore, they each “see” equal voltage and current, and magnitudes and phase angles, when averaged over one alternating cycle.

2.7 STATOR WINDINGS

Wedge and driver Ripple spring or flat wedge Groundwall insulation Wedge depth packing

Solid copper strands Strand insulation Middle separator strip

Flat/ripple side packing or CRTV coating or wrapper and RTV filler Semiconducting layer

Slot bottom strip

Figure 2.7-3 Cross section of stator bar in the stator slot.

Wedge and driver Ripple spring or flat wedge Groundwall insulation Wedge depth packing

Dedicated turn insulation

Solid copper strands Strand insulation

Middle separator strip

Flat/ripple side packing or CRTV coating or wrapper and RTV filler

Semiconducting layer

Slot bottom strip

Figure 2.7-4 Cross section of a stator multi-turn coil in the stator slot.

67

68

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Slot wedge

Under wedge ripple spring Conductive slot side spring

Laminated stator core

Copper strands (20 per turn)

RTD three wire lead

Turn to turn insulation Ground insulation

Conductive bottom slot strip Conductive middle slot strip Integrally moulded corona conductive layer

Figure 2.7-5 Shows a two turn bar in the stator slot (teeth are adjacent to bar). Source: Courtesy of Dr. Michael Znidarich & Engineers Australia [3].

The stator winding in any particular phase group are arranged such that there are parallel paths that overlap between top and bottom bars or coils, see Figures 2.7-6 and 2.7-7, for examples of reversing group jumpers. The overlap is staggered between the top and bottom bars or coils. The top bars or legs in the first pole group are connected to the bottom bars or legs in the next pole group in one direction, whereas the bottom bars or legs in the first pole group are connected in the other direction on the opposite side pole group. This connection with the bars or legs on progressive pole groups around the stator creates a “reach” or “pitch” of a certain number of slots. The pitch is, therefore, the number of slots that the stator bars or coils have to reach in the stator bore arc, separating the two bars or coils to be connected. This is almost always less than one pole pitch and it is done to assist in reducing the harmonics induced in the stator winding. Once locally connected, bars or coils form a group. A group may be a parallel circuit, or a full phase. Parallel circuits may be connected in series with other parallel circuits to form a full phase. The total width of the overlapping parallels is called the “breadth.” The combination of pitch and breadth create a “winding or distribution factor.” The distribution factor is used to minimize the harmonic content of the generated voltage. In the case of a two parallel path or more winding,

2.7 STATOR WINDINGS

69

Bottom bar Top bar jumper jumper

Figure 2.7-6 Stator endwinding showing reversing group jumper connection on a bar winding.

Circuit rings

Back legs

Front legs

Figure 2.7-7 Stator endwinding showing reversing group jumper connections on a multi turn winding.

these are connected in series or parallel via the circuit rings outside the stator bore, see Figures 2.7-6 and 2.7-7. The connection type will depend on a number of other design issues regarding current-carrying ability of the copper in the winding.

70

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Consider for simplicity a two parallel path, three-phase-winding. Alternating voltage is created by the action of the rotor field as it moves past these windings. Since there is a plus and minus, or north and south, to the rotating magnetic field, opposite-polarity currents flow on each side of a given pair of poles in the distributed winding. The currents normally flowing in large hydro generators can be on the order of thousands of amperes. Due to the very high currents, the conductor bars or coils in a hydro generator have a large cross sectional area. The high current capacities of copper in the stator bars or coils generate significant heat. The losses due to the flowing currents are called I2R or “copper” losses in the winding. Controlling the losses in the stator winding requires careful design consideration because of the variance in magnetic field from the stator bore toward the slot bottom. The magnetic field tends to be more intense toward the top of the slot and, therefore, the top bars or coil legs generally produce more voltage than the bottom bars or legs. This difference in voltage can produce circulating losses if not properly managed. Bars will always have Roebel transpositions within the bar to balance the voltage between strands before joining the ends. Multi-turn coils will be transposed, whether internally or externally to closely manage the voltage difference between strands to minimize circulating loss. Within the bars or legs themselves, there are also eddy currents flowing in the individual strands of the conductors caused by the localized-leakage magnetic field. It is important to choose the conductor size in order to reduce these eddy current losses. An exaggerated example of a single strand versus many individual strands and the magnitude of eddy currents is shown in Figure 2.7-8. This figure illustrates the concept that smaller individually insulated strands will reduce the magnitude of the overall eddy current thus reducing the losses. Figure 2.7-9 further illustrates this concept with progressively smaller individual strand cross sections. To further reduce the effect of the eddy currents within each Multi-turn coil, the manufacturer may choose to use a technique called an inverted turn. This

Figure 2.7-8 Showing reduced eddy current losses with individual strands of copper instead of one large piece. Source: Courtesy of Dr. Michael Znidarich & Engineers Australia [4].

2.7 STATOR WINDINGS

Case 1

Case 2

Case 3

2 Turns/coil

2 Turns/coil

2 Turns/coil

6 Strands/turn

10 Strands/turn 5.40 mm × 4.50 mm

20 Strands/turn

5.40 mm × 7.70 mm

71

5.40 mm × 2.10 mm

Figure 2.7-9 Individual strand cross section to reduce losses from right to left in the figure. Source: Courtesy of Dr. Michael Znidarich & Engineers Australia [4].

Inverted turn

Coil loop

Figure 2.7-10 Looping of a coil with inverted turn. Source: Courtesy of Dr. Michael Znidarich & Engineers Australia [4].

inverted turn is done in the endwinding of the multi-turn coil, and depending on the design, it can be done on any turn of the coil, wherever the most reduction in eddy currents is calculated during the design stage, see Figure 2.7-10 during the looping process of coil manufacturing.

72

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Similarly, to reduce the effect of the eddy currents within each individual stator bar, the conductors are made up of numerous copper “strands” (refer back to Figure 2.7-2). This is analogous to the reasoning behind the stator core being made up of very thin insulated laminations rather than a solid mass of steel. However, although the strands are insulated from one another in the bar or coil, they are eventually connected at each end of the stator bar or coil. Therefore, additional circulating current could flow from the top to the bottom strands in a single bar or coil. This is due to the difference in the magnetic field from the top to bottom of the slot. To reduce the effect of the circulating currents, the strands are “Roebel transposed” in each bar (see Figures 2.7-11 and 2.7-12). Roebel transposition of the copper strands refers to the repositioning of each strand in the stator bar stack such that it occupies each position in the stack at least once over the full length of the stator bar. Roebel transpositions are mainly 360 and 540 . A 360 transposition means that each strand occupies each position once over the length of the bar, and a 540 transposition means that each strand occupies each position one-and-a half times. The 360 transposition is generally done in the slot only and the 540 transposition includes the very ends of the stator bars, and in the curved endwinding portion as well.

5 6 7 8

1 2 3 4

1 5 6 7

2 3 4 8

2 1 5 6

3 4 8 7

3 2 1 5

4 8 7 6

4 3 2 1

8 7 6 5

8 4 3 2

7 6 5 1

7 8 4 3

6 5 1 2

6 7 8 4

5 1 2 3

5 6 7 8

1 2 3 4

Figure 2.7-11 Roebel transposition 3D view. Source: Courtesy of Dr. Michael Znidarich & Engineers Australia [4].

2.7 STATOR WINDINGS

1′

2′

3′

4′

1′

2′

3′

4′

1′

2′

3′

4′

1′

2′

3′

4′

1′

2′

3′

4′

1′

2′

3′

4′

1′

2′

3′

4′

1′

2′

3′

4′

73

Figure 2.7-12 Roebel bar principle. Source: Courtesy of Dr. Michael Znidarich & Engineers Australia [4].

There is another problem with circulating currents that occurs in doublestack stator bars, that is, designs where there are two separate Roebel-transposed stacks side by side, thus giving four strand widths in the bar. This double stacking is rare for a hydro generator, but there are machines that have this arrangement. Although it appears that the stacks are so close together that there would be no difference in magnetic field from one side of the bar to the other, this is not true. In fact, there is a significant difference, because the magnetic field does cut between the stacks such that a certain amount of circulating current occurs in a double-stack stator bar. The amount of circulating current from one stack to the other in a single bar can cause temperature differentials from on stack to the other up to 10 C on average. Figure 2.7-13 shows a normal type of temperature profile for a double stacked stator bar with separate Roebel transpositions for each stack. The temperature difference from one side to the other has the overall effect of reducing the available stator current output because the maximum hot-spot temperature is raised by about 10 C during operation. Obviously, eliminating this temperature difference would allow higher output from the same slot dimensions of a bar if the temperature hotspots could be reduced. A method has been developed that does allow for a more even temperature distribution across the strand stack and it is termed cross-Roebel (see Figure 2.7-14). The method simply has all strands transposed such that they occupy both sides of the bar stack and not just one side. In this method, the two strand stacks are balanced electromagnetically from side to side as well, and this eliminates the stack-to-stack circulating currents (see Figure 2.7-15).

74

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Double roebel bar

Temperature profile for separate transpositions

°C 40.0 35.0 30.0 25.0 20.0 15.0 10.0 5.0

1 5 9 13 17 21 25 29 33 37 41 45 48 53 57 61 65 69 73 77 81

0.0 Side A

Side B

Side A

Side B Strands

Figure 2.7-13 Temperature profile of a double-stack stator bar with separate Roebel transposed stacks. The average temperature difference between side A and side B of the bar example shown is about 10 C. Source: Courtesy of Alstom Power Inc.

L/4 180°

L/2 +

180°

L/4 +

180°

= 540°

Figure 2.7-14 Cross Roebel transposition-temperature profile of a double-stack stator bar. Source: Courtesy of Alstom Power Inc.

Cross roebel bar

Temperature profile for cross roebel transposition

°C 40.0 35.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0 1

5

9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 Strands

Figure 2.7-15 Cross Roebel transposition – temperature profile of a double stack stator bar. All cooling strands are temperature equalized due to the elimination of stack-to-stack circulating currents. Source: Courtesy of Alstom Power Inc.

2.7 STATOR WINDINGS

75

There are many ways of designing stator conductor bars, depending on the size and cooling method required for the machine. Cooling is particularly critical in designing machines for higher outputs. In this regard, direct cooling is the most desirable type of cooling because it increases the generator stators current carrying capability considerably. The advantage of this is to reduce flux levels and, hence, the physical size and weight of the generator. The basic limit for conventionally cooled generators (i.e. indirect cooling with air) is now in the 944.5 MVA range. Hydro generators up to 855 MW have been built with direct conductor cooling. In indirectly cooled machines, the strands within the conductor bars are all solid and the heat generated in the conductors is removed by conduction through the groundwall insulation to the stator core. The size of the generator is significantly limited by the temperature conduction through the groundwall insulation to the stator core. In direct water cooled windings, the copper strands are made hollow, to carry liquid coolant. The stands are generally rectangular in shape to allow stacking and they are each individually insulated from one another and Roebel transposed. Figure 2.7-16 shows a three-dimensional representation of what a typical stator bar looks like when inserted into the stator core as well as the strand arrangement for this particular design. In this mixed strand arrangement, the hollow

Figure 2.7-16 Shows a 3D representation of a typical water cooled stator bar in the slot section of the stator core. Source: Courtesy Voith.

76

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

strands are evenly interspersed among solid strands. The strands can be arranged in various combinations to produce more efficient winding designs. In directly cooled stators, it is possible to increase the current density in the copper winding of the stator to achieve higher ratings. Trade-offs are also made between slot sizes and winding configurations to find the optimum terminal voltage level versus the current flowing in the stator winding, all in consideration with keeping magnetic flux densities in the stator iron at manageable levels. Because the stator current densities in directly cooled windings are so much higher than in indirectly cooled windings, designers must also consider the effect of transients and temperature rise. Considerations of reactance and stability also come into play and, therefore, so do short circuit ratio and excitation performance. Some modern generator designs mix solid copper stands for conduction of the electrical current and hollow stainless steel strands for carrying the coolant. Figure 2.7-17 shows an arrangement for a direct cooled winding where the headers and hoses carry the coolant to the winding. This design has been in service for the last 30 years and has been successful. The use of stainless steel strands for cooling has eliminated certain industry problems of copper erosion and corrosion in the stator bars. The mixed steel and copper stator bars also tend to be more rigid than fully copper bars and allow higher wedging pressures in the slot. In direct water cooled machines, the cooling method dictates the need for an external system to remove the heat picked up by the stator cooling water after it passes through the stator winding. Therefore, an external system is attached to the generator that employs heat exchangers to accomplish this function. To circulate the water, pumps and a piping system are provided. In addition a filtering system is provided to remove any large particles suspended in the stator cooling water

Figure 2.7-17 of Voith.

Direct cooled stator winding (rotor poles removed). Source: Courtesy

2.7 STATOR WINDINGS

77

that can cause blockage within the stator windings inside the generator. Since the water is in contact with current-carrying copper conductors, which are also operating at voltage levels from ground potential up to 23 kV, the water must be kept absolutely as pure as possible to avoid flashovers by conduction through the water. To maintain pure water, a de-ionizing system is provided. See Chapter 3 for a description of the stator cooling water system. The basic functions of electrical insulation in the stator winding are to maintain ground insulation between the conductors and the stator core and other grounded objects, and to maintain insulation between turns of multi-turn coils and between the strands within a turn. The groundwall insulation must be designed to withstand line-to-line AC voltages over the entire life of the generator. In addition, it must be capable of withstanding overvoltages from system faults. The turn insulation must withstand normal coil voltage over its lifetime, with substantial short time overvoltages in the event of a steep-front voltage surge such as system faults or lightning strikes local to the generating station. Strand insulation is exposed to only a few volts with brief overvoltages during occasional high-current transients. A high resistance coating or “semiconducting” system is applied on top of the groundwall insulation in the slot to control the voltage distribution over the length of the slot for machines with terminal voltages in excess of 6 kV depending on the designer’s preference (see Figure 2.7-18). In addition, a special “grading” system is applied to the bars or coils over a short distance starting a few inches from the bar or coil exit from the slot to part way into the endwinding area. This grading system is typically a Silicon Carbide or Iron Oxide coating. The grading system allows for a gradual voltage drop in the endwinding to the stator core. The endwinding is at line potential and surface current flows from the endwinding through the grading system to the stator core. Some manufacturers apply the grading material over the full length of the end turn.

Figure 2.7-18 Shows the black semiconducting material and the grey Silicon Carbide grading material.

78

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

To ensure good contact between the stator winding and the core in the slots, a side packing filler is inserted between the coil and core section inside the along the entire length. The side filler is impregnated with semiconducting material to assist with the electrical contact to the stator core. The base material is usually made up of strong resin-filled woven glass material. It may be a flat piece but a ripple-spring filler is now commonly used to ensure continual pressure and contact over the life of the winding (see Figures 2.7-19 and 2.7-20). Another popular method used by some manufacturers to side pack the winding is to use a sheet of soft semiconducting material with an adhesive applied to it, and wrap the slot section of the coil or bar forming a blanket over the winding. The adhesive is applied to one side of the blanket only. To illustrate the application of the adhesive, imagine a piece of cloth that is as long as the coil or bar and wide enough to wrap three sides of the coil or bar. Then, place a bead of adhesive in an “S” shape covering the cloth. The cloth is then wrapped around three sides of the coil or bar, the bare side (without the blanket) faces the airgap of the machine. The coil or bar with this new “blanket” is then inserted into the slot section and the adhesive conforms to the slot section securing it into place. The adhesive is between the winding and the blanket only, the core does not come into contact with the adhesive. Only the dry side (where no adhesive has been applied) of the blanket contacts the core. This type of semiconductive installation requires the installer to apply just the right amount of adhesive. Too little will allow for a loose winding, and too much will be wasteful and create a much larger cleanup than necessary. Further, if the adhesive material smears into the wedge groove, wedging will be difficult. Some manufacturers apply the bead of conducting or nonconducting

Figure 2.7-19 Flat side-packing (top) and ripple spring (bottom) with semiconducting impregnation.

2.8 STATOR WINDING WEDGES

Figure 2.7-20

79

Shows stand up view of Figure 2.7-19 to illustrate the ripple spring.

adhesive on the blanket, fold it in half, and then wrap this strip in a spiral fashion around the coil or bar. Due to the current flowing in the stator bars or coils, there is a reaction force in each slot, which varies according to the level of current and direction of flow at any instant. This creates forces between bars or coils that are both repulsive and attractive at any given time in the alternating cycle. Therefore, the slot section of a stator conductor bar or coil “sees” significant and constant (mainly radial) vibration forces at the twice-per-revolution frequency. The stator bars or coils tend to vibrate in the slot, a phenomenon called “bar or coil bouncing” (see Chapter 5). Therefore, the stator bars or coils must be tightly wedged in the slot to eliminate the relative motion and avoid fretting damage from contact against themselves and the stator core and bar or coil packing systems. Coil side packing typically should be on the trialing side referred to the direction of rotation of the unit. Stator windings have been known to fail quickly once they become loose in the slot.

2.8 STATOR WINDING WEDGES There are many different wedging systems employed by different manufacturers, too numerous for all to be covered in this book, however, all have the common purpose of keeping the stator bars tight in the slot. For an example of a few types of common hydro generator wedges see Figures 2.8-1 and 2.8-2. Starting with Figure 2.8-1, the far left of the picture is the depth packing material of varying thicknesses that is placed between the bar or coil surface and the drive or stationary wedge, or just the wedge body depending on the wedging system. The amount of

80

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Slab wedges

Packing materials Stationary and drive wedges

Figure 2.8-1 Typical packing material and wedge assemblies.

packing material depends on how much space exists to be filled. Sufficient packing is installed to tighten the wedge assembly to the OEM installation procedures. These procedures are normally supplied when wedges are purchased for a stator re-wedge (ask for the procedure) or when the machine is new and wedges are being installed. If there is insufficient packing material, the wedge system will be loose. If there is too much packing material, the wedge may crack under the extreme pressure that is placed on it, this is particularly true for systems with drive wedges. To the right of the packing material in Figure 2.8-1 is a typical “slab” wedge made of some sort of insulating material from the early days such as Micarta™ or even wood such as Maple. Split Maple wedges are split axially with opposing tapers and were very popular in the early days of hydro generators. The slab wedge will have different methods used to tighten it against the top coil or bar in the slot. For example, a stationary or drive wedge between it and the packing material can be used or the slab can be driven over flat filler. The final wedge assembly to the right of the slab wedge in Figure 2.8-1 is a three-part wedge system consisting of the main wedge itself, a drive wedge, and a stationary wedge. The drive and stationary wedge are fixed in length and taper. All materials in modern systems typically consist of some sort of epoxy mixed with fiberglass and processed to form the shapes as shown. This system typically allows for the main wedge itself to deflect slightly when in its final position providing the “spring” action pushing against the coil surface to keep everything tight. The amount of “spring” provided by this design is minimal due to the small amount of deflection of the wedge. In this system, it is very easy to put too much pressure on the main wedge by putting too much packing material and overdriving the “drive” wedge into its final position. Cracking of the main wedge over time will result if too much deflection is allowed particularly since this system has a very thin wedge. When designing slab type wedges over flat sliding fillers, it is important to know the flexural strength of

2.8 STATOR WINDING WEDGES

81

the wedge material to ensure it can provide the necessary retention force for the coils in the slot. For this reason one has to carefully choose the “fiberglass laminate” that is used. Not all laminates are the same. During installation, it is important to utilize the gauge that is supplied by the manufacturer during the winding process in order to get the optimal deflection of the wedge. The gauge is manufactured on site by using a dial indicator placed on the wedge surface and driving the “drive wedge” into position. Depending on how much packing is in place, the drive wedge will need to be driven a longer length or a shorter length to get the dial indicator to the desired reading as the main wedge deflects outwards. It is important to observe that the dial indicator reading is actually measuring the deflection of the wedge outwards and not the movement of the wedge outwards because it is not seated properly in the groove. A gauge can then be made out of aluminum with various numbers of lines marked on it that can be inserted into the drive wedge position to determine if enough or not enough packing has been installed to get the proper wedge deflection. It is important to recognize that when this main wedge deflects (crowns) into the airgap, the amount of wedge left “holding on” in the wedge groove is reduced and thus the coil retention effectiveness may also be reduced. To summarize, the amount of packing material is critical in ensuring the wedge remains tight for as long as possible for the life of the winding. It is quite likely that a re-wedge may be necessary more than once in the life of the winding using this system. Figure 2.8-2 has four different styles or components of wedge assemblies which will now be discussed. The top wedge assembly is a typical fiberglass wedge body with a ripple spring and drive wedge assembly to compress the spring once installed. As with the other systems discussed, the depth packing is critical in ensuring the right amount of spring compression occurs during in installation. Most wedge spring assemblies should be compressed to 80% of their original value to ensure proper pressure over the life of the winding. Of course, consult the spring manufacturer or generator OEM for the exact compression amount for the specific installation at hand. Checking the amount of spring compression is done in two ways. The first way is to insert a feeler gauge into the air vent notches in the wedge to see how much compression the spring has. Again, the spring manufacturer or generator OEM will have the feeler gauge thickness to use as a “go-no-go” gauge. Another way to measure spring compression is to use a wedge with holes drilled into the wedge assembly as shown in the last wedge body in Figure 2.8-2. A gauge with a small needle head is used to measure the spring compression in each successive hole. These numbers should be within a tolerance set by the spring manufacturer or generator OEM. A variation of these holes drilled into the wedge body is to machine a groove the same length and width as all of the holes so the same gauge can now slide along the groove length to measure the spring compression. The slot will contain at least three wedges of this type (top, middle, bottom) or more depending on the stack length. The wedge itself, having a groove or holes machined into it, is obviously weaker than a wedge that is solid. However, if designed correctly, it should be sufficiently strong to provide its intended function along with the rest of the solid wedges in the slots. Consultation with the generator OEM is recommended if there are any doubts about the wedge integrity. The small wedge at the

82

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Figure 2.8-2 Various types of wedge assemblies.

bottom of Figure 2.8-2 has an interesting design feature; it has Kevlar® wrapped around it. The Kevlar® provides a less abrasive and softer surface against the stator iron so if the wedge were to become loose and vibrate in the groove, damage to the core would be minimized (see Figures 2.8-3 and 2.8-4). The typical fiberglass and epoxy wedge is quite abrasive when it moves relative to the core and can cause abrasion of the iron if left unattended. The Kevlar® wrap is applied in a mold to the wedge (the wedge is made in a mold as well, it is not a machined wedge) and is a very expensive, not very common, and extremely difficult to source. It has been used successfully in machines with a very small wedge groove where machining a fiberglass wedge proves difficult with the tight tolerances required for a proper fit. The second and third wedge assemblies are typical ripple spring and driver types. There are two important fundamental differences between the two wedge assemblies. The first is that the ripple spring in the third wedge sits between the depth packing and the wedge body with the driver in the wedge groove. The third wedge assembly has the ripple spring in the wedge groove with the tapered driver against the depth packing surface. Either assembly will perform well over the life if the depth packing is correctly installed and the wedge assembly remains tight. The

2.8 STATOR WINDING WEDGES

Kevlar wrapped wedge

Fretting dust

Core packet

Figure 2.8-3 Shows core packet and fretting at the interface of wedge groove.

Wedge groove after cleaning

Figure 2.8-4 Shows minimal damage to wedge groove after cleaning.

83

84

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

second difference is that the third wedge assembly is Kevlar® wrapped, while the first, second, and third assembly are simply machined epoxy glass. Stator bar or coil looseness is one of the main reasons that tight stator wedges in the slots is so important. The resulting vibrations of the bars in the slot due to looseness can quickly wear the groundwall insulation on the bar or coil right through to the copper and cause a stator ground failure. Maximum instantaneous bar/coil bounce force per unit length of stator winding in the slot occurs when the top and bottom bars or coils in the same slot are in phase and carrying maximum stator current as shown in Equation (2.9): Total bar or coil leg bounce force, F total = F bottom + F top

(2.9)

where, F top =

3μo I 2 N m ws k2

F bottom =

μo I 2 N m ws k 2

Therefore, F total =

4μo I 2 N m ws k 2

μo = 4π × 10−7 H/m I = stator phase current in amperes ws = stator slot width in meters k = number of parallel stator circuits per phase This force is toward the bottom of the stator slot and is also sinusoidal in nature, due to the fact that it is proportional to the current squared. This means that the force is associated with the pole-pass forcing function and produces vibration at 120 (100) Hz, similar to the vibration forces on the stator core and frame. Since the magnetic field in the slot is highest near the top of the slot and diminishes toward the bottom of the slot, it can be shown that the resulting difference between the forces on the top and bottom bars or coil legs is substantial. In fact, the top bar or coil leg forces can be up to three times that of the bottom bar forces when both bars are in the same phase. The net effect for maximum bar bounce forces is described above. Wedging of the stator bars or coils, however, is not strictly concerned with just the bar or coil leg bouncing effects. Since there is considerable heat generated in a stator bar or coil, there are also thermal expansion, contraction, and insulation shrinkage issues to consider. Thermal expansion and contraction can easily loosen

2.9 ENDWINDING SUPPORT SYSTEMS

85

bars in the slot if they are not wedged properly, and the heat impact on the insulation systems can also be a factor if the insulation is not preshrunk (winding can lose a very small amount of mass as it continues to cure in service) prior to wedging. To elaborate on the preshrunk condition, when a winding is new, there is a possibility of some very minute amount of shrinkage of the insulation system due to the continued curing of the resins at specific temperatures depending on the manufacturer. It is important to realize this shrinkage is extremely small, so if the wedges are installed on the lower end of being tight, then this extra shrinkage could put the wedge into the loose category. It is more probable, however, that that the bar or coil has not been properly bottomed in the slot when wedged and the vibrations have assisted in this task and now the wedges are loose.

2.9 ENDWINDING SUPPORT SYSTEMS In addition to the slot, significant forces are present in the end regions of the stator winding as well. The endwinding geometry is also complex and requires a support structure that is flexible in certain modes and stiff in others, all at the same time, to restrain the endwinding under all modes of normal and abnormal operation. In addition, the strong electric fields in the end region require that nonconducting supports be used. Most support systems use blocks, tension devices, and rings, which together with the bars or coils themselves form a substantially rigid structure. Support in the radial direction is generally made to be very stiff, to keep vibration levels minimized. In the axial direction, it may be required that the endwinding structure be allowed to move axially to accommodate the thermal expansion of the slot section of the winding. Sudden phase-to-phase short circuits are the most significant transient behaviors in which high forces are developed in the stator winding. These must be accounted for in the design of the winding and in its support structures in the slot and the endwindings. Spacers, blocks, and wedges associated with the stator endwinding should be made of material that will not buckle, shrink, absorb moisture, or otherwise allow the windings to become loose and unsupported. All parts of the stator endwinding and associated connections and support structures should be designed so that they will be capable of withstanding full line-to-line and threephase short circuit at the generator terminals for 30 seconds as outlined in Ref. [5]. Vibration forces in the endwinding of some large hydro generators under normal load are also high and must be kept under control to ensure that there is no wear incurred on the endwinding as a consequence of rubbing or impacting. Thermal cycling and shrinkage effects can also promote advanced loosening and high vibration. The maximum vibration level of the endwindings and associated support structures, once the machine is installed and operating, should be less than 50 μm peak to peak [6]. This is unfiltered, with no natural resonances in the frequency ranges of 48–72 and 96–144 Hz for a 60 Hz system (40–60 and 80–120 Hz for a 50 Hz system). See Figures 2.9-1 and 2.9-2 for typical arrangements of endwinding supports.

86

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Figure 2.9-1 Stator endwinding support system for a bar winding.

Figure 2.9-2 Stator endwinding support system for a multi-turn coil.

2.10 STATOR WINDING CONFIGURATIONS Stator windings are designed to optimize the relationship between operating voltage and current-carrying ability. This goes back to the basic MVA relationship, which is a combination of the stator terminal voltage and the stator winding current. For the same level of MVA, as the terminal voltage of the winding is increased, the stator current required is reduced. The opposite is also true. As the terminal voltage is reduced, the stator current would have to be increased to

2.10 STATOR WINDING CONFIGURATIONS

87

keep the MVA rating constant. This relationship has significant consequences for generator design. For example the number of coils in series for a given core length will increase with voltage and the number of parallel circuits will increase for higher line current leading to the requirement for addition number of slots. These factors of core length and the number of slots is the basis for machine design. The product of these two is proportional to the MVA capability of the machine (see Figures 2.10-1 and 2.10-2).

A

B

Figure 2.10-1

A

Figure 2.10-2

C

Two parallel Y connected winding.

B

Four parallel Y connected winding.

C

88

CHAPTER 2

A

Figure 2.10-3

GENERATOR DESIGN AND CONSTRUCTION

B

C

Two parallel paths with long series jumpers.

If the connection is parallel, the terminal voltage tends to be lower and the stator current higher. For the same MVA rating, if the connection of the stator winding is in series, the terminal voltage will be higher and the current lower. The physical consequence of this is that the higher voltage machine requires a thicker groundwall insulation to withstand the higher voltage. For parallel connected winding, there would need to be a large amount of copper and increased cooling to accommodate the higher stator current. In another example, there are four parallel paths in the stator winding as shown in Figure 2.10-2. Figure 2.10-3 shows a Y-connection comprised of two of the parallels connected in series for a 720 slot machine. These configurations described above (and there are many more) allow flexibility in design to achieve a machine with a smaller overall size, lower cost, and lowest losses for best efficiency.

2.11 STATOR TERMINAL CONNECTIONS All generators require a means to deliver the power produced inside the machine, out to the main transformer, via an isolated phase bus (IPB) system, copper bus or cables. Since there are three phases in the generator, three-phase lead connections are required, commonly called stator terminal connections. These are used to make the connection from the stator winding inside the generator, out through the generator frame and casing, to the system. Each stator terminal carries the same current as the sum of the currents of all the parallels in a single phase. Since the terminals are at the rated voltage of the generator they need to be insulated, and generally, the same type of materials used for the stator winding insulation are used for the terminals as well (see Figure 2.11-1). In this photograph, the main leads are insulated with mica tape and epoxy resin, then painted with a beige protective paint which allows for easy cleaning, is essentially cosmetic in nature, and has almost no insulating capability. The high voltage terminals may also have a series of split phase

2.11 STATOR TERMINAL CONNECTIONS

89

Figure 2.11-1 Generator main leads – new installation with tags to show testing of the CT’s completed.

Figure 2.11-2

Shows split phase CT’s on main output leads.

current transformers in addition to the conventional CTs used for protection of the stator winding from turn to turn faults associated with multi-turn windings as shown in Figure 2.11-2. In this photo, the winding is a two parallel or 2Y connection and is painted with red protective paint similar to the beige paint. Each leg of the phase goes through the CT, thus the term “split phase,” and the CT monitors the current in each phase as they should be equal if the machine has no coils cut out and the airgap is perfectly balanced. In reality, the airgap is never perfect, and there will

90

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

always be some split phase reading on the winding. These numbers should be documented when the machine is new or the winding is new and monitored thereafter and trended. In addition to the high-voltage terminals, there are also three neutral terminals that make up the common connection point at the zero voltage or wye/star connection of the stator winding. Although these are essentially at zero or ground potential, they do carry the full stator current that the high-voltage connections carry and so must be given the same cooling as the high-voltage terminals. They are also insulated from ground, except at the actual connection or “star” point, to ensure no circulating currents or faults occur anywhere else in the winding system. This end of the stator winding (neutral end) will also have conventional CTs used for protection of the stator winding as shown in Figure 2.11-3. This arrangement can contain a number of redundant CTs for a “B” series protection if the “A” series were to fail. In protection systems nowadays, there are normally two redundant duplicate protection schemes in case one fails, these are “A” and “B.”

Figure 2.11-3 Generator neutral leads – new CT’s tested and ready.

2.12 ROTOR RIM

91

2.12 ROTOR RIM The rotor rim is visually a seemingly simple laminated steel structure that holds the field pole assemblies in place. The main function of the rotor rim is to house the field poles and to transfer the mechanical energy from the spider or drum to create a rotating magnetic DC field. When the machine is in service, the rim is subjected to various magnetic and mechanical forces that can cause complex behaviors while in service. There are a few main components of the rotor rim that will be discussed in detail later in the chapter but will be identified briefly now for reference: 1. Individual steel segments that together form the laminated rim (Figure 2.12-1) – these form the rim as they are piled in a circle around the spider or drum assembly. 2. Rim studs and nuts (Figure 2.12-8) – these are made of steel and hold the rim together by compressing the finished stack of steel segments. 3. Rim end plates (not all designs have this) (Figures 2.12-1 and 2.12-9) – used as a thicker steel segment at either end of the rim stack and works together with the studs and nuts to provide more uniform compression on the finished stack. 4. Rim keys (Figures 2.12-1 and 2.12-9) – these full or partial length keys transmit torque from the spider/drum assembly to the rim and provide the shrink interface for a rim that is designed to have shrink applied.

Rim keys

Field pole

Pole keys

Rim end plates

Laminated rim

Spider/drum assembly

Figure 2.12-1

Shows a modern day rotor. Source: Courtesy of Dr. Michael Znidarich.

92

CHAPTER 2

Figure 2.12-2

GENERATOR DESIGN AND CONSTRUCTION

Shows rotor from 1925 without the field poles installed.

5. Torque blocks (not all designs have this) (Figure 2.12-3) – these transmit the torque from the spider/drum assembly to the rim but do not provide a shrink interface and are partial length of the rim stack. Rotor rims have evolved over time from the late 1890s when the rim was actually an integral part of the cast spider assembly (see Figure 2.12-2) to being a separate laminated structure interfacing with the spider or drum via rim keys on a modern day machine as shown in Figures 2.12-1 and 2.12-9. Of course, depending on how large the machine is in diameter and the speed, integral type rims are still used today as they do prove economical in the right design circumstance. A rim and spider assembly that is integral behaves differently in service than a rim that has a laminated assembly constructed from steel segments (see Figure 2.12-8) and separate from the spider or drum assembly. When the integral rim is constructed the shape is set by the cast process, so the rim itself is circular and concentric when rotating providing the cast process was done accurately. The circularity or concentricity does not change when the machine is in service since it is part of the spider assembly and really has nowhere to move. The rim cannot expand independently from the spider on this type of arrangement. Supposing that the stator is circular as well, then the airgap for the machine remains stable and consistent all the way around the machine. The reason a more stable and consistent airgap is desirable will be discussed later in the book. The rim is normally constructed in a separate area of the powerhouse that has ample room for scaffolding and measuring devices to ensure the circularity, verticality, and concentricity is maintained in accordance with the design standard from the manufacturer. The construction procedure for assembling a rotor rim is proprietary to the manufacturer but a typical sequence is something like this. To begin, the bottom rim end plate segments as shown in Figure 2.12-9, or simply steel segments as shown in Figure 2.12-3, are installed onto the spider or drum assembly all the

2.12 ROTOR RIM

Rim support shelves

Torque block

Butting steel segments

Figure 2.12-3

93

Torque block with keys installed

The beginning of rim piling.

Large rim keys go here

Piling pin

Overlapping steel segments

Figure 2.12-4

Shows rim stack at an early stage of progression.

way around the circumference and are supported by stationary stands that can be adjusted for height. The height adjustment is critical in making sure the rim is erected as level as possible. The rim end plates are typically a very heavy construction that is much thicker than the steel segment. This is because in some designs, the rim end plates will be acting as the pressing plates when the bolts are tightened at the end of the assembly process instead of just the steel segments. Either design serves the same purpose and is equally effective. Once the initial circle is made, piling pins and adjustable thickness rim piling keys are installed, and the steel

94

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

segments are piled in a circle in a specific sequence, with overlap, see Figure 2.12-4. The sequence of overlap is a proprietary feature of the OEM and this is what will determine the operating behavior of the rim while in service. The rim piling keys are temporary and are typically a tapered set that are in place of the permanent rim keys. These temporary keys during construction provide a couple of important functions. First, they provide circular shape control for the steel segments when being piled and secondly provide concentric control of the piled rim stack. Both of these quantities are extremely important when the final assembly is complete so that the rim rotates as close to on center with the spider or drum assembly as possible. Further, both of these quantities can be adjusted as the piling is ongoing (by adjusting the tapered keys) and thus is checked after so many inches of piled rim is achieved, usually coinciding with the intermittent pressing operation. Once the rim reaches a specified height, the piling pins (same diameter as the rim bolts and shown in Figure 2.12-4), are removed and replaced with the rim bolts as shown in Figure 2.12-5. Depending on the size of the steel segments, more than one person is required to place the steel segments over the rim bolts. The tolerance of the rim bolts to the holes in the steel segments is very tight, thus, sand-filled mallets are used to strike and move the steel segment down the rim bolts to its final position. One person in the erection process will be going around the piled rim as each steel segment is installed and marking it with a grease pen to ensure the next segment goes in the proper location. After piling one steel segment after another, it can become confusing where the next segment should be installed in the sequence, so the person marking with the grease pen is a great way to check the process. At set points in the piling process, when a certain amount of rim height has been

Figure 2.12-5 in place.

Rim being piled showing the stacks of steel laminations and rim bolts finally

2.12 ROTOR RIM

Figure 2.12-6

95

Rim press operation during the piling process.

piled, pressing of the stack is done to encourage proper settling and compression and also serves as a checkpoint for various dimensional controls such as level, verticality, and stack thickness as shown in Figure 2.12-6. If the level or stack verticality is not within the required tolerances, now is the time to fix it. The more steel segments that are stacked, the more difficult it will be to adjust the verticality or level to within specified tolerances. At some critical point, it is no longer to possible to adjust these quantities and the rim will forever remain with these characteristics. The piling continues until the proper stack height is reached, then, the final pressing is completed. The proper amount of stack height is critical to ensure there is sufficient stack height to accommodate the tightening of the core bolts since they are of fixed length and require so many threads to be engaged once the nut is put on and torqued to the final value to give the desired rim stack pressure. For example, if one too few layers are installed, when it comes time to tighten the nuts on the rim bolts, the nuts will bottom out onto the non threaded portion of the rim bolt and no more compression can be achieved without adding more steel segments (if any spares are available) or some other arrangement. Conversely, if one too many layers are stacked, then there may not be sufficient thread engagement of the rim bolt when the nuts are torqued to the specified value. What is presented in the erection procedure is only one way of erecting a rim, there are many different ways this can be done, so consult the OEM for the procedure that was used for the rim in question. Now that the rim has been erected, it is appropriate to discuss in detail the main components of an assembled rim: • Individual steel segments that together form the laminated rim • Rim bolts and nuts

96

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

• Rim end plates (not all designs have this) • Rim keys • Torque blocks (not all designs have this) The steel segments are typically made of medium carbon steel typically 3.98 mm (0.157 ) to 6.35 mm (0.250 ) thick and are under considerable stress while in service and while at standstill depending on if the rim is a shrunk design or not, more on this shrunk idea later in the book. Each steel segment is manufactured using a die and punch press or may be laser cut depending on the economics, quantity, and delivery time required see Figure 2.12-5. The steel segment as mentioned above must accommodate a number of different components in order for the entire structure to operate properly when in service. Also, the constructed rim may have two or more “donuts” or separate sections stacked on top of each other when completed depending on the height required by the design as shown in Figure 2.12-7. This picture was taken from outside the rotor with the field poles removed. This particular rim design has three distinct sections stacked on top of one another. There is no mechanical mechanism other than the mass in the axial direction that holds the sections together. The rim keys tie the sections together from a tangential movement point of view.

Figure 2.12-7 rim donuts.

Shows rotor

2.12 ROTOR RIM

97

The rim bolts and nuts vary in diameter depending on the size and speed of the machine but are typically made from a higher strength steel since they are tensioned when the rim is completed. The bolts are machined to a very precise tolerance and the steel segment holes are punched or laser cut to a similar tight tolerance. The typical clearance for the bolt and the hole in the steel segment is approximately 0.076–0.127 mm (0.003–0.005 ). Piling these steel segments that can weight 100 lb or more in a special sequence onto the rim bolts to make a laminated structure is no easy task. As previously mentioned, not all rim designs have a rim end plate as shown in Figure 2.12-8; the steel segments are used instead. The rim end plates, if so equipped (Figure 2.12-9), are typically made from medium carbon steel and are much thicker than the steel segments, but are the same pattern. They vary in thickness depending on the machine diameter, speed, and compression required in the finished rim assembly. A typical pressure for a rim when compressed is 500 PSI but can vary depending on the size and speed of the generator. The rim keys are the critical interface between the rim itself and the rotor spider or drum assembly. Manufactures will use different grades of steel for rim keys depending on their rim to spider fit design. The key sits in the keyway that is punched or laser cut for the laminated steel segment and the keyway that is machined for the spider arm or drum assembly as shown in Figures 2.12-10 and 2.12-11. There are many different rim key designs, and we will touch on a few examples here. In Figure 2.12-10, the rim key is a solid rectangular piece that is driven

No end plate just steel segments

Figure 2.12-8

Rim bolts and nuts

Fully piled rim with no end plate – just steel segments on top.

98

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Rim key

Spider post

Top and bottom rim end plates

Spider arm

Steel segments Spider hub

Figure 2.12-9

Shows another rim design with end plates.

Rim end plate Spider post Rim key

Steel segments

Figure 2.12-10 Rim key in final position.

into the keyway when the rim (steel segment) are heated sufficiently to allow the key to be inserted. When the rim is cooled down to ambient temperature, an interference fit is established, known as a “shrink” fit. This design also allows for a shim, a thin piece of steel stock, to make up the difference in interference fit if needed.

2.12 ROTOR RIM

99

Black reference line

Figure 2.12-11 Tapered rim keys set in final position by hydraulic jack.

Shims are used when the key itself has not been machined to the exact radial dimension needed for the final interference fit. It is desirable to have a single piece key to keep things very simple unless the shim is being used as a liner or shrink dimension on the rim (steel segment) side in order to provide a smooth surface for the key to ride on when being inserted. As previously discussed, this interference fit, depending on the OEM, is maintained even during load rejection speeds. In Figure 2.12-11, a tapered key system is being used to make up the interference fit required for the designed shrink to be applied. These keys (one stationary and one drive key) are machined with graduated slope in mils/in or μm/m, so when the key is driven down or pulled up, a radial displacement in terms of key thickness can be calculated. In this scenario, there are two tapered keys, the stationary key on the rim side (just the tip visible in the picture) and the drive key which has the hydraulic jack underneath to pull the key outwards a little bit to achieve the correct fit. The drive key is normally installed using a plastic sledge hammer to push the key inwards for a proper fit. When the rim is sufficiently heated, the drive key is inserted until the black reference line is the same level as the stationary key. If the key is driven too far down, since a mallet and human force is being used, a hydraulic jack can be used to reverse the key insertion slightly. This system is advantageous and simple as one can calculate how much the drive key must be adjusted to have the interference fit required based on the taper. Referring back to Figures 2.12-3 and 2.12-4, for this particular rotor design, there are rim keys which we have discussed but also a torque keys as well. The torque keys on this particular design are what transmit the torque from the shaft to the rest of the drum assembly. These torque keys are not full axial length like the rim keys; they are smaller in length and do not have an interference fit. The keys

100

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Figure 2.12-12 Torque key installation in progress.

are driven in with a hammer until refusal and locked into position by welding. A torque key installation in progress with the final key to be hammered into position is shown in Figure 2.12-12. In some designs, the rim keys take the torque from the spider or drum and transmit it to the rim, or a combination thereof. The rim keys as previously mentioned control the circularity and concentricity of the rim as well as take the shrink forces (if the rim is a shrunk design) from the rim and transfer it to the spider or drum assembly. If the rim is a shrunk design, the spider/drum, rim key, and rim are always in contact even while the machine is at rated speed or even higher such as load rejection speeds. If the rim is a floating design or nonshrunk, the rim keys or torque keys are in place to transmit the torque from the spider/drum to the rim assembly and to maintain rim-to-spider concentricity. These keys are typically driven to refusal during the installation process at ambient temperatures. The radial and tangential tolerances on the key assembly are very tight, in the neighbourhood of (0.050 mm) 0.002 . The concentricity and circularity is now a function of the rims sole ability to maintain rigidity and shape with respect to the airgap while in service. Since, in this design, the spider/drum, rim keys and torque keys, and rim are not in contact when the machine is at speeds other than stand still, airgap uniformity is of paramount importance. Nonshrunk rims are typically used when the normal airgap magnetic forces are not sufficiently high to change the circular shape of a sufficiently mechanically stiff laminated rim. Normal assumes a uniform airgap and, of course, some deviation from perfection as no airgap is 100% uniform. Should this airgap become compromised to the point where the magnetic forces exceed the rim mechanical stiffness, the magnetic forces can distort the shape of the rim where it is in line with the critically small airgap on a once per revolution frequency. The nonshrunk rim will slide on the spider ledge as it is pulled by the narrow airgap’s high magnetic force and then will

2.12 ROTOR RIM

101

slide back into its original position when the airgap is sufficiently large again. This type of activity will cause potentially severe fretting on rim support and key components leading to mechanical vibration of the stator and rotor assemblies as the rim loses circularity and concentricity. The fretting of the rim support contact surface can cause failure of the support structure which represents a serious risk to the safe operation of the generator. On a shrunk rim design, compromised airgaps can also cause the rotor assembly as a whole to migrate toward the smallest airgap and can also cause vibrations depending on the circularity and concentricity of the rim. Since the rim is shrunk onto the spider/drum assembly during normal operating speeds (may float during an overspeed event), it will take more force to pull the entire rotor assembly, which is restrained by the guide bearings, toward the stator as opposed to the unshrunk design where the rim can move independently of the spider/drum assembly. As mentioned previously, there are many different designs for the rim-to-spider/ drum interface, too many to discuss in this book. The important thing to recognize is if the machine has a floating or shrunk rim design, consultation with the OEM may be required if there is any uncertainty. When the rotor rim is first assembled and compressed, the frictional forces between the steel “lamination” segments in some rim designs is not enough to prevent segmental movement in the radial direction for the first time the machine achieves an overspeed condition such as a load rejection. This depends on the machine diameter, speed, lamination thickness, and mass of the rim. The design of the rim can take into account that the steel segments are allowed to move and become tight up against the rim bolts, this is also known as “rim slip.” In this condition, the steel segments move in the radial direction outwards toward the airgap and the clearance between the steel bolts and the steel segment holes is taken up. In other words, the steel segments are butt up against the steel bolts. The circumference of the rim is now slightly larger than the original thus making the airgap slightly smaller. The formula to calculate the amount of radial airgap reduction once slip occurs is shown in Equation (2.10): AGAS = AGBFS − LCTRB × RSPC

2π

(2.10)

where, AGAS = Airgap after slip – which is the calculated airgap after slip AGBFS = Airgap before first spin of the machine – this is the airgap after the rim has been piled and before the rotor has been spun for the first time – this can be measured at site during construction or on an OEM drawing LCTRB = Lamination clearance to rim bolts – the clearance on the rim lamination can be found on the rim segment drawing and the bolt size can also be found on a drawing or both dimensions can be acquired from the OEM RSPC = is the number of rim segments per circle – which is the number of rim segments it takes to complete one circle of the rim. This information should be on the rim segment drawing or can be counted at site or can be acquired from the OEM.

102

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Let us take an example of a 140 MVA machine with an airgap of 0.796 (20.21 mm), 12 rim segments per circle with a clearance of 0.005 (0.127 mm) between the lamination and the rim bolts. Using Equation (2.10) yields the following result: AGAS = 0 796 − 0 005 × 12

6 283 = 0 786

19 96 mm

Here, we can see that the radial airgap reduction with the machine at standstill would be 0.010 (0.25 mm). This may seem like a small amount for a machine that has this large of an airgap, but the smaller the airgap at the beginning, the more this slip reduction will impact the final airgap. This exercise is simply to make the reader aware that if rim slip occurs, this is the magnitude of the reduction given the parameters listed above. Notice that the more the clearance between the rim lamination and the rim bolts and the more segments per circle, the more rim slip will affect the final value. Once this “slip” has occurred, it is not reversible, and the speed at which this occurs is design-specific. Not all rims slip as described, some OEMs claim their designs are not subjected to this event, thus a conversation with the OEM may be in order. This brings up another question about the airgap when the machine is assembled versus after the machine has experienced this “rim slip.” It is now appropriate to discuss the airgap and the different types of airgaps that may be listed on a drawing, operating manual, or in a proposal from the OEM. From experience, there are at least three designations for airgap: Erected airgap – this is the airgap that exists when the machine is first assembled before it spins for the first time. This value may be present on an assembly drawing from the OEM as construction is ongoing at site. Design airgap – this is the airgap that exists after the machine has been spun for the first time having experienced a load rejection. A load rejection is at some speed higher, typically between 25 and 40% above rated speed. This allows the rim steel segments (if they are going to move) to settle into a final position radially, the concept of rim slip. It is important to recognize that if the machine reached a runaway speed condition, that is typically twice rated speed, the rim segments may settle a bit more if the load rejection speed did not butt the steel laminations against the steel bolts. The design airgap value is likely on a drawing or operating manual that is issued to the customer. For the laminated rim designs that do not suffer rim slip (friction rim) because they have sufficient friction between laminations to remain stable for all operating modes including overspeed, the erected airgap and design airgap may be the same. Running airgap – this is the theoretical airgap that is present dynamically while the machine is in service running at rated load and temperature. The only way to measure this is with a dynamic airgap monitoring system. This theoretical calculated value is not normally given to the customer as it is an internal number for the OEM for design purposes. This airgap will be the result of centrifugal forces expanding the rim, temperature of the machine, and the expansion difference between the rotor and stator structures.

2.13 ROTOR SPIDER/DRUM

103

These three airgap types are invaluable when trying to understand drawings, dynamic operation, and discussions about the machine with the OEM. Discuss with the OEM which airgap is being referenced on the drawing as each OEM may have a different interpretation of the definitions as given here.

2.13 ROTOR SPIDER/DRUM These components can be some of the most complex and tricky to understand depending on the design and manufacturer. There are many types of spiders and drums in service today, and it is not possible to touch on each and every unique design, instead a more general description of a rotor spider/drum design and construction and its intended function will be presented. For a more in depth understanding of a specific design, the reader is encouraged to consult with the drawings that came with the machine (if any) and the OEM. The main purpose of the spider/drum is to transfer the torque from the main shaft to the rotor rim assembly. Many spider/drums made in the early 1900s made use of cast technology as shown in Figure 2.12-2, and as technology improved they were made from fabricated steel components that were welded and/or bolted together as shown in Figure 2.12-1. Some spiders even incorporated both a cast and fabricated steel design together as shown in Figure 2.13-1. The spider is composed of a hub and arms that act as a single unit to transmit the mechanical energy to the rim. Spider arms take on many shapes and sizes depending on how many arms there are, rim weight that needs to be supported, speed, diameter of the spider, and whether or not the rim is floating or a shrunk design and the amount of shrink that is applied.

Cast

Fabricated

Figure 2.13-1

Shows a cast and fabricated spider assembly.

104

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Machines with cast spiders have the advantage that the spider and hub assembly and sometimes even the rim (cast as well, no steel segments as shown in Figure 2.12-2) are all one integral piece. There are no welds to worry about, and if the original casting was done properly with no inherent defects, there is little that can go wrong if the machine is operated and maintained within the design nameplate rating. A fabricated spider, similar to the one shown in Figure 2.13-1 (except it is all fabricated steel with no cast pieces), have many more considerations in the way they are designed and constructed. Starting with the hub it can be constructed as a single cylinder section or many sections welded together to make one cylinder. The hub, when finally completed, will ultimately have to carry the entire weight of the spider arms, rotor rim, field poles, and field winding. The hub will have to endure the high compressive forces of a shrunk rim without collapsing at standstill as well as the torsional forces when the machine is stopped and started. The welds that make up the hub assembly will also have to endure a portion if not all of these forces, depending on the hub design. The spider arms are attached to the hub by heavy welds. Anytime a weld is used to attach one piece of metal to another there are inherently going to be stresses developed in the pieces that are being joined. These stresses must be managed by stress relieving techniques or by the welding procedure so that stress cracks do not develop while the machine is in service and undergoing cyclical loading. The spider arms must support the weight of the rotor rim, field poles, and field winding on a small shelf at the bottom of each arm and transfer this weight to the hub assembly. The spider arms will also have to endure the high compressive forces from a shrunk rim design without buckling at standstill. It is very important not to overcompress the arms during the shrink process as this yielding is irreversible. This activity is better left to the OEM if there is any uncertainty of the shrink value or procedure to apply the shrink to the machine. A drum assembly is typically made up of an upper and lower steel disk separated and held together with contoured steel webbings or vanes that are welded in place. Ultimately, the vanes form part of the powerful fan assembly for the rotor which will circulate air inside the generator. The lower disk couples with the generator shaft and transfers all or a portion of the torque to the rim via the torque blocks and/or rim keys in cooperation with the top disk if so designed as shown in Figure 2.13-2. In this design, the rim keys transfer the torque and absorb the shrink forces. Depending on the manufacturer, the lower disk may transfer 80% of the torque and the upper disk may transfer 20% of the torque or some other percentage variation, depending on the design. The lower disk also has the rim shelf that will accommodate the steel segments of the assembled rim. This rim support shelf area in contact with the steal segments or bottom end plate varies greatly from machine to machine. For example, on the drum design as shown in Figure 2.13-3, the area seems larger than the rim support shelf area at the end of the spider arm as shown in Figure 2.13-4. On the other hand, looking at the support shelf on the drum design in Figure 2.12-3, the area seems closer to the spider arm design. It all depends on the diameter of the rotor, weight of the rim, speed, and if the rim is a shrunk or floating design. As with the spider design, if the rim is shrunk, the drum will have to endure

2.13 ROTOR SPIDER/DRUM

Figure 2.13-2

105

Shows spider drum design. Source: Courtesy of Dr. Michael Znidarich.

Steel segments

Support shelves

Figure 2.13-3

Rotor drum support shelves.

high compressive forces to accommodate the rim shrink. Depending on how large these forces are, it will drive the designer to re-enforce the drum components to accommodate the additional loading. In all cases, it is more economical from a spider and drum point of view to have a floating rim since the components do not have

106

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Bottom rim end plate

Spider arm support shelf

Figure 2.13-4

Rotor spider support shelf.

to be as robust in order to accommodate the additional forces from the rim shrink. The rim keys on a full floating rim may be driven cold by a sledge or pneumatic hammer, so there is little compression between the rim and spider/drum assembly at standstill, and there is no heating of the rim during installation of keys on this style. Other floating designs may incorporate some rim heating during key installation to give a rim float at some speed lower than synchronous. It is important to keep in mind that later in the operating life of the machine, the airgap as previously mentioned will no longer be as uniform as when first assembled (assuming the OEM has erected the machine properly) and a floating rim will be more susceptible to the unbalanced airgap and vibrational issues may arise. Keep in mind that a floating rim is more sensitive to eccentricities in the airgap than a shrunk design. In Chapter 9 of the book, the consequences of a loose rim on the support shelf structure will be discussed in detail.

2.14 ROTOR POLE BODY The rotor pole body is made up of many components and its main purpose is to house and keep the copper field winding in place as well as to provide a DC flux path that is created in the rotor field as shown in Figure 2.14-1 [7]. The main components of the rotor pole body are • Punchings • End plates • Through bolts and nuts or rivets • Amortisseur winding • Amortisseur shorting plate

2.14 ROTOR POLE BODY

107

Figure 2.14-1 Example flux distribution through pole body. Source: Courtesy of H.C. Karmaker.

Through bolts and nuts

Amortisseur bar (winding)

Punchings

End plate

Amortisseur shorting plate

Figure 2.14-2 Shows parts of field pole being assembled. Source: Courtesy of Dr. Michael Znidarich.

These components are illustrated in Figure 2.14-2. The pole punching is the main component of the pole body assembly. It is the piece that holds the copper field winding from moving out in the radial direction while in operation and houses the amortisseur bars, ground insulation, and the through bolts which hold the entire

108

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Figure 2.14-3 Shows finished rotor pole end plate with the “L” shaped portion. Source: Courtesy of RPR Hydro.

pole body assembly together. The pole punching is typically made of medium carbon steel (0.3–0.8% carbon content) that has no insulation except oxidation on either side of the punching and is typically 1.5 mm (0.060 ) thick. In some cases, since there are a large number of pieces required to make up the pole body assemblies for one generator, the pieces are made using a die and punch press. In other cases, if the quantity required is sufficiently low, laser cutting may be used to produce the pole lamination, but this is generally a more expensive process than die and punch, however, in recent years, the price has come down significantly. The end plate, as appropriately named, is installed on both ends of the pole body assembly. The end plate is typically made from forged or cast steel depending on the speed of the machine and the size of the rotor pole body. The “L” shaped part of the end plate is under quite a bit of stress from centrifugal forces since it is at the outer radius of the rotor assembly, see Figure 2.14-3. Its purpose is to provide a pressing surface for the copper at each end in conjunction with the through bolts and nuts. The end plates help hold the copper field winding in place in the radial direction to prevent distortion during operation. Higher speed machines may have an interpole wedge style brace Vee blocks to prevent the copper winding from distorting into the space between poles during operation. Depending on the design and size of the field poles and the speed the machine achieves during normal operation, load rejection, or even runaway, the copper winding wants to occupy the interpolar space which is more tangential in direction than radial. The wedge prevents this from happening (see Figures 2.14-4 and 2.14-5). The wedge design typically has minimal centrifugal loading by utilizing high-strength aluminum alloys to make them as light as

2.14 ROTOR POLE BODY

109

Cast aluminum alloy bracket Holding down bolt

Epoxy glass laminate insulation

Rectangular slot punched in rim plates

Figure 2.14-4

Concept of interpolar wedge. Source: Courtesy of Dr. Michael Znidarich.

Figure 2.14-5

Example of an interpolar wedge. Source: Courtesy of Dr. Michael Znidarich.

possible. Another important feature is the minimal obstruction to the cooling path along the coil side, allowing maximum air passage in this area. Normally, the wedge is fitted in the center of the pole side while larger pole lengths will have 2 or more spaced for equal loading.

110

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

The through bolts as previously mentioned hold the entire pole body assembly together. They are typically made of a higher strength, more durable steel than the pole punching since these bolts are under tension when the pole body is completed. There are a many different ways the through bolts are secured to the pole body. One modern and common method is to apply nuts on each end of the through bolt and then tighten and weld the nuts to the end plate. The construction of the pole body assembly is proprietary to the manufacturer but a typical example follows. One end plate and pole punchings are stacked on a special horizontal table, positioned with a jig to get the proper alignment for the through bolts and amortisseur bars, and pressed every so many feet of stack length to ensure proper compression. Once the final stack length is reached, the stack is compressed to a specified value to properly seat all the pole punchings before the amortisseur bars are installed and brazed to the shorting plate. A typical pressure the stack is compressed to is 580 PSI (4 MPA) or higher depending on the design of the pole. The other end plate is installed and the through bolts are secured either by swaging or with nuts that are tightened and welded. The pole body is now ready to be insulated and accept the copper field winding. Laminar insulation is not required since the pole body is experiencing a DC field while in synchronous operation and not an AC field. The only location where the pole body will experience an AC field (cross slot leakage flux or tooth ripple flux) during synchronous operation is at the pole-face since typical hydro generator airgap dimensions are small relative to the stator slot pitch. With regards to the cross slot leakage flux, one of the issues a designer must account for when selecting the size of the generator airgap is the width of the stator slots. During operation, there is slot-to-slot leakage flux in the stator and it is important that this leakage flux linkage with surface of the rotor field poles be kept to a minimum so as not to cause additional and potentially excessive pole-face losses (this will all depend how close the pole-face gets to the cross slot leakage flux). The previous statement goes back to ensuring the airgap on the machine remains as close as possible to design parameters. In older and smaller MVA machines where the design airgap is very small, the stator will have a closed slot or nearly closed slot to keep this interaction to a minimum. The amortisseur winding and shorting plate are described later in this chapter and the reader is referred to Section 2.16.

2.15 ROTOR WINDING AND INSULATION The rotor winding and insulation make up the main electrical portion of the pole assembly. On salient pole rotors, there are several types of rotor coils. One type is a wire wound coil, which are made from rectangular film insulated wire and then wrapped around the insulated pole body. The other two more commonly used types are the edge bent copper coil or the fabricated brazed joint copper coil. These edge bent

2.15 ROTOR WINDING AND INSULATION

111

coils are made with large rectangular continuous copper strap which is edge bent as shown in Figure 2.15-1. In this manufacturing process, long pieces of copper strap are coiled using a special machine. When the machine gets to the end and needs to make a turn, the copper is bent around the edges thus giving the profile in the figure (thicker at the small radius against the pole body and thinner at the outer radius). Brazed joints in the copper coil are made along the straight portion if needed when the long piece of copper strap runs out. The coil shown in Figure 2.15-1 may only have two or three brazed joints in the entire assembly along the straight portion. The only real consequence to edge bent is the thinning of the turns at the ends of each coil along the outer edges as shown in Figure 2.15-2. The effect shown in this figure is an extreme case of what it can look like, most edge bending results in some variation less that what is presented. The other method is to braze pieces of copper strap (butt or interlocking joint) to form the coil as shown in Figures 2.15-3 and 2.15-4. In this example, the interlocking joint is first subjected to a very high clamping pressure to put the pieces together and then brazed. The difference is the fabricated joint is more uniform in coil thickness at the ends as shown but more brazed joints are required. A butt braze (copper segments are brazed perpendicular to each other to form the coil) has similar features of uniformity of coil thickness and more brazed joints like the interlocking. Experience has shown that all methods are equally effective, and there should be no discernable difference in reliability provided the manufacturing has been done correctly. During a reinsulation is a good time to check the copper sections for any type of cracking or deformity. The brazing process can be done using a torch or automated process using special machinery. In either case, a common brazing material called Silfos® which is a

Thicker copper here

Thinner copper here

Figure 2.15-1

Copper strap edge bent coils.

Edge bent portion of copper strap

112

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Gaps due to edge bending and thinning of copper

Figure 2.15-2

Edge bend consequence of gaps between copper segments.

Jigsaw joint

Finned turns for cooling Jigsaw joint

Figure 2.15-3 Assembled copper pieces at the jigsaw (interlocking) joint similar to a puzzle. Source: Courtesy of Dr. Michael Znidarich.

copper alloy containing copper, silver, and phosphorous can be used to fuse the copper segments together under high heat. Strap wound coils need turn insulation inserted between the copper segments prior to the consolidation of the coils. Some older machines, pre-1970s had turn insulation consisting of an asbestos paper in sheet or tape form bonded with shellac

2.15 ROTOR WINDING AND INSULATION

113

Uniform copper thickness

Brazed jigsaw (interlocking) joint

Figure 2.15-4 Brazed jig saw (interlocking) sections and uniform copper. Source: Courtesy of RPR Hydro.

and heat press cured. The asbestos paper was thicker (more like a sponge consistency) and provided a nice filler for the edge bent uniformity problem, hence the gaps as shown in Figure 2.15-2. The insulation shown in this figure is a thin Nomex®, so the copper thinning is much more evident. If desired, these gaps are easily mitigated with an insulating filler to prevent ingress of moisture and contaminants as well as for esthetic purposes. Adding more insulation between turns is not economical nor will it likely completely eliminate the gaps at the edges of the coils. As mentioned, modern strap wound coil designs use Nomex® or similar materials since their thermal, mechanical, and electrical properties are more than adequate for this application. Thin strips of this insulating material are placed between turns and epoxy is used to consolidate everything in place under a heat cure process. Since there are large rotational mechanical forces acting on the coils,

114

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

particularly during overspeed and runaway conditions, it is essential that they are well bonded into a rigid structure. The ground insulation is between the copper coil and the pole body assembly and can be composed of many materials depending on the vintage of the machine. Some such materials are asbestos, epoxy-bonded glass fibers, micafolium, and Nomex® [8]. This insulation can be applied in several different ways depending on the manufacturer and vintage of the machine. In the early days of field pole manufacturing, the ground insulation was applied to the copper winding along the inside of the coil and then placed over the pole body assembly which may or may not have had a collar around the pole body. The other more traditional way is to wrap the pole body material with ground insulation such as an appropriate grade of Nomex® or equivalent and installing an insulating collar made from synthetic resin-bonded glass or Daglas® backed mica to insulate between the copper winding and the pole body tips as well as increase leakage distance from the copper to the pole body tips [8]. Figure 2.15-5 shows the bottom collar installed during the field pole refurbishment. These collars are preferably made in once piece, but for larger poles, a pinned lapped joint may be necessary. This insulating collar is normally sealed to the pole body insulation with silicone to isolate the copper winding from the grounded pole and prevent contamination ingress. Once the coil assembly is placed onto the pole body, it is desirable that the coil does not move while the machine is in service as this would cause abrasion of the pole body insulation. To prevent this, the manufacturer would secure the coil onto the pole body using wedges and a gluing compound to secure the wedges in place as shown in Figure 2.15-6. It is important to realize here that bonding of the

Pole body

Nomex groundwall insulation

Silicone sealant

Bottom insulating collar

Figure 2.15-5

Shows groundwall insulation, insulating collar, silicone sealant.

2.15 ROTOR WINDING AND INSULATION

Figure 2.15-6

115

Shows wedges on pole body to secure copper coil.

coil to the pole body assembly is done to varying degrees depending on the coil-tocoil connections used on the field winding. If the coil-to-coil connection is a bolted and soldered connection (consolidated with no flexibility), then the bonding to the pole body must not be so rigid, allowing for a finite amount of flexibility as the connection is stressed in operation. If the connection is the flexible type, then the manufacturer may choose to more solidly bond the copper to the pole body by using more wedges and even fill the void between the copper and pole body with epoxy. Consult the manufacturer whenever a reinsulation of the pole body assembly is required to determine which system best suits the machine. Finally, a top collar (the one closest to the rim) may also be used to ensure adequate creepage distance to the pole body/rim assembly, but not all designs incorporate the top collar. Silicone is used to seal the gap between the coil collar and pole body assembly to prevent contamination ingress, see completed pole in Figure 2.15-7. The outside surfaces of the copper coil may be left bare to provide the most efficient cooling. Sometimes the copper is painted with an insulating paint for protection and ease of cleaning. Lastly, some designs of copper strap coil utilize a high-low approach or finned turns in order to improve heat transfer into the ventilating air (see Figure 2.15-3).

116

CHAPTER 2

Figure 2.15-7

GENERATOR DESIGN AND CONSTRUCTION

Finished field pole with red insulating varnish applied.

Some very large machine designs have springs inside the rim assembly that push against the inner pole collar to keep it radially in place. On some generators, the rotor coils are designed to carry the current in one direction on one pole, and in the other direction on the next pole. These are called open or crossed coils. On other generators, all the rotor coils are wound in the same direction and special connections are installed on top or bottom of the rotor rim to carry the current in one direction in one pole and in the other direction in the next pole.

2.16 AMORTISSEUR WINDING Most rotor field poles employ a damper (also called amortisseur or damping) winding to dampen torsional oscillations and provide a path for induced currents to flow. The amortisseur winding is essentially a separate winding installed under the face of the pole body that is connected in a way similar to the squirrel-cage of an induction motor. The winding is typically made from a tough pitch copper, brass, Everdur (copper 95%, silicon 4%, manganese 1%), aluminum, or iron. The amortisseur winding is typically buried in the pole body steel and is not always visible over its length (see Figures 2.16-1 and 2.16-2). It produces an opposing torque when currents flow in it and this helps dampen torsional oscillations and add to the stability of the rotor during system excursions from normal operating conditions. Negative sequence currents in the stator winding will also cause the amortisseur bar to be active. The limit for negative sequence currents while in operation

2.16 AMORTISSEUR WINDING

Figure 2.16-1

117

Amortisseur bar visible through the pole punching.

Figure 2.16-2 Pole under construction with the amortisseur shorting bar shown – amortisseur bar (not installed yet) will not be visible under pole punching when installed, see also Figure 2.14-2 for where the amortisseur bars reside.

is 5% of the rated current for nonconnected and 10% amortisseur windings with interpole connections, as outlined in the Ref. [5]. On most machines, the amortisseur winding segments are only connected to each other in one pole in a noninterconnected fashion. In special machines, such as

118

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

a pump generator, the amortisseur winding segments are interconnected with the adjacent pole, forming a “squirrel cage.” This is due to the large starting currents the amortisseur winding will be exposed to when the rotor is started from a standstill by applying system voltage to the stator, when a starting motor is not used. Some manufacturers have a design philosophy where the amortisseur is interconnected in all machines that are designed due to better damping performance (subtransient reactance X d ) for little cost difference. During manufacture, the amortisseur bars are installed into the pole body such that there is intermittent contact between the pole body and the amortisseur bar itself through the entire length. This helps ensure that the bar does not come loose while in service. The bars are then joined at the ends of each pole by brazing them together using a single copper plate or multiple layers of copper plate as shown in Figure 2.16-3. These plate assemblies offer some flexibility since when the bars are active in service, the current flowing through each bar may not be equal and thus axial thermal expansion may be different from bar to bar. If the brazed connection in combination with the copper plate(s) does not offer the correct amount of flexibility, cracking of the brazed connection or plate(s) will result affecting the performance of the amortisseur circuit [7]. It is very important that proper periodic inspections are done on these connections to ensure their integrity. Figure 2.16-1 is an example of a less flexible arrangement for axial expansion for the amortisseur connections. There are many different connections that can be used to connect one set of amortisseur bars to the adjacent set on the next pole. A flexible connection is normally used in order to accommodate expansion and contraction between shorting bars from operating stresses. Typical connections may include leaf copper, solid omega shaped, and flexible braid style as shown in Figure 2.16-3. These flexible connections may be solidly bolted, bolted and brazed, or just brazed depending on the manufacturer. Other solid designs as shown in Figure 2.16-4 can be solidly

Figure 2.16-3

Flexible connection between amortisseur circuits on adjacent poles.

2.17 SLIP/COLLECTOR RINGS AND BRUSH GEAR

Figure 2.16-4

119

Shows solid amortisseur connection between adjacent poles.

bolted, bolted and brazed or just brazed as well. Inspection should include ensuring the bolts have not worked themselves loose and that there is no cracking or fraying anywhere on the connector. Signs of overheating on the connector could be an indicator of a poor or a high resistance connection. It is generally understood that the flexible braided connections can be more prone to filament breakage due to thermal and stop/start cycling and centrifugal forces at operating speeds. Make sure that this type of connection has been authorized by the OEM and that the OEM has done an analysis of the conditions in which this connection must function.

2.17 SLIP/COLLECTOR RINGS AND BRUSH GEAR A DC current is supplied to the rotor winding to create the rotating magnetic field. This can be done by a brushless excitation system as shown in Figure 2.17-1 and schematically as shown in Figure 2.17-2 or by a set of positive and negative slip or

120

CHAPTER 2

Figure 2.17-1

GENERATOR DESIGN AND CONSTRUCTION

Brushless excitation.

EXC 3~

Figure 2.17-2 of Voith.

Traditional rotating brushless exciter with diodes. Source: Courtesy

collector rings as shown in Figures 2.17-3 and 2.17-4. For the slip/collector ring type of current delivery system, the rings are supported by an insulating block which is generally made of epoxy glass or some other insulating system. The rings are typically made of mild steel, but other materials such as brass and copper alloys were also used. Each ring is opposite in polarity to the other as one conducts current into the rotor winding and the other collector ring brings it back out. The current transfer to the rings takes place using a sliding contact surface by carbon-loaded brushes that slide along the rotating surface of the rings as the rotor spins. The brushes in more modern systems utilize a constant pressure spring to maintain a consistent pressure against the ring surface during operation as shown

2.17 SLIP/COLLECTOR RINGS AND BRUSH GEAR

Brush and brush holder

121

Constant pressure spring

Slip ring

Figure 2.17-3

Shows modern style constant pressure spring.

Adjustable pressure spring using box rachet

Figure 2.17-4

Slipring assembly with old style adjustable spring.

in Figure 2.17-3. In older designs, the spring pressure must be adjusted every so often since the spring pressure is dependent on the setting on the brush box as shown in Figure 2.17-4. This type of system requires much more observation and adjustment than the constant pressure type. Good contact is difficult to achieve if the surface of the rings and brushes is not properly prepared when installed.

122

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

The friction between the ring and brush surfaces and the I2R across the brushcollector contact resistance generates heat. To theoretically maintain reasonable current sharing between multiple brushes on some rings there are helical grooves cut into the ring surface to wipe the brush surface in operation. The rings themselves may be machined with a slight radial offset or the rings themselves may be mounted slightly off-center so as to move the brush in and out slightly to ensure the dust from the brush does not jam the brush in the holder. It is very important that the brushes move freely and unobstructed inside the brush box assembly. Failure to maintain this freedom of movement can lead to many problems with the sliprings including unequal loading of brushes and excessive sparking and eventually a ring/ brush gear failure. More discussion on the brush boxes, brushes, clearances to the ring, and so forth, later in the book.

2.18 COOLING AIR Most hydro generators are cooled with air, either directly by passing air from the inside or outside the powerhouse into the machine or by passing air cooled by an air to water heat exchanger into the machine. An example ventilation diagram for a specific machine is shown in Figure 2.18-1. There are many variations of ventilation diagrams specific to the OEM and a particular machine design. The OEM should be able to provide the ventilation diagram upon request. There are a few types of ventilation schemes, the first is to draw air from outside the powerhouse or within the powerhouse directly, and the second is to have a totally enclosed ventilation system. In the first system, powerhouse ambient air or air from outside of the powerhouse is circulated by the rotor fan assembly into the machine, through the core, over the endwindings and out the back end of the machine into the powerhouse or back outside again. Taking air from outside the powerhouse usually results in a large amount of dust and debris along with humid air being introduced into the generator. Design standards typically state that the air being circulated into the machine should not exceed 40 C. Powerhouse ventilation may be required if the ambient air temperature is too high for generator operation. In the second system (TEWAC, Total Enclose Water to Air Cooling), the generator is enclosed by some sort of air tight enclosure so the air within it is completely separate from the powerhouse air and the two do not mix unless louvers are purposely opened for this reason. The air inside the enclosure is circulated by the rotor fan and pushed through the core and out of the back end of the core where the air is directed by baffles in the stator frame to the air to water heat exchangers. These heat exchangers are typically designed to take generator hot air and reduce the temperature to no more than 40 C and recirculated back into the generator. The amount of air circulated inside a machine depends on the design requirements. The manufacturer will calculate the losses of the machine and based on that determine how many cubic feet per minute (CFM) or cubic meters per minute

2.18 COOLING AIR

123

Collector enclosure filter

Collector

Upper air deflector

Brush rigging

Removable covers

Stator winding and connections Upper bracket

Station heating duct

Rotor coil

Air baffle Rotor rim

Rotor spider

Collector support

Stator frame

Rotor pole Stator core

Lower air deflector Thrust and guide bearing

Generator shaft

Lower bearing bracket

Stator soleplate

EL. 283 FT. 0.00 IN.

Pit diaphragm

Shaft seal

Figure 2.18-1

Bracket soleplate

Ventilation diagram for a rim ventilated machine.

(CMM) of air needs to be circulating in order to meet the temperature rise guarantee. It comes down to how close to the maximum total temperature of the machine the manufacturer wants to operate to. The way in which the airflow is directed through the machine should never be adjusted or modified unless the manufacturer is consulted. Changing the airflow in one area may give rise to an airflow deficiency in another. Conversely, if it is evident that the machine is not being cooled uniformly, preferably during the heat run tests during commissioning (while the machine is still under warranty), the manufacturer should be consulted as soon as possible. Both systems are efficient provided the air that is being circulated within the machine is kept clean and free from insects and other airborne contaminants

124

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

including grinding, sandblasting, and welding by-products, as well as vapors from paints, and solvents. When the exhaust louvers are opened for heating, intake louvers of equal volume of flow are opened bringing in whatever powerhouse air is available.

2.19 ROTOR FANS/BLOWER The purpose of the rotor fan or blower is to circulate or draw air into a main central cavity and then distribute that air throughout the machine. There are two main types of rotor fans in a vertical hydro generator, a rotor centrifugal fan or a rotor axial fan. Pictures of each type of rotor fan are shown in Figures 2.19-1 and 2.19-2. The ventilation diagram for the rotor centrifugal fan is shown in Figure 2.18-1. The rotor centrifugal fan is situated in the center of the rotor assembly with large vanes that pump air from the center of the generator in through the rim, into the airgap and through the core. The ends of the airgap on the rotor are sealed with covers preventing air escape. In this design, the rim has airflowing through it and is designed accordingly and is known as a ventilated rim. The rotor axial fan is situated on the top and bottom end of the rotor where the airgap is. The fan forces air into the airgap from the top and bottom of the rotor pressuring this area and thus forcing air through the core. In this case, the rim itself does not have the ability to allow air passage, so it is known as a nonventilated rim. The air pumping loss component in the generator affecting overall windage loss is directly proportion to the volume of air pumped. In more recent hydro generator designs, the amount of air being pumped with modern ventilation shrouds/ baffling is in the range of 55–75 CFM/kW (1.55–2.1 m3/min/kW) loss within the

Opening at center of rotor spider/drum for air entry

Figure 2.19-1

Rotor centrifugal fan.

2.20 ROTOR INERTIA, TORQUE, AND TORSIONAL STRESS

125

Axial fans

Figure 2.19-2

Rotor axial fan.

generator ventilation circuit, that is, not including bearing friction losses that are extracted from the unit by other means. Older open ventilated units, with less air baffling and large core ventilation ducts, could have windage loss of around 120 CFM/kW (3.39 m3/min/kW).

2.20 ROTOR INERTIA, TORQUE, AND TORSIONAL STRESS Rotor inertia is a very important consideration when it comes to system events in an effort to maintain system stability. The inertia of the machine is used in determining the effects on the system when transient events occur. In general, normalized inertia constant or H factor can be calculated as follows from Equation (2.11): H=

0 231∗10 − 6 WR2 N 2 MWs MVA MVA

(2.11)

where, W = rotor weight in kg R = radius of gyration in meters N = rotor speed in RPM MVA = generator apparent power The normalized inertial H factor is a comparative measure of the amount of energy stored in the generator rotor at rated speed and is the time in seconds that this energy would sustain rated megawatt output without any additional mechanical torque supplied to the rotor and can be expressed as megawatts seconds divided

126

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

by MVA – (MWs)/MVA. In some cases of inertia discussion, acceleration time is used. This is defined as the time it would take to accelerate the generator to rated speed if the mechanical shaft input was equal to rated torque. It turns out that this time is equal to 2H seconds. In some machine designs, in order to achieve the specified inertia, rim height was added in order to achieve the required mass. This would be evident during an inspection if there are extra rotor laminations at the bottom of the rim well past where the pole body bottom ends as shown in Figure 2.20-1. Notice the colored laminations beneath where the field pole would normally be installed on the rim. This is the extra height that would add inertia to the generator. When installing a new machine, it is a good practice to check with the electrical system operator to ensure the generator has the required H constant. It is common for the system operators to have models of the electrical grid under their control and interconnected systems and can run simulations to ensure system disturbances or emergency conditions are accommodated by the new generator.

Extra laminations Pole body ends here

Figure 2.20-1

Extra rim laminations at the bottom (colored in red).

2.20 ROTOR INERTIA, TORQUE, AND TORSIONAL STRESS

127

The torque on the rotor as seen at the rotor surface in the airgap is as follows in Equation (2.12): Airgap torque (at rotor surface), T=

BAπDr L N−m 2

B=

RPϕ Wb m2 πDr L

A=

N ph N c I a A m πDr

(2.12)

where, Flux density,

Electric loading,

where P = number of poles R = ratio of the pole-face width to the pole pitch Φ = flux per pole in Webers Dr = rotor diameter in meters L = core iron effective length in meters Nph = number of phases Nc = number stator conductors per phase Ia = stator phase current in amperes. The above formula is an elegant result derived from basic physical principles. However, the typical user of a hydro generator can make use of the following very simple expression for the shaft torque as shown in Equation (2.13): Torque~

MW output Speed × Efficiency

(2.13)

The efficiency of a typical hydro generator is above 98%; thus, an approximate and conservative simplification of the equation yields: Torque~

MW output Speed

Using Imperial units Torque =

kW output × 7000 , ftlb RPM

128

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Using SI units: Torque =

kW output × 60 000 , Nm 2πRPM

This torque on the rotor shaft can be significant and creates torsional forging stresses all along its length. Torsional stresses are basically shear stresses in the rotor shaft due to the twist in the shaft that is created by the action of the rotor’s magnetic coupling to the stator magnetic field, as opposed to the opposite force imposed by the water flow to the turbine. Increasing the water flow to the turbine causes the rotor load angle to increase and, hence, MW load to increase, and produces increased mechanical torque in the shaft. The magnetic coupling between the rotor and stator is what inhibits the rotor from running away and keeps the turbine and generator system in synchronous equilibrium. Increasing or decreasing the rotor magnetic field causes the load angle to increase or decrease, but does not actually change the torque applied, only the angle of the torque, or, in electrical terms, the power factor and reactive power output of the machine. Under some electrical fault conditions, the airgap torque can be significantly higher than the rated torque. It is not unusual to have faulty synchronization of the unit onto the grid (phase angle difference between the generator and the system grid) being 10 times the rated torque. During this event, the rotor shaft can be exposed to additional torque loading that is dependent on the ratio of the generator rotational inertia to that of the rotational inertia of the turbine runner.

2.21 THRUST AND GUIDE BEARINGS 2.21.1

Introduction

All generators require bearings to operate with minimal friction and vibration. For a vertical hydro generator, there are a several bearings that are part of the design and variations of these bearings can be found depending on the manufacturer. The types of bearings are, the thrust bearing, thrust bearing/guide bearing combination, and additional guide bearings along the shaft length if required particularly for long shaft lengths where extra support is required, both styles are shown in Figures 2.21-1 and 2.21-2. The upper and lower bracket, and the thrust and guide bearings, make up the generator load bearing structure. Bearing structures support axial and radial loads. For vertical generators, the axial load consists of the weight of the rotating components and of the hydraulic thrust. The hydraulic thrust is exerted by the water flow through the water passage and the turbine. The hydraulic thrust is transferred to the headcover, spiral casing, and along the shaft to the generator thrust bearing. The generator and turbine guide bearings manage the radial forces exerted from water transients and flow disturbances occurring in rough zones of operation, stops and starts, along with load rejections [9]. As well, any mass and magnetic imbalances that exist on the rotor are transferred to the powerhouse foundation through the guide bearings and their support structures [8].

2.21 THRUST AND GUIDE BEARINGS

129

Overspeed device (or P.M.G. for governor control not supplied with generator) Pilot exciter Main exciter Exciter inspection platform Upper oil reservoir Combined upper guide and thrust bearing assembly Stator coil

Bearing cooling coil Upper bracket

Stator laminations Air housing Cooler section Field coil Rotor spider Laminated rotor rim Stator sole plate Foundation bolt Combined brakes and jacks

Shaft and coupling flange

Lower oil reservoir and guide bearing assembly

Lower bracket sole plate

Lower bracket

Figure 2.21-1 Shows a cross section of a conventional two guide bearing generator. Source: Courtesy of Voith.

Overspeed device (or P.M.G. for governor control not supplied with generator) Air housing Pilot exciter Main exciter

Stator coil

Stator laminations

Upper bracket Cooler section Rotor spider Field coil Laminated rotor rim Stator sole plate

Combined guide and thrust bearing assembly

Foundation bolt Combined brakes and jacks

Oil reservoir

Lower bracket sole plate Lower bracket Bearing cooling coil

Shaft and coupling flange

Compression tube jack screws

Figure 2.21-2 Shows a cross section of an umbrella style generator with one guide bearing. Source: Courtesy of Voith.

130

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

There are a couple of different materials that can be used for thrust bearings, Babbitt or Teflon™. Babbitt was invented back in 1839 by Isaac Babbitt who was a goldsmith at the time and contains multiple alloys [10]. Polytetrafluoroethylene (PTFE) is a synthetic fluoropolymer of tetrafluoroethylene that has numerous applications. The best known brand name of PTFE-based formulas is Teflon™ by The Chemours Company. The Chemours Company is a 2015 spin-off of DuPont Co., which discovered the compound in 1938 [11]. Because of the relative softness of pure Teflon™ many PTFE-based bearing liners used for higher speed and highly loaded thrust bearings today have a mixture of Teflon™ and carbon (carbon-filled PTFE) to provide better dimensional shape stability under high temperature exposure [8]. 2.21.1.1 History of the Popular Kingsbury Thrust Bearing In the late 1880s, experiments were being conducted on the lubrication of bearing surfaces. The idea of “floating” a load on a film of oil grew from the experiments of Beauchamp Tower and the theoretical work of Osborne Reynolds who was a British engineer, physicist, and educator best known for his work in hydraulics and hydrodynamics [12]. Reynolds showed that “if an extensive flat surface is rubbed over a slightly inclined surface, oil being present, there would be a pressure distribution with a maximum somewhere beyond the center in the direction of motion.” Prior to the development of the pivoted shoe thrust bearing, marine propulsion relied on a “horseshoe” bearing which consisted of several equally spaced collars to share the load, each on a sector of a thrust plate. The parallel surfaces rubbed, wore, and produced considerable friction. Design unit loads were on the order of 40 PSI. Comparison tests against a pivoted shoe thrust bearing of equal capacity showed that the pivoted shoe thrust bearing, at only 1/4 the size, had 1/7 the area but operated successfully with only 1/10 the frictional drag of the horseshoe bearing [13]. In 1896, inspired by the work of Osborne Reynolds, Albert Kingsbury conceived and tested a pivoted shoe thrust bearing. According to Dr. Kingsbury, the test bearings ran well. Small loads were applied first, on the order of 50 PSI (which was typical of ship propeller shaft unit loads at the time). The loads were gradually increased, finally reaching 4000 PSI, the speed being about 285 RPM [13]. 2.21.1.2 First Application of the Pivoted Shoe Thrust Bearing In 1912, Albert Kingsbury was contracted by the Pennsylvania Water and Power Company to apply his design in their hydroelectric plant at Holtwood, PA. The existing roller bearings were causing extensive down times (several outages a year) for inspections, repair, and replacement. The first hydrodynamic pivoted shoe (or pad) thrust bearing was installed in Unit 5 on 22 June 1912. At start-up of the 12 000 kW unit, the bearing wiped. In resolving the reason for failure, much was learned about tolerances and finishes required for the hydrodynamic bearings to operate. After properly finishing the runner and fitting the bearing, the unit ran with continued good operation. This bearing, owing to its merit of running 75 years with negligible wear under a load of 220 tons, was designated by ASME as the 23rd International Historic Mechanical Engineering Landmark on 27 June 1987 [13].

2.21 THRUST AND GUIDE BEARINGS

2.21.2

131

Important Concepts

2.21.2.1 Oil Film and Oil Film Thickness Thrust and guide bearings are the most important generator subassemblies since they transfer and bear the rotational masses and loads. The most important condition for bearing these loads is the oil film separation in between the two bearing surfaces. Having the oil film wedge in between bearing surfaces allows the friction to be exerted within the oil film, also known as “liquid/fluid friction.” In order to assure bearing of this load, there is a necessity to form an oil wedge (hydrodynamic wedge) within the bearing that narrows down toward the sense of rotation [9]. The forming of the oil film can be explained as follows. The oil adheres to the surface (via friction force), and at the beginning of the runner plate rotation in the case of thrust bearing, or shaft journal rotation in the case of guide bearing, the laminar layer of the oil is wedged by the rotational motion and has a tendency of flowing in all directions. Due to the viscosity of the oil, the flowing effect is exerting pressure within the oil and forms the oil wedge. As the gap between the bearings gets smaller due to the wedge, the flowing of the oil is more obstructed and oil pressure in the oil wedge increases. The pressure in the oil wedge lifts the shaft journal or the thrust bearing runner plate increasing the speed at which the oil flows. Balance in the thrust bearing is reached once the quantity of the oil is equal in the all directions within the bearing gap. Thrust and radial loads are balanced by summarized (equalized) oil pressure within the oil wedge. The oil wedge in the thrust bearing is created by tilting the bearing pads. In the guide bearings, the gap in-between the bearing pads with a beveled edge and the shaft journal provides sufficient space to accommodate oil entrance between the sliding bearing/shaft surfaces [9]. Bearing oil balance depends on the load and rotational speed. As the load increases, or the speed of rotation decreases, the thickness of the oil film decreases as well [9]. There is a minimal thickness of oil film that should be sufficient to accommodate load fluctuations as the surface roughness should not allow oil film disruption and direct contact between the bearing surfaces. In the case of direct contact of bearing surfaces, the generation of heat results in permanent bearing damage. The oil film thickness in thrust bearings should be a minimum of 30–35 μm (1.18–1.38 mils) for long term safe bearing function [9]. 2.21.2.2 Loads and Load Displacement Hydraulic thrust on large units can transfer loads as large as 6 745 000–7 869 000 lbf (30–35 MN), whereas the axial load, is in principal, constant. At the time of this writing, some of the largest generators in the world have thrust bearings capable of 13 218 770 lbf (58.8 MN) [14]. Load fluctuations are influenced by hydraulic transient pressure pulsations due to mechanical geometrical imperfections in alignment, the variable surface of the bearing pads, or thrust bearing runner plate imperfections. The axial position of the rotor with respect to the stator could also exert a vertical component of electromagnetic unbalance. The total variable component of axial load could reach as high as 10–15% of the total axial load [9].

132

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Radial loads on vertical generators could be the result of poor static balance of the generator rotor, incorrect shaft alignment, an ambiguous hydraulic force due to transient behavior, or a magnetic unbalance of the rotor caused by a compromised airgap in the generator. Compromised airgaps can be the result of poor circularity or concentricity of the rotor or stator or both. All of these loads are supported by radial guide bearings housed within their respective brackets and transfer the load to the generator foundation [9]. Theoretically, in the case of radial forces, the radial loads on the guide bearings are minimal if not negligible. Within the normal operational sequence, guide bearings are indeed exposed to constant fluctuating radial loads that are minimal in nature. However, in runaway mode or in the case of a load rejection, these loads are transferred to the two (two bearings is common) generator guide bearings. The third guide bearing (turbine guide) in this case is not taken into consideration when calculating the load support. In the case that generator has only one guide bearing, the turbine guide bearing is then taken into consideration for the load support calculation [9]. In modern hydro-generator design, the ultimate design load is that which would result from a short circuit of the rotor poles [8].

2.21.3

Thrust Bearings

There are numerous thrust bearing designs and an exhaustive list of designs is beyond the scope of this book, thus, only a few will be discussed here. The thrust bearing takes all the vertical weight of the rotor and the hydraulic downward thrust produced by the runner. This is a very large load when you consider the surface area this load is distributed on. The bearing can be located under or above the rotor. A typical design uses many pie-shaped sections of bearing called “shoes” that have Babbitt metal or Teflon™ surfaces as shown in Figures 2.21-3 and 2.21-4. As shown, these shoe sections form a single ring and in cases where segments are centrally pivoted, may require that there are inner and outer rings used depending on the radial dimension of the segment. One way for support of the segments is that each shoe is supported by a network of load cells (jack screws) which has to support the shoe equally with respect to each other so the bearing is then “loaded” equally when in service. A typical load cell arrangement is as shown in Figure 2.21-5. Incorrect loading of the load cells could result in a bearing “wipe,” so setting up the load cells correctly is paramount. The load cell is an adjustable jack screw which makes it possible to distribute the load equally between individual shoes. These shoes, which now constitute the thrust bearing, operate in a bath of oil which ensures that the shoe surfaces will be separated by a wedge-shaped oil film from the runner plate (surface that carries the weight of the rotor) as shown in principle in Figure 2.21-6. This figure also shows the pressure distribution of the oil wedge. The practical application of this principle is that the loaded plate is actually the runner plate which is attached to the shaft and rotates in the same bath of oil at shaft speed as shown in Figure 2.21-7. This figure shows an older Westinghouse design and is not the current Voith design. The jack screws under the shoes are the supports or pivots, and they are

2.21 THRUST AND GUIDE BEARINGS

133

Grooves and hole for oil

Figure 2.21-3 Shows Babbitt bearing pie pieces or “shoes.” Source: Courtesy of Ryan Gillespie of Ontario Power Generation.

Figure 2.21-4

Shows Teflon™ bearing pie pieces or “shoes.”

in the center of the shoe (circumferentially) [15]. A similar arrangement using springs is shown in Figure 2.21-8 [8]. The runner plate is shown in Figure 2.21-9, and it is this precisely prepared surface that sits on top of the shoes which has the oil wedge in between while in service. This bearing design when

134

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Shoes

Load cells

Figure 2.21-5

Shows load cells supporting the thrust bearing shoes.

Rotation of runner plate

Hydrodynamic oil wedge Stationary bearing shoe

Pivot point

Resultant

Centerline of shoe

Oil pressure

Pmax

Figure 2.21-6 Shows oil pressure distribution between runner plate and pivoting bearing shoe.

2.21 THRUST AND GUIDE BEARINGS

135

Weight

Oil level

Loaded plate Oil film

Oil film Shoe

Pivot

Speed V

Oil film

Shoe Pivot

Pivot

Shoe

Figure 2.21-7 Sketch showing oil wedge principle for a pivoting shoe on jack screw (pivot). Source: Courtesy of Voith. Load

Pressure Rotating surface Oil film Segment Spring support

Figure 2.21-8 Oil pressure distribution between the runner plate for a spring supported thrust bearing design.

Figure 2.21-9 Shows runner plate that sits on top of thrust bearing shoes. Source: Courtesy of Ryan Gillespie of Ontario Power Generation.

136

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

using Babbitt may also employ an oil lift system used during routine start-ups and shut-downs as well as after prolonged shutdown periods. The oil lift system works by pumping high pressure oil between the mating surfaces introducing a thin film of oil which will then produce the “hydro dynamic wedge” and properly lubricate the surfaces. This is achieved by having circular grooves and holes strategically placed on each shoe as shown in Figure 2.21-3. When the machine is slowing down or shut down, the oil between the mating surfaces is slowly being squeezed out. There will be a point in time during and/or after shutdown when enough oil is squeezed out that there will not be enough oil to initially prelubricate the bearing surfaces during start-up rotation. In this case, insufficient oil is in between the surfaces and will quickly overheat and the bearing will wipe. The manufacturer will be able to determine what timing after shutdown will require an oil lift system to be activated. Another thrust bearing design that is more simplified employs special spring assemblies to form a bed for each shoe as shown in Figures 2.21-10 and 2.21-11. Using compressible springs as the segment support eliminates the need for high tolerances of the spring assemblies and for perfect alignment of the shaft system components as the springs act as a pressure relief should the loading or the part condition not be as initially assembled [8]. This allows the bed of springs to support each shoe equally around the ring. If Babbitt is used an oil lift system would be used in this design as well. Cooling of the thrust bearing oil can be done in several ways. The most traditional method is to have a cooling coil inside the oil reservoir as shown in Figures 2.21-12 and 2.21-13. In the case of Figure 2.21-13, the thrust bearing is on the inside of the cooler tubes shown. There are many variations of cooling tubes depending on the water conditions present such as the degree of silt in the water, micro-organisms that attach themselves to the cooler tubes, etc. Stainless steel or other alloys may also be used to make the tubes and may be combined with the nickel.

Figure 2.21-10 Shows spring assembly to support the shoes.

2.21 THRUST AND GUIDE BEARINGS

137

Figure 2.21-11 Shows the bearing shoe installed.

Figure 2.21-12 Shows the copper cooling coil at the bottom of the oil reservoir.

Water is circulated inside the cooling coil and the heat is exchanged through the tubes. The oil is self-pumping in these enclosed cooling designs so the oil mixes inside the oil pot to avoid stratification of oil at different temperatures. Fins may be added to increase the efficiency of the cooling coil. Another method is to pump the oil to an external cooler and cycle it back into the oil pot using external electric pumps. The oil reservoir size depends on the machine design and the specifications required by the customer. Considerations such as available water temperature, speed of the machine (bearing losses), available space for the thrust bearing (size restraints), and sustained time for runaway speed conditions all play a role in the amount of oil in the reservoir and cooling capacity the bearing must have. The seal for the oil reservoir is discussed in Chapter 3.

138

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Figure 2.21-13 Shows thrust bearing cooler (extruded aluminum fin CuNi) tubes inside the oil reservoir.

2.21.4

Thrust Bearing Pressure

When plane-type thrust bearings replaced roller bearings for use in vertical axis hydro units, the average bearing pressure was limited to about 40 PSI (0.28 MPa). Improvements to the segment support allowed average pressures to get into the 290 PSI (2 MPa) range. With more operating data, confidence in these bearings quickly followed and by 1930, several large capacity thrust bearings operated with average design pressures in the 450 PSI (3.1 MPa) range. Between 1930 and 1990, the average thrust bearing increased slowly to about 400 PSI (2.8 MPa), aided by better Babbitt bonding technology, and the use of high pressure oil injection systems for start/stop operation and with the use of PTFE liners. In the last 20 years, finite element analysis has allowed the mathematical evaluation of thrust bearing segments, and their support, when exposed to the combined loading of hydrodynamic forces and thermal gradients. This enhanced thrust bearing assessment has improved the performance predictability of the bearing and has enabled thrust bearing pressures of 725 PSI (5 MPa), and above, and to achieve higher reliability than was experienced in the early use of this bearing type, which had even lower pressure designs [8].

2.21.5

Guide Bearings

The previously discussed concept of a hydrodynamic wedge also applies to guide bearings. Professor Osborne Reynolds showed that oil, because of its adhesion to the journal and its resistance to flow (viscosity), is dragged by the rotation of the journal so as to form a wedge-shaped film between the journal and journal bearing

2.21 THRUST AND GUIDE BEARINGS

139

Journal bearing

Rotation

Journal

Oil

Adhesion

Oil wedge

Figure 2.21-14 Demonstrates the principle of the hydrodynamic wedge in the guide bearing. Source: Courtesy of Kingsbury Inc.

as shown in Figure 2.21-14. This action sets up the pressure in the oil film which thereby supports the load. This wedge-shaped film was shown by Reynolds to be the absolutely essential feature of effective journal lubrication [13]. The guide bearings are there to make sure the entire shaft and rotor assembly once aligned, remains in that position. The guide bearings are typically made from segmented Babbitt metal as shown in Figure 2.21-15 and are lubricated and cooled using oil. Some guide bearings are part of the thrust bearing assembly and share the same oil reservoir as shown in Figure 2.21-16. 2.21.5.1 Recommended Guide Bearing Gap and Gap Calculation The guide bearing gap directly influences the bearings capacity to support the load, as well as the positioning of the shaft journal inside the bearing body. Bearings with larger loads require specific definitions of the relative and absolute gaps as shown in Equations (2.14) and (2.15). D−d Relative gap d

(2.14)

Z = D − d Absolute gap

(2.15)

φ=

140

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Figure 2.21-15 Typical shaft guide bearing (sleeve type).

Figure 2.21-16 Guide bearing shoe in the thrust/guide bearing combination assembly (segmental type).

where, D = guide bearing internal diameter d = shaft bearing journal diameter The absolute gap can be defined as the need for handling small variations of the shaft journal positioning within the bearing.

2.21 THRUST AND GUIDE BEARINGS

141

Bearings with hydrodynamic oil film lubrication have values of the Relative Gap defined as follows: 0 0003 < φ < 0 0005 The quoted minimum and maximum respective values are for power transmission shafts [9]. 2.21.5.2 Bearing Gap Sample Calculation [9] Turbine shaft journal diameter is d = 1295.4 mm (51 ) and from Equation (2.14): φ=

D−d re − arranging yields d

φd = D − d and from Equation 2 15 Z = D − d and substituting yields Z = φd

(2.16)

Substituting values of φ for the minimum and maximum (0.0003 and 0.0005): Z = 0 0003 × 1295 4 = 0 388 mm 0 0153 − Minimum absolute gap calculated for 51 diameter hydrodynamic lubricated bearing Z = 0 0005 × 1295 4 = 0 647 mm 0 0255 − Maximum gap calculated for 51 diameter hydrodynamic lubricated bearing From rearranging Equation (2.15), the turbine bearing diameter may now be calculated as follows: D=Z+d

(2.17)

D = 1295 4 mm + 0 388 – 0 647 mm The turbine bearing diameter will be 1295.78 mm (51.0153 ) minimum to 1296.04 mm (51.0255 ) maximum. 2.21.5.3 Bearing Insulation and Bearing Current Variations in reluctance in the magnetic circuit of the generator may cause periodic changes in the amount of flux which links the shaft. This change in flux may generate sufficient voltage to circulate current through the circuit consisting of the shaft, bearings, brackets, and frame. This condition is encountered only in machines having guide bearings located in positions relative to the electromagnetic axial centerline that would support this voltage. If precautions are not taken to prevent the flow of circulating current, it will have a destructive effect on the shaft journals and bearings.

142

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

It is not always practical to provide control of the generation of shaft voltages, so it becomes necessary to insulate one of more of the bearings from their supporting members. This insulation interrupts the path for circulating currents. The insulation consists of a suitable thickness of modern-day fiber and epoxy glass laminate or Micarta placed between the bearing shell or bearing support and the bearing brackets. Further, in order to avoid short circuiting the insulation, all water and oil piping and temperature detector bulbs must also be insulated [15].

2.21.6

Deterioration and Failure of the Bearing Surface

The terms “bearing wipe” or “the generator wiped a bearing” are commonly heard among persons that operate hydro generators. The following five items form a brief introduction to what the aforementioned terms actually mean. 2.21.6.1 Wiping Defined Wiping is a form of damage that occurs whenever a substantial (visual) amount of Babbitt is displaced or removed, likely by direct contact with the journal or runner. Often, this material is redeposited at another location on the bearing surface or on an edge [16] (see Figures 2.21-17 and 2.21-18). It is also possible for a PTFE bearing to wipe the remnants as are shown in the spring bed in Figure 2.21-19. 2.21.6.2 Mechanism of Wiping Three mechanisms can be visualized as causing wiping. First, sufficient softening of the Babbitt can occur as a consequence of direct contact (friction) between the journal, or runner, and the Babbitt. Second, the bearing metal may be plastically

Figure 2.21-17 Shows a Babbitt bearing wipe. Source: Courtesy of Ryan Gillespie of Ontario Power Generation.

2.21 THRUST AND GUIDE BEARINGS

143

Figure 2.21-18 Shows another bearing wipe. Source: Courtesy of Ryan Gillespie of Ontario Power Generation.

Figure 2.21-19 Shows what is left of a PTFE bearing when shoe removed and spring bed exposed. Source: Courtesy of Ryan Gillespie of Ontario Power Generation.

deformed by mechanical cold working by the journal. Third, abnormal hydrodynamic pressures developed near the hmin may cause local plastic deformation of the Babbitt. The term hmin refers to the total static and dynamic load for which a bearing can support and maintain a given value of minimum film thickness [16].

144

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

2.21.6.3 Classification of Wiping Damage Wiping (full or partial) is probably a familiar kind of damage encountered in Babbitted bearings. The general appearance of a wiping in a thrust bearing is a broad polished area with a buildup of displaced Babbitt downstream of the polishing. Wiping is a generic term, being evidence of underlying causes that can be divided into primary and secondary categories. The primary category consists of wiping from causes such as • Misalignment • Tight clearances • Startup conditions related to lubricant supply • Differences between cold and hot alignment • Elastic and thermal distortions • Overload • Rotor dynamic instability • Shocks • Improper assembly • Unexpected load angle • Oil starvation • Boundary lubrication [16] The secondary category consists of a wipes that are a consequence of other damaging mechanisms. These mechanisms include • Fatigue wiping • Hydrogen blisters • Tin oxide • Electrolysis • Electromagnetic spark tracks [16] 2.21.6.4 Appearance of Wiping Since wiping can arise from direct contact between the runner and the bearing, this sort of damage is characterized by the physical displacement of the Babbitt material. The heaviest signs of this displacement are usually near the operating hmin, noting that the location of the minimum film thickness may have been influenced by other causes of bearing damage. This displacement of the Babbitt extends over a fairly wide angular region and is characterized by irregular edges at the end of the wiped area, commonly where the displaced material has been deposited on top of undamaged material. If the temperatures during the wipe are not high, the damaged area has a polished appearance. If high temperatures are generated during the wiping, portions of the wiped area may look dark and burnished, often as a result of damage to the lubricant [16].

2.21 THRUST AND GUIDE BEARINGS

145

2.21.6.5 Causes of Wiping Wiping is another generic family, possibly related to more severe wear. The causes of overheating, namely, oil starvation, faulty bearing geometry, external heat sources, and misalignment apply equally here. Additional direct causes of a bearing wiping are the following: • Excessive static and/or dynamic bearing load • Excessive synchronous (imbalance) or nonsynchronous (for example, instability) vibration • Shock loading during operation • Loss of Babbitt strength (softening), especially as a result of high operating temperatures • Insufficient operating oil viscosity, especially as a result of high operating temperatures [16]. It is beyond the scope of this book to get into all of the mechanisms that can cause or contribute to bearing failures. The reader is referred to the references at the end of the chapter that offer a comprehensive discussion on bearings. Next is a list of the most common items that cause or contribute to bearing failures. 2.21.6.6 Abrasion A bearing surface exhibiting circumferential scratches is the result of abrasion damage. Abrasion is caused by hard debris, which is larger than the film thickness, passing through the oil film. The debris may embed itself in the soft Babbitt, exhibiting a short arc on the shoe surface, ending at the point where the debris becomes embedded as shown in Figure 2.21-20. Depending on the debris size, the scratch may continue across the entire shoe surface. Abrasion damage becomes worse with

Figure 2.21-20 Bearing shoe with surface abrasion. Source: Courtesy of Kingsbury Inc.

146

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

time. Surface scratches allow an escape for lubricating oil in the oil wedge, decreasing the film thickness. This will eventually lead to bearing wipe. Another source of abrasion damage is a rough journal, collar, or runner surface. Roughness may be due to previous abrasion damage. It may also be from rust formed after extended periods of down time. New bearings should not be installed when the rotating component is visibly damaged. In order to eliminate abrasion damage, the lubricating oil must be filtered. If the oil cannot be filtered or has degraded, it should be replaced. It is important to evaluate the filtering system, since the problem may be an incorrectly sized filter. The filter should only pass debris smaller in size than the predicted bearing minimum film thickness. In addition to filtering/replacing the oil, the entire bearing assembly, oil reservoir and piping should be flushed and cleaned [13]. It should be noted that even oil filtering cannot extract some of the “heavier” debris that may rest on the floor of the oil pot. For this reason, if significant Babbitt abrasion is encountered, the oil should be drained from the pot and a manual clean-up of the reservoir and the bearing parts should be completed [8]. 2.21.6.7 Tin Oxide Damage This is one of several electrochemical reactions which eliminate the embedability properties of a fluid-film bearing. Tin oxide damage is recognizable by the hard, dark brown or black film that forms on the Babbitt as shown in Figure 2.21-21. Tin oxide forms in the presence of tin-based Babbitt, oil and salt water, beginning in areas of high temperature and pressure. Once it has formed, it cannot be dissolved, and its hardness will prevent foreign particles from embedding in the bearing lining. This allows abrasion damage to occur. Pieces of tin oxide may break off during operation and score the journal, collar, or runner. The formation of tin oxide will also eliminate bearing clearance. This damage may be stopped by eliminating some or all of the contributing elements. The lubricating oil must be replaced.

Figure 2.21-21 Bearing shoe with tin oxide damage. Source: Courtesy of Kingsbury Inc.

2.21 THRUST AND GUIDE BEARINGS

147

A reduction in oil temperature may also discourage the formation of tin oxide. In addition to replacing the oil, the entire bearing assembly, oil reservoir, and piping should be flushed and cleaned with mineral spirits [13]. 2.21.6.8 Overheating Overheating damage may represent itself in many ways, such as Babbitt discoloration as shown in Figure 2.21-22, cracking, wiping, or deformation. Repeated cycles of heating may produce thermal ratcheting as shown in Figure 2.21-23, a type of surface deformation that occurs in anisotropic materials.

Figure 2.21-22 Bearing that exhibits overheating. Source: Courtesy of Kingsbury Inc.

Figure 2.21-23 Bearing shoe exhibiting thermal ratcheting. Source: Courtesy of Kingsbury Inc.

148

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

These materials possess different thermal expansion coefficients in each crystal axis. Overheating may be caused by numerous sources, many of which concern the quantity and quality of the lubricant supply. Among the possible causes are • Improper lubricant selection • Inadequate lubricant supply • Interrupted fluid film • Boundary lubrication The following conditions may also cause overheating: • Improper bearing selection • High pressure oil lift system failure • Poor collar, runner, or journal surface finish • Insufficient bearing clearance • Excessive load • Overspeed • Harsh operating environment [13] 2.21.6.9 Oil Starvation It is often possible to distinguish oil starvation (i.e. the total absence of lubricant) from a less than adequate oil flow by closely examining the Babbitt surface of the bearing. In Figure 2.21-24, the Babbitt has been completely removed from the shoe surface. If the Babbitt has been completely removed from the shoe surface, and there is no accumulation of Babbitt in between the shoes, then there was no oil flow to the shoe surface. If there is Babbitt accumulation between the shoes, there was at least some oil flow on the shoe surface to cool and solidify the molten Babbitt. With

Figure 2.21-24 Bearing show exhibiting oil starvation. Source: Courtesy of Kingsbury Inc.

2.21 THRUST AND GUIDE BEARINGS

149

Figure 2.21-25 Show exhibiting outer edge Babbitt erosion. Source: Courtesy of Kingsbury Inc.

no oil coming in at the leading edge of the shoe, this area typically shuts down, resulting in the Babbitt being eliminated in the corner near the outer diameter as shown in Figure 2.21-25 [13]. 2.21.6.10 Electrical Pitting Electrical pitting as shown in Figure 2.21-26 appears as rounded pits in the bearing lining. The pits may appear frosted, or they may be blackened due to oil deposits. It is not unusual for them to be very small and difficult to observe with the unaided eye. A clearly defined boundary exists between the pitted and undamaged regions, with the pitting usually occurring where the oil film is thinnest. As pitting progresses, the individual pits lose their characteristic appearance as they begin to overlap. Pits located near the boundary should still be intact. The debris that enters the oil begins abrasion damage. Once the bearing surface becomes incapable of supporting an oil film, the bearing will wipe. The bearing may recover an oil film and continue to operate, and pitting will begin again. This process may occur several times before the inevitable catastrophic bearing failure. Electrical pitting damage is caused by intermittent arcing between the stationary and rotating machine components. Because of the small film thicknesses relative to other machine clearances, the arcing commonly occurs through the bearings. Although the rotating and other stationary members can also be affected, the most severe pitting occurs in the soft Babbitt [13]. The hazing of the Babbitt surface is typically the result of electrical discharge with very low voltage that exists with very thin oil films. If electrical discharge takes place in larger oil films, where the discharge voltage can be much higher, there may be evidence of additional “crater-type” damage on the journal or rotating ring [8]. 2.21.6.11 Fatigue Fatigue damage, as shown in Figures 2.21-27–2.21-29, may represent itself as intergranular or hairline cracks in the Babbitt. The cracks may appear to open in the direction of rotation. Pieces of Babbitt may spall out or appear to be pulled away in the direction of rotation. The cracks extend toward the Babbitt bond line,

150

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Figure 2.21-26 Bearing shows signs of electrical pitting. Source: Courtesy of Kingsbury Inc.

and may reveal the shoe backing. A combination of causes may contribute to fatigue damage, but concentrated cyclic loading is usually involved. The fatigue mechanism involves repeated bending or flexing of the bearing, and damage occurs more rapidly with poor bonding. As well, fatigue is more prevalent in bearing designs where the Babbitt thickness is larger [8]. It is important to note that fatigue damage will occur without poor bonding. Fatigue can occur when conditions produce concentrated cyclic loads, such as • Misalignment • Journal eccentricity • Imbalance • Vibration • Thermal Cycling • Bent Shaft [13] 2.21.6.12 Cavitation Cavitation damage appears as discreet irregularly shaped Babbitt voids which may or may not extend to the bond line as shown in Figure 2.21-30. It may also appear as localized Babbitt erosion. The location of the damage is important in

2.21 THRUST AND GUIDE BEARINGS

151

Figure 2.21-27 Edge load pivoted show showing Babbitt mechanical fatigue. Source: Courtesy of Kingsbury Inc.

Figure 2.21-28 Edge load journal shell with mechanical fatigue. Source: Courtesy of Kingsbury Inc.

152

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

Figure 2.21-29 Shoe segment showing fatigue. Source: Courtesy of Kingsbury Inc.

Figure 2.21-30 Shows result of cavitation erosion on the bearing surface. Source: Courtesy of HydroTech Inc.

determining the trouble source. Often called cavitation erosion, cavitation damage is caused by the formation and implosion of vapor bubbles in areas of rapid pressure change. Damage often occurs at the outside diameter of thrust bearings due to the existence of higher velocities. This type of damage can also affect stationary machine components in close proximity to the rotor. Based on its source, cavitation can be eliminated in a number of ways:

2.21 THRUST AND GUIDE BEARINGS

153

• Radius/chamfer sharp steps • Modify bearing grooves • Reduce bearing clearance • Reduce bearing arc • Eliminate flow restrictions (downstream) • Increase lubricant flow • Increase oil viscosity • Lower the bearing temperature • Change oil feed pressure • Use harder bearing materials • Lower bearing operating pressure [8] • Eliminate pressure disruptions, such as gaps between split rotating ring sections [8] The lubricating oil must be filtered or replaced. In addition to filtering/replacing the oil, the entire bearing assembly, oil reservoir, and piping should be flushed and cleaned. Depending on the extent of damage, voids in the Babbitt can be puddlerepaired. The original bearing finish must be restored. Journal shoes may also be puddle-repaired and refinished. If this cannot be done, shoes must be reBabbitted or the shoes replaced [8, 13]. 2.21.6.13 Oil A quick visual examination of the oil or oil filter may be all that is required to determine that a problem exists and that further investigation is necessary. Cloudy or discolored oil indicates that a problem exists. A thorough oil analysis can provide very useful data to assist in diagnosing bearing or machine distress. Be aware that the usefulness of the analysis is directly related to the information you request. As a minimum, the following should be supplied: • Particulate density • Particulate breakdown • Viscosity • Water contamination • Chemical breakdown The amount of particulate, as well as its content, can identify potential trouble spots. Oil viscosity will decrease in time, and whether or not distress is suspected, it should be periodically evaluated. Water contamination is extremely unwanted, since it can cause rust and oil foaming, and if it is drawn into the oil film, bearing failure. A chemical breakdown of the oil will help to determine the integrity of additive packages and the presence of unwanted contaminants [13].

154

CHAPTER 2

GENERATOR DESIGN AND CONSTRUCTION

2.21.6.14 Operational Data Perhaps the most important source of diagnostic information is unit operational data. Identifying periods of load or speed changes, recent maintenance, pad temperature, and vibration level trends, or the performance of related machinery may also help determine the root cause of distress. Vibration data or an analysis may help discover existing problems, as well as examining the remaining bearings in a troubled unit [13]. 2.21.6.15 Upgrades and Uprates The most severe impact to thrust bearing aging comes with generator upgrade and uprate. Usually, after the upgrade and uprate, hydraulic thrust is increased, and it is directly transferred to the thrust bearing surface. Sometimes the increase of the hydraulic thrust is such that it surpasses the load-carrying capability of the bearing. The most significant bearing defect showcased in these cases is bearing cavitation close to the oil lift pump ports or at the overloaded bearing pad areas which exhibit missing Babbitt material [9]. Another negative impact on thrust and guide bearing life is the increased use of hydro-generators as peaking units. These more frequent start/stop cycles increases the cyclic stress on the bearing, thereby introducing a higher probability of Babbitt fatigue [8].

2.22 REFERENCES 1. ASME, Hydro Power Technical Committee (1996). The Guide to Hydropower Mechanical Design, Kansas City, HCI Publications, 0-9651765-0-9. 2. Znidarich, M. (2008). Hydro generator high voltage stator windings: Part 1 – essential characteristics and degradation mechanisms. Australian Journal of Electrical and Electronics Engineering, Vol 5(1), 1–17. 3. Znidarich, M. (2009). Hydro generator high voltage stator windings: Part 3 – stator winding slot support systems. Australian Journal of Electrical and Electronics Engineering, Vol 6(1), 1–10. 4. Znidarich, M. (2009). Hydro generator high voltage stator windings: Part 2 – design for reduced copper losses and elimination of harmonics. Australian Journal of Electrical and Electronics Engineering, Vol 5(2), 119–135. 5. IEEE (2015). IEE C50.12-2005, IEEE Standard for Salient-Pole 50 Hz and 60 Hz Synchronous Generators and Generator/Motors for Hydraulic Turbine Applications Rated 5 MVA and Above, New York, IEEE. 6. IEEE (2009). IEEE 1665: Guide for the Rewind of Synchronous Generators, 50 Hz and 60 Hz, Rated 1 MVA and Above, Piscataway, NJ, IEEE. 7. Karmaker, H. (2003). Broken damper bar detection studies using flux probe measurements and time-stepping fintie element analysis for salient pole synchronous machines. Proceedings of the 4th IEEE International Symposium on Diagnostics for Electric Machines, Power Electronics and Drives (SDEMPED), Atlanta, GA (24–26 August 2003) (p. Unknown). Atlanta: SDEMPED. 8. Contribution by Wayne Martin P. Eng, Andritz Canada 9. Contribution by Tim Maricic P. Eng, Ontario Power Generation, with Refs. [17–19].

2.23 FURTHER READING

155

10. Wikipedia (2018). https://en.wikipedia.org/wiki/Babbitt_(alloy) (accessed 19 March 2020). 11. Wikipedia (2018). https://en.wikipedia.org/wiki/Polytetrafluoroethylene (accessed 19 March 2020). 12. Encyclopedia Britannica (2018). Searched for Osborne Reynolds. https://www.britannica.com/biography/Osborne-Reynolds (accessed March 2018). 13. Reproduced from, “Kingsbury A General Guide To The Principles Operation And Troubleshooting Of Hydrodynamic Bearings, Philadelphia PA, March 2015.” 14. Schafer, D. and Liangwei, S. (2001). Investigations into a 6000 tons thrust bearing with Teflon or Babbitt layer for the Three Gorges units. Proceedings for the 5th International Conference on Electrical Machines and Systems, 2001. ICEMS 2001, Shenyang, China, (18–20 August 2001), pp. 131–136. 15. Westinghouse Vertical Water Wheel Synchronous Generators Instruction Book, Westinghouse Electric Corporation, East Pittburgh, PA, 1950. 16. Reproduced from, “EPRI Manual of Bearing Failures and Repair in Power Plant Rotating Equipment: 2011 Update. EPRI, Palo Alto, CA: 2011,1021780”. 17. Hirs, G. G. (1962). The load capacity and stability characteristics of hydrodynamic grooved journal bearings. ASLE Vol 8, 296–305. 18. Shigley, J. and Mischke, C. (1989). Mechanical engineering Design, Boston, MA, McGraw Hill. 19. Engineering Mechanical Handbook, University of Belgrade, 1962. 20. The Centre for Energy Advancement through Technological Innovation (CEATI), (2015). Hydroelectric Turbine-Generator Units Guide for Erection Tolerances and Shaft system Alignment.

2.23 FURTHER READING Culbert, I.M., Dhirani, H. and Stone, G.C. (1988). Handbook to Assess Rotating Machine Insulation Condition, Vol. 16, EPRI.

CHAPTER

3

GENERATOR AUXILIARY SYSTEMS

All large generators have auxiliary systems to handle such things as lubricating oil for the thrust and guide bearings, water systems for stator bar direct cooling and supplying air to water heat exchangers, and excitation systems for field current application. Not all generators require all these systems and the requirements depend on the size and nature of the machine. For example, a smaller open ventilated generator does not require cooling water for the stator cooling. This chapter discusses the general nature of the three major auxiliary systems that may be in use in a particular generator: • Oil systems • Stator surface air cooling system • Stator winding direct cooling water system • Excitation systems Each system has numerous variations to accommodate the many different generator configurations that may be found in operation. But regardless of the generator design and which variation of a system is in use, they all individually have the same basic function as described in the first paragraph. Readers are required to review, understand, and apply the safety procedures found in Section 7.2.1.

3.1 OIL SYSTEMS The oil system provides oil for all of the generator bearings (thrust, guide, and oil lift system). The main components of the lubricating oil system consist generally of the oil reservoir, pumps (if equipped), heat exchangers, filters and strainers, centrifuge or purifier, vapor extractor, and various check valves and instrumentation. Handbook of Large Hydro Generators: Operation and Maintenance, First Edition. Glenn Mottershead, Stefano Bomben, Isidor Kerszenbaum, and Geoff Klempner. © 2021 The Institute of Electrical and Electronics Engineers, Inc. Published 2021 by John Wiley & Sons, Inc.

157

158

CHAPTER 3

GENERATOR AUXILIARY SYSTEMS

Normally, station service water (raw water) from a local lake or river is used inside the cooling coils for all oil systems associated with the generator. The water is processed through strainers to ensure debris that can cause problems within the system is removed, thus it is important to check the strainers from time to time and clean them. There are a few different philosophies for designing the main lubricating oil system for the thrust bearing. The main goal in the thrust bearing oil system is to maintain uniform lubrication of the thrust surface, oil cleanliness, and uniform temperature distribution inside the bearing housing. Stratification of the oil inside the thrust bearing housing will cause incorrect oil temperature readings, localized heating of different areas of the bearing, and could cause a bearing wipe. It is very common for the bearing housing to have an internal oil to water cooler strategically placed to remove heat from the oil. The turning of the thrust assembly alone causes the oil to circulate and distribute uniformly inside the housing. Some machines use a motor-driven pump to move oil from the bearing housing to an external cooler assembly and back again. Typical Babbitt thrust bearing shoes are designed as multiple segmented pads that will self-lubricate once the shaft is turning above a required minimum speed. When operating at the required speeds, the thrust bearing runner will pump sufficient quantities of oil over the bearing to balance the outflow of oil leaving the bearing. When operating below the minimum speed, oil will leave the bearing surface faster than the runner can pump it in. In order to protect the thrust bearing from damage during slow rotation while shutting down or starting, a high pressure oil lift system is often used to ensure sufficient lubrication. Sufficient lubrication of the bearing during these periods of slow rotation will prevent wiping of the Babbitt material. The oil lift is an external system that feeds high pressure oil between the bearing shoes and runner plate, thus creating a thin layer of oil to rotate on. It is important to recognize that all of the weight of the rotating rotor assembly is on the thrust bearing so when the machine is at rest, the oil between the thrust shoes and the runner plate (the part that is attached to the rotating mass) is squeezed out, leaving insufficient amounts of oil for the next start up. This is why it is dangerous for a unit to “creep” or rotate extremely slowly without introducing oil into this area. Damage to the thrust bearing is almost certain without lubrication. The oil lift system is also useful during overhauls when the rotor must be rotated manually for whatever reason; it makes rotating much easier. Where oil lift systems are not installed, thrust bearing damage on shutdowns is avoided by minimizing the amount of low speed operation. This is accomplished by applying the rotor brakes to bring the rotor quickly to a full stop from speed typically below 25% synchronous speed. For restarting a generator without oil lift there will be manufacturer’s suggested minimum standstill times for safe restarting, after which hydraulic lifting of the rotor will be required to strip the thrust bearing shoes from the bearing runner plate to re-introduce oil. Lifting the rotor is often done by pumping high pressure oil into the generator’s braking system. The high pressure oil can be supplied by a motorized pump or by

3.1 OIL SYSTEMS

159

a manual hydraulic pump. When the rotor is lifting, it is not uncommon to hear the thrust bearing shoes drop one at a time from the thrust bearing runner plate as oil slowly makes its way back into the shoe surface. Shoes can stick to the runner surface for a long time after the rotor has been lifted and shoe stripping devices have been used to pull downward on the shoe to minimize this time. As mentioned in Chapter 2, an alternative to Babbitt bearings is to use PTFE (Teflon™) which has many benefits and may be more appropriate for a given machine. It is highly recommended the OEM is consulted when converting to this bearing material, to ensure any minor design changes if required are implemented. It should be noted that essentially dry PTFE rubbing against the rotating ring has a friction coefficient of about 0.1. The rotor of a unit supplied with a PTFE-lined thrust bearing, without a high pressure oil injection system, cannot be easily turned by hand. Therefore, service operations, such as inspection and shaft alignment checks, which require the rotor to be turned, requires that the rotor be jacked off the thrust bearing to allow oil to flood the segment surface in order to enable the rotor to be manually rotated. For units using PTFE thrust bearing segment liners, consideration should be given to implementing a permanently installed rotor turning device for use during unit servicing [1]. The generator guide bearings in a vertical generator can be made up of multiple segmented shoes or a shell or sleeve type bearing. In many designs, the guide bearing may be located in the same housing structure and oil reservoir as the thrust bearing [1]. If the guide bearing is separate to the thrust bearing, a segmented shoe bearing will typically have a self-contained oil pot which has all the lubrication necessary for the bearing. The oil is self-pumped by the bearing rotation inside the oil pot to ensure proper lubrication and temperature distribution. For higher speed generators, the oil can be pumped out and cooled by an external cooler and then pumped back into the bearing oil pot. On many older units (pre-1940) the sleeve type guide bearing is typically supplied with a drip through oil system that pumps oil up to the top of the bearing from a lower reservoir and allows it to run down through one or more bearings on the shaft. Generators driven by a variable blade pitch Kaplan turbine assembly will have a separate oil system to power the blade rotation control device in the turbine hub. The oil for the turbine must travel through an oil head arrangement above the generator, then down the generator and turbine shafts to the turbine hub. External oil cooling systems can have full-flow filters and/or strainers for removal of debris from the lube oil. Strainers are generally sized to remove larger debris and filters remove debris in the range of a few microns and larger. They can be mechanical or organic-type filters and strainers. Debris removal is important to reduce the possibility of scoring the bearing Babbitt material or plugging of the oil lines or other small orifices that are critical to the operation of the entire system. Self-contained oil pot systems generally do not have continuous oil filtering and therefore require routine oil sampling to monitor oil cleanliness and quality. It is very important that water is not present in the oil and thus, the monitoring of oil quality and the use of oil level alarms is critical. If an oil-to-water heat

160

CHAPTER 3

GENERATOR AUXILIARY SYSTEMS

exchanger develops a water leak into the oil pot reservoir, the oil level would increase setting off an alarm in a control room to notify operators that action is required. The oil reservoir around a vertical generator’s thrust or guide bearing is often called an oil pot. This oil pot is supported by a bracket and has the primary function of keeping the oil inside where it is lubricating the bearing surfaces. The pot itself is complex, and the detailed design varies greatly from one machine to the next, but there is a common function that can be discussed here. The upper portion of the oil pot will provide the base on which the rotating shaft vapor seals are located as shown in Figure 3.1-1. These seals are typically designed to have a reasonably small clearance to the shaft. Some designs have an air handling arrangement that compliments the seal and is intended to passively collect oil vapor that can pass through the seal clearances. This vapor handling can be in the form of applied air pressure to the interior of a multi-chamber seal, or vapor extraction from a motor-driven vacuum system. It is not always obvious how the original manufacturer had intended to manage vapor from the oil pot, but it is important to determine this intent when troubleshooting oil vapor leakage, since making changes in an attempt to reduce leakage often worsens the problem. Sealing below the oil pot is typically done with a high hat or chimney tube sometimes called oil well tube, which stands well above the oil fill level inside the overhung bearing runner as shown in Figure 3.1-2. Oil vapor and liquid oil can be a problem if the high hat to shaft has misalignment, or air pressure differences across the oil pot, pull vapor into the generator. When oil vapor can not be adequately contained by the original shaft clearance seals consideration can be give to custom shaft contact brush seals designed by some OEM and independent experts.

Figure 3.1-1 Shows oil pot labyrinth seal at the top.

3.2 STATOR SURFACE AIR COOLING SYSTEM

161

Figure 3.1-2 Shows a typical high hat or chimney tube.

3.2 STATOR SURFACE AIR COOLING SYSTEM The surface air coolers are used when the machine is indirectly cooled, that is, the windings do not have water cooled conductors. Rather, the water is passed through cooling tubes/plate fins via a radiator that is placed on the periphery of the stator frame. In some older designs, the coolers may be mounted on separate supports and not directly on the frame. The number of coolers per unit can be as few as one to as many as eight or more, depending on the kilowatt of losses to be extracted, see Figure 3.2-1 [1].

3.2.1

Construction

The surface air cooler has three main parts, the main frame, the water box, and the extruded cooling tubes with fins or plate fins to perform the heat exchange. The frame is usually fabricated welded steel in the shape of a rectangle that will form the basis for attaching the water boxes at both ends and fastening the cooler to the stator frame. The cooler frame can be galvanized, stainless steel, or treated with rust inhibiting epoxy paint given the nature of the environment it is in. The water box, as shown in Figure 3.2-2, is where the cold water enters and the warm water

162

CHAPTER 3

GENERATOR AUXILIARY SYSTEMS

Figure 3.2-1 Typical tube type surface air cooler.

Figure 3.2-2 Shows typical water box on one side of surface air cooler.

3.2 STATOR SURFACE AIR COOLING SYSTEM

163

exits the surface air cooler. The cooling tubes terminate inside the water box and carry the cold water across the hot air passing by the cooling tubes or plate fins. Depending on the velocity of the water, pH, and the amount of silt in the water, the water box will erode accordingly. To ensure maximum life of the water box, a good choice is stainless steel. However, consultation with the surface air cooler manufacturer about the water conditions as mentioned above may yield other alternatives alloys. In the early days of generator manufacturing, the cooling tubes were simply just made of extruded copper. At that time they were manufactured this way, because the heat transport properties of copper is efficient at a relatively low cost. As technology progressed, more advanced alloys were used such as 90/10 Cu/Ni tube which improved the strength and durability, and has better anti-fouling and corrosion resistant properties than copper alone [2]. One very important factor in choosing the material for the cooling tube is to consider microbiologically induced corrosion or MIC for short. Very tiny micro-organisms in the water attach to the cooling tubes of the cooler especially during stagnant water periods when the machine is not running. There are many factors that determine how much MIC corrosion takes place and at what pace the erosion occurs. It is a good idea to get a water sample every two years from the raw water system and have it analyzed to determine what micro-organisms are present in the water and what materials would best resist these living things. Figure 3.2-3 shows an example of damage from a failed surface air cooler due to MIC. This cooling tube was comprised of 90/10 Cu Ni, had a flow rate of 1.5 m/s (5 ft/s), and was in service for less than 5 years, and had stagnant water for long periods of time during maintenance outages ranging from a few weeks to a few months. During maintenance periods, it is a good idea to flush the coolers periodically if water is maintained inside or drain the coolers completely. There is much literature written on the topic of MIC, and the reader is encouraged to research further if desired [3].

3.2.2

Function

The purpose of the surface air cooler is to dissipate heat that is generated by all the losses in the generator while in service with the exception of those from the bearings. These losses are mainly, but not limited to, stator core loss, I2R losses from the rotor and stator windings, and with the windage losses at rated temperature and MVA. The generator hot air coming into the cooler from the stator frame passes through the cooling tubes/plate fins and becomes generator cold air that is then recirculated into the machine. The generator hot air coming into the cooler is a matter of manufacturer design, typically between 55 and 70 C at rated load. The air exiting the cooler is called generator cold air and is limited to 40 C by design. This 40 C maximum cold air temperature is used as the intake air into the machine to be reheated to 55–70 C after absorbing heat from the core and coils and sent back out to be cooled back to 40 C. The cooler designer must take into account the amount of air the machine is circulating, the pressure drop across the cooler, water velocity through the cooler, and maximum water temperature into the cooler in order to achieve this 40 C limit. This 40 C cold air has been a standard for almost a

164

CHAPTER 3

GENERATOR AUXILIARY SYSTEMS

Outer cooling fins

Inner tube

Figure 3.2-3 Cross section of a heat exchanger tube, showing the inner tube surface pit and the corrosion damage on the fin. Magnification is 7×.

century and is followed by IEEE and IEC standards. In some northern climates, where the maximum water temperature does not get much above 15 C, the air out of the coolers may be limited to 30 C rather than the 40 C standard [1]. Surface air cooler design is usually based on achieving a specific differential temperature between cold air out of the cooler to cold water into the cooler. Typical temperature difference is 10–15 C. As this differential gets smaller, the size of the cooler generally needs to get larger and would be more expensive. Cooler designs also must take into account the probability of future reduced heat transfer and cooler performance due to contamination, whether on the water side (inside the tubes) or on the air side (cooler fins). Extra cooler fin area is typically part of the original cooler design to account for the possibility of future developing contamination and lack of cleanliness fouling factor) [1].

3.3 BEARING COOLING COILS AND WATER SUPPLY

3.2.3

165

Replacement Surface Air Coolers

It is inevitable at some point in time that the coolers will need to be replaced. It is important to remember when replacing coolers in kind, that all the losses as outlined in the machine’s original test data are specified to the cooler supplier including a 10% fouling factor for good measure on the cooler design. The test data is what the manufacturer would have produced to the owner of the machine to prove the guarantees as cited in the purchase agreement. Equally important is when uprating the machine, the additional losses are specified to the supplier so the generator does not overheat when at the new rating. Again, always check with the OEM or qualified expert when uprating the machine to ensure it is capable of the new rating, particularly the mechanical increase in power.

3.2.4

Maintenance

The surface air coolers, typically, are using raw river water that has gone through some sort of strainer system to remove debris. It is possible for debris to get past the strainers particularly if they become damaged while in service. It is quite common to find pieces of tree branch and the like when looking inside the tubes of the cooler. In order to clean the cooler thoroughly, removal of the water box covers at both ends is required. Depending on the material used for the water box, water velocity, water chemistry, and how long the cooler has been in service will determine how much work may be required to clean the internal surface of the water box. Stainless steel is more resistant to water abrasion due to silt mixed in the water and corrosion than, say, a galvanized or treated steel surface. Even epoxy-coated surfaces fall victim to this abrasion and corrosion, although they are cheaper to manufacture initially. In some cases, where erosion is prevalent, titanium material is selected for tube material [1]. Particular attention should be paid to the inlet and outlet water separator plate at the one end of the cooler, ensure it has not separated from the water box as this will mix inlet and outlet water before it has had a chance to go through the cooling tubes. Hard water deposits and corrosion should be carefully removed with water jet or dry ice blast. The tubes should be cleaned with a bristle brush and water from one end to another. Pressure washers may also be used but caution should be used to ensure no damage is done to the tubes when cleaning at elevated pressures. During re-assembly, new gaskets should be used and when completed, a pressure test should be performed to ensure there are no leaks. Consult with the cooler manual or the supplier for a test protocol after maintenance is complete.

3.3 BEARING COOLING COILS AND WATER SUPPLY The cooling water for the generator bearings is typically taken from the main station service water in the plant. This water is the same as the surface air cooling water and thus has strainers to remove any debris that would be present. The

166

CHAPTER 3

GENERATOR AUXILIARY SYSTEMS

Figure 3.3-1 Thrust bearing cooling coil.

cooling tube for the thrust and other bearings on the generator can be made from the same material and configuration as the surface air cooling tubes see Figure 3.3-1. Once again, keeping in mind that MIC can cause leaks in the system, these coils should be designed so that leak detection is available. Water in the bearing oil is highly undesirable, as failure of the bearing will eventually result. The thrust bearing in the generator that is lubricated and cooled using oil, utilizes two main methods to cool the oil. The first method is to put the cooling coil right inside the oil pot where the oil is circulating, and circulate the station service water inside the cooling tube as shown in Figure 3.3-2. In this particular case, the cooling tubes are just simple copper with no fins (pre-1950), placed in a very precise location inside the oil pot in order to cool the oil as evenly as possible and avoid temperature stratification of the oil. After this time period, installations started having fin type tubes, and in the modern day, this is the definitely the case. Temperature stratification can occur if the oil circulation around the cooling coil is not optimal, thus stranding hot oil and causing bearing failures, if the stratification is substantial enough. This can be the case when cooling coils are replaced or their location changed within the oil pot during an overhaul. It is very important to remember replacing the cooling coils in kind and put them back in their original location. If cooling coil modifications are planned for the overhaul, it is recommended the OEM is consulted. The second method is pumping the oil to an external cooling system close to the generator, using a physical pump or by the natural rotational movement of the bearing. The oil is cooled and then returned to the oil pot for the next cycle of heat transfer. In this method, the oil exit and entrance points are strategically located by the original designer to take the hot oil out and feed the cold oil back in using natural mixing within the oil pot. Again, during an overhaul, it is important not to relocate these inlet and outlet points on the oil pot. If a bearing heating issue is

3.4 STATOR WINDING DIRECT COOLING WATER SYSTEM

167

Figure 3.3-2 Simple copper tubing used for cooling of the oil inside an oil pot.

known, it is recommended the OEM be consulted for resolution. Stratification of oil can occur in this scenario as well. This setup has a few more risks associated with it versus the internal cooling coil such as electrical or mechanical failure of the oil pump and the added risk of leaks due to the oil travel to and from the oil pot. Each bearing or sets of bearings typically have a flow meter to verify water flow. Bearings are not as forgiving as a surface air cooler in that water flow restriction or lack of water flow all together has a more immediate effect on this equipment. Thrust bearings, for example, due to the high compressive load from the rotor combined with the rotational speed, generate large amounts of heat that must be dissipated continuously without fail. In this case, cooling water flow failure alarms and subsequent trips can be added to the protection systems of the generator to avoid bearing wipes if using Babbitt.

3.4 STATOR WINDING DIRECT COOLING WATER SYSTEM The stator cooling water system is used to provide a source of pure demineralized water to the generator stator winding for direct cooling of the stator winding. Most systems are provided as package units, mounted on a single platform, which includes all of the system components which are generally made from stainless steel or copper.

3.4.1

System Components and Functions

An example of a pure water system, which is installed separately from the generator, is shown in Figure 3.4-1 and is comprised of the following main components [1]:

168

CHAPTER 3

GENERATOR AUXILIARY SYSTEMS

(7) (8) (7) (9) (6)

(5) ( 11 )

H+ OH– ( 12 )

(4) M2

( 10 )

P2

M

(1)

(2)

M1

(3)

P1

M Cold pure water pipes – DN 150

N2 pipes

Hot pure water pipes – DN 150

Bypass valve

Auxiliary pipes – DN 80

N2 pressure reducing valve

Auxiliary pipes – 1/2”

Strainer

Figure 3.4-1 Simplified Schematic of the pure water system piping. Source: Courtesy of Voith.

1. Centrifugal pumps with nonreturn valves 2. Heat Exchangers 3. Three-way motor valve 4. Mechanical filter 5. Strainer 6. Stator winding bypass 7. Stator winding manifolds with PTFE hoses 8. 9. 10. 11. 12.

Stator winding Expansion tank Nitrogen supply system Water treatment circuit – ion exchange deionizer Water treatment circuit – Alkalization unit

3.4 STATOR WINDING DIRECT COOLING WATER SYSTEM

169

Several built-in transducers distributed along the cooling circuit continuously monitor the following [1]: • Water volume flow • Water pressure • Water differential pressure at stator winding • Temperatures • Conductivities The water coming from the stator winding circulates through the water pump, passes through one of the heat exchangers, through the mechanical filter and finally enters the stator winding again, this is called the main circuit. A bypass loop that is parallel to the main circuit is designated as the treatment circuit. This circuit consists mainly of the ion exchange deionizer, one alkalization unit, a conductivity sensor, and flow meter. An additional bypass is located outside of the stator winding inlet and outlet flanges. The bypass serves as a return loop for commissioning the system. The bypass is also required to treat the demineralized water in the external section before the stator winding is ready to be filled with pure water with the specified chemistry (conductivity, oxygen content, and pH). Variations in water volume caused by temperature changes are managed by the expansion tank [1]. 3.4.1.1 Centrifugal Pumps (1) The water is permanently circulated in this closed-loop system. To ensure uninterrupted generator operation, two 100% capacity pumps are provided with one in operation and one always on standby which cuts in automatically if needed. Water backflow is prevented using the nonreturn valves and shut-off valves are fitted for isolation purposes [1]. 3.4.1.2 Heat Exchanger (2) Two 100% capacity heat exchangers are provided but only one is required during operation. The water in the heat exchanger that is not being used is maintained to ensure the conductivity is within specification if a changeover is required. Stand still corrosion in the nonactive heat exchanger is also prevented [1]. 3.4.1.3 3-Way Motor Valve (3) This valve is installed at the inlet pipe of the heat exchangers. The purpose of this valve is to mix hot and cold water to set the stator winding inlet temperature to the specified value. This is done using PLC control [1]. 3.4.1.4 Mechanical Filter and Strainer (4+5) In order to prevent clogging of the stator winding cooling ducts by impurities, a 100% capacity mechanical filter is installed in the main circuit of the water system. A strainer located parallel to the filter will be activated in case the mechanical filter has to be serviced (replacement of the 5 μm filter cartridges). The water in the

170

CHAPTER 3

GENERATOR AUXILIARY SYSTEMS

nonactive strainer is maintained to avoid conductivity changes in the case of a filter change over [1]. 3.4.1.5 Expansion Tank (9) This tank is required as the water volume of the system changes as the temperature of the stator winding varies. It is used as storage and compensation of water in the system. A pressurized nitrogen blanket (200 mbar) is introduced above the water to avoid water to air contact [1]. 3.4.1.6 Nitrogen Supply System (10) The 99.99% pure nitrogen supply is maintained using a commercially available tank. This supply system is comprised of a nitrogen cylinder, pressure reducing valve, spring-loaded relief valve, and a water condensation trap [1]. 3.4.1.7 Water Treatment Circuit: Ione Exchange Deionizer (11) Since the water is in direct contact with the hollow copper conductor surfaces on the high voltage stator bar as well as the grounded water manifolds, the conductivity must be kept at a low level. During operation, copper ions are released from the stator winding and would gradually increase the conductivity. The pH value is maintained at 8.8–9.0, which is equivalent to a conductivity of approximately 2 μ S/cm. In order to achieve this, a water flow of 100 L/h continuously passes through the ion exchange deionizer, which is installed in the bypass circuit of the system. For full deionization, the water has to remain in the ion exchange deionizer for a certain duration and thus the water flow rate has to be adjusted and monitored [1]. 3.4.1.8 Water Treatment Circuit: Alkalization Unit (12) Corrosion of the stator winding copper cannot be completely avoided even though low oxygen content water was used at the time of commissioning. This corrosion can be reduced to a negligible level by alkalizing the water. This is accomplished by keeping the pH between 8.8 and 9 by continuously injecting diluted sodium hydroxide solution (5 g of NaOH/dm3) into the ion exchanger outlet pipe of the water treatment bypass circuit. The ion exchanger (with H+ cation exchangers and OH− anion exchangers) is designed for continuous service. It removed all copper, iron, chlorine, carbon dioxide ions, etc., from the water. However, it also eliminates all Na+ ions from the water and must be compensated by the continuous feeding of the diluted sodium hydroxide solution [1]. The relationship between conductivity and pH under ideal conditions is shown in Figure 3.4-2 [1].

3.5 EXCITATION SYSTEMS

171

Conductivity as a function of pH in water at 18 °C

Conductivity (ms / cm)

10

NaOH 1

HCI 0.1

0.01 6

7

8

9

10

pH

Figure 3.4-2 Relationship between conductivity and pH under ideal conditions. Source: Courtesy of Voith.

3.5 EXCITATION SYSTEMS For the machine to actually function as a generator, magnetizing current or excitation must be supplied to the rotor winding by the excitation system. The system is designed to control the applied voltage and the field current to the rotor, which in turn gives control of the generator terminal voltage. Subsequently, this is what provides reactive power and power factor control between the generator and the system. Voltage requirements range typically from 125 to 250 V DC for the larger generators. DC field currents as high as 3000 A or more can be found on the largest hydro generators. Excitation response time must be fast so that the automatic voltage regulator can control the generator during system disturbances or transients for which rapid changes of excitation are necessary. “Field forcing” is the term generally used for this mode of operation and requires the exciter to be capable of forcing the field voltage from two to three times the rated value. Therefore, the rotor winding must be insulated for these voltage levels.

3.5.1

Types of Excitation Systems

The excitation system types are as follows: 1. Rotating 2. Brushless with diodes (shaft mounted)

172

CHAPTER 3

GENERATOR AUXILIARY SYSTEMS

3. Brushless with Thyristor (shaft mounted) 4. Static pilot 5. Static main 3.5.1.1 Rotating Within the family of rotating exciters, there are numerous types that can be found operating on all types and sizes of generators. The basic kinds of rotating exciters are motor- or shaft-driven and separately self-excited or bus-fed systems. The subject of excitation systems is a book in itself, and it is not our intention to focus on exciters in this book. They are discussed in brief as an auxiliary system to the generator. Older rotating excitation systems consisted of a rotating pilot exciter at the very top of the generator, which would then feed into the main exciter directly below it on the same shaft which would then power the main rotor of the generator as shown in Figure 3.5-1. In this particular case, the pilot exciter is 7 kW compound wound, and the main exciter is 145 kW shunt wound separately excited. The compound

Figure 3.5-1 Shows pilot and main rotating exciters.

3.5 EXCITATION SYSTEMS

173

wound pilot exciter was the most common form of constant voltage excitation. These pilot exciters are invariably a 125 or 250 V machines with a self-excited shunt field and a series excited field, adjusted to give substantially flat compounding. Thus, regardless of the load on the pilot exciter, the magnitude of its terminal voltage is practically constant. A rheostat, either under the control of a voltage regulator or under manual control, is connected in series with the output circuit of the pilot exciter to regulate the voltage applied to the field of the main exciter as shown in Figure 3.5-2 [4]. Both exciters have a full set of brushes to maintain and areas to be kept clean which requires much attention. The shaft-driven rotating excitation system has been the most widely used excitation source in the past for large hydro generators. The most basic configuration is a stationary field as shown in Figure 3.5-3 and a rotating armature as shown in Figure 3.5-4. This configuration of stationary and rotating is used for the pilot and main excitation components, but the pilot is just a smaller version of the main exciter. Main field Exciter field

+ –

Static pilot exciter

AC generator

Exciter

Rheostat

Potential transformer

V

Figure 3.5-2 Schematic of a shaft-driven, rotating pilot and main excitation system.

Figure 3.5-3 Typical stator or stationary field.

174

CHAPTER 3

GENERATOR AUXILIARY SYSTEMS

Figure 3.5-4 Typical rotating exciter armature.

For the rest of the excitation options, brushless with diodes, brushless with thyristors, and static pilot and static main, the reader is referred to Chapter 12 as they are discussed in detail there.

3.6 EXCITATION SYSTEM PERFORMANCE CHARACTERISTICS The principal function of the excitation system is to furnish DC power (direct current and voltage) to the generator field, creating the magnetic field. The excitation system also provides control and protective equipment that regulates the generator electrical output. Excitation voltage is a key factor in controlling generator output. One desirable characteristic of an excitation system is its ability to produce high levels of excitation voltage (the ceiling) very rapidly following a change in terminal voltage. IEEE 421.2 defines a high initial response (HIR) excitation system as one that reaches 95% of the specified ceiling voltage in 0.1 seconds or less. For units tied into a power system grid, such quick action to restore power system conditions reduces the tendency for loss of synchronization [5]. The other important performance feature of an excitation system is the level or amount of ceiling voltage it can achieve. The response ratio is the term for quantifying the forcing or ceiling voltage available from the exciter. The response ratio is the average rate of rise in exciter voltage for the first one half second after change

3.7 REFERENCES

175

initiation, divided by the rated generator field voltage. Thus, it is expressed in terms of per unit (pu) of rated field voltage. A standard level of exciter response ratio is 0.5 pu. This level has been found to be adequate for the large majority of industrial and utility applications. Power system studies have shown that some applications benefit from higher response ratios or more powerful exciters. In general, it can be observed that conventional rotating exciters, such as the classical rotating and the brushless type rotating diode exciters have slower response times due to the time constants of the rotating magnetic components. In fast acting brushless rotating thyristor exciters and full static exciters, maximum exciter output is available almost instantaneously by signaling the controlling thyristors to provide full forcing. The machine owner should check with the system operator when overhauling an excitation system to ensure the existing system response and performance is still adequate. There is more discussion in Chapter 12 on exciter performance.

3.7 REFERENCES 1. Tavares M. (2010). Grand Coulee Dam, G22, G23, and G24, Pure Water System (PWS) General Description, Revision 1.0, Voith Hydro Holding Gmbh & Co, Corporate Technology VHEC, Heidenheim. 2. Nickel Institute (1982). The story of nickel. http://www.nickelinstitute.org/~/Media/ Files/Technical/Literature/Copper_NickelAlloys_PropertiesandApplications_12007/_. pdf, Retrieved 14 May 2017 from nickelinstitute.org: http://www.nickelinstitute.org (accessed 19 March 2020). 3. Goszczynski, G. (2002). Examination of the G17 Heat Exchanger Tube from Sir Adam Beck II Generating Station, Toronto, ON, Kinectrics Inc. 4. Central Station Engineers (1964). Electrical Transmission and Distribution Reference Book, Pittsburg, PA, Westinghouse Corporation. 5. IEEE (2014). 421.2-2014: IEEE Guide for Identification, Testing, and Evaluation of the Dynamic Performance of Excitation Control Systems. New York, IEEE.

CHAPTER

4

OPERATION AND CONTROL

Chapters 2 and 3 describe the very complex construction aspects of a hydro generator and the peripheral equipment required for its operation. One can thus imagine that this is a very costly system, in addition to being a very critical component in a power plant. Given the importance of this machine, the need to operate it reliably for many years, and the initial large capital expenditure, it goes without saying that a large effort must go into preparing a comprehensive and detailed purchasing specification. Assuming this has been done, and a well-designed and wellmanufactured generator has been delivered, the long-term availability and reliability of the unit will depend greatly on how the machine is operated and maintained. A well-designed and well-manufactured unit can become compromised by serious operational challenges, some of which, discussed below, are outside the control of the plant operators. However, in such a case, several ameliorating actions can be taken by a proactive station management. Other challenges are directly related to how the operators run the unit. All these operational aspects are discussed in this chapter; the very important maintenance issues are covered starting with Chapter 7. Readers are required to review, understand, and apply the safety procedures found in Section 7.2.1.

4.1 BASIC OPERATING PARAMETERS All generators are designed such that they have a rating. The rating of the machine is a series of parameters that describe the generator in engineering terms. These parameters tell the owner about the available power output of the generator and its capability with regard to electrical, thermal, and mechanical limits. With enough experience, the trained person can also often infer other information about such things as the generator size and basic construction features. Like any industrial apparatus, large generators are specified, designed, and constructed to meet a number of requirements. These requirements are predicated

Handbook of Large Hydro Generators: Operation and Maintenance, First Edition. Glenn Mottershead, Stefano Bomben, Isidor Kerszenbaum, and Geoff Klempner. © 2021 The Institute of Electrical and Electronics Engineers, Inc. Published 2021 by John Wiley & Sons, Inc.

177

178

CHAPTER 4

OPERATION AND CONTROL

on customer needs, as well as mandatory industry standards and “best industry practice” guidelines. The requirements are given in the form of performance parameters and dimensional standards. The performance parameters of a large generator are defined in a number of standards. In the United States, the leading standard defining hydro generator performance variables is the Ref. [1] among other IEEE standards that outline operation and maintenance guidelines. In other countries, standards such as IEC may also apply. In the following subsections, a definition and, when required, an explanation of all performance parameters is given.

4.1.1

Machine Rating

A machine is described by giving it a rating at the generator’s capability point of maximum continuous power output at full rated lagging power factor. The terms generally used to provide the rating are as shown in Table 4.1-1: Each of these parameters signifies a finite design quantity that describes a certain capability or limitation of the generator. In some cases, they also provide operating limits that, if exceeded, will cause excess stress in the generator (mechanical, thermal, or electrical) on one or more of its components. All large generators are designed with these parameters in mind, and they are all reflected in the design standards for generators [1]. There are specific ranges for the abovementioned parameters, and these are outlined in the design standards and discussed in documents regarding good operating practice for hydro generators [2]. The ratings of large generators have increased dramatically over the years as designers have learned to optimize newer and better materials in their designs. As discussed in Chapter 2, the rate of increase of generator ratings over the years has been logarithmic.

TABLE 4.1-1 Rating quantities for generators

Apparent power Real power Reactive power Power factor Stator terminal voltage (±5%) Stator current Field voltage Field current Frequency Speed/Overspeed capability Stator/Field winding insulation class Stator/Rotor temperature rise

MVA MW MVAR PF Vt Ia Vf If Hz RPM/RPM Class (F,H)/Class (F,H) C/ C rise

4.1 BASIC OPERATING PARAMETERS

179

Figure 4.1-1 Typical nameplate for a hydro generator.

Hydro generators are presently being built with ratings up to 944.5 MVA at the time of this writing. An example of a nameplate that may be found on a large generator is shown in Figure 4.1-1.

4.1.2

Apparent Power

Apparent power refers to the rating of the generator. In large generators, it is usually given in units of MVA, although it may also be stated in kVA. Although machines are commonly talked about in terms of real power (MW or kW), it is the apparent power that best describes the rating. This is because the product of the voltage and the current (MVA) largely determines the physical size of a machine. In a three-phase power system, the MVA is given by the following expression: MVA = √3 Generator line current in kA × Line voltage in kV Alternatively, MVA = 3 Generator s line current in kA × Phase voltage in kV Also, MVA =

MW Power factor

(4.1)

Using the preceding expressions, one can find the maximum current that can be supplied by the generator at a given terminal voltage. This is important for sizing the conductors or busses that carry the power into the system, as well as for setting protection relays. For a theoretical explanation about the origins of apparent power, see Section 1.3.

180

4.1.3

CHAPTER 4

OPERATION AND CONTROL

Power Factor

It was shown in Section 1.3 that the power factor is a measure of the angle between the current and the voltage in a particular branch or a circuit. In mathematical terms, the power factor is the cosine of that angle. Within the context of a generator connected to a system, the power factor describes the existing angle between the voltage at the terminals of the generator (Vt), and the current flowing through those terminals (Il). In the workings of generators, by definition, the angle between the current and the voltage is deemed positive when the current lags the voltage, and vice versa, it is defined as negative when the current leads the voltage. Therefore, the power factor is used to describe the generator as operating in the “lagging” or “leading” power factor range. A positive power factor indicates that the unit is operating in the lagging region, it is generating VARs. A negative power factor indicates that the unit is operating in the leading region, and it is absorbing VARs from the system. Additional terms for describing if the unit is producing or consuming VARs are “overexcited or inductive” for lagging power factor operation, and “under excited or capacitive” for leading power factor operation. Unity power factor refers to a power factor of 1.0 where no reactive power is being produced or absorbed, the machine is only producing real power in MW. It is common for generator operators to say the unit is “boosting” or “bucking” VARs. Boosting in this context is synonymous with overexcited or inductive, and bucking means underexcited or capacitive. These different terms for defining the same mode of operation can be confusing to the uninitiated. A simple way out is just to remember that if the generator is overexcited (i.e. if field current is increased), it will export more VARs into the system. On the other hand, if it is underexcited (i.e. if the field current is reduced), the generator will absorb VARs from the system in order to maintain the required airgap flux density. The rated power factor is the operating point that maximizes both watts and VARs delivered, and it is a design variable. Increasing excitation from that point onward requires the unit to significantly reduce the active output (watts), in order to remain within the allowable operating region (more about that later). For most hydro generators, the rated power factor is in the range of 0.80–0.90 lagging (overexcited). The power factor (actually reflecting the flow of reactive power) has a big influence on the power system in that it can change the system’s voltage. The change in voltage in turn affects the ability of the system to carry the required levels of power and, consequently, its stability. To illustrate this important concept, a very elementary example is offered. Figure 4.1-2 depicts a generator supplying a single radial circuit, with a load at the end of it. Let us assume two cases: Case 4.1 has a line impedance of 1 + j5 Ω and a load of 5 Ω, and Case 4.2 has the same line impedance but the load has, in addition to the 5 Ω resistance, a 5 Ω reactance (a reactance is denoted by preceding its value with the letter j). Assume that the generator voltage is maintained at 100 V in both cases.

4.1 BASIC OPERATING PARAMETERS

181

Generator Line

Load

~ Current (I)

VL

VG = 100 v

Figure 4.1-2 Schematic representation of a generator feeding a load through a line.

Case 4.1 The impedance of the line is 1 + j5 Ω and the load is equal to 5 Ω. The magnitude of the current delivered by the generator is then I=

100 √ 6 + 52 2

= 12 8 A

The magnitude of the voltage at the load terminals is V L = 12 8 A × 5 Ω = 64 V And the power delivered to the load is P = 12 82 × 5 Ω = 819 W The load power factor (PF) is: PF = MW/MVA = 819/(64 × 12.8) = 1 Case 4.2 In this case, the line impedance has not changed, but the load now has an additional inductive reactance of 5 Ω. The magnitude of the current delivered by the generator is now I=

100 √ 6 + 102 2

= 8 57 A

The magnitude of the voltage at the load terminals is V L = 8 57 A × √ 52 + 52 Ω = 60 6 V And the power delivered to the load is P = 8 572 × 5 Ω = 367 W The load power factor is PF =

MW = 367 60 6 × 8 57 = 0 71 MVA

182

CHAPTER 4

OPERATION AND CONTROL

As a result of the addition of a load reactance, the power factor of the load has been reduced from unity to 0.71, the voltage at the load terminals dropped 5.6%, and the real power delivered to the load is reduced to approximately 45% of the original value. This simple exercise illustrates the significant impact on a system of an addition of inductive reactance (i.e. in reducing the power factor). Increasing the excitation of the generator in the simple case of this example would increase the generator terminal voltage, driving the load voltage higher and somewhat compensating for the voltage drop introduced by the reduced power factor.

4.1.4

Real Power

The rated power (in MW) of the generator is the product of the rated apparent power (in MVA) and the rated power factor. The turbine determines the rated power of the hydro generator, as a whole unit. The rated power of the generator is often specified and designed to be somewhat higher than that of the turbine to take advantage of additional output that may become available in the future from the turbine. This parameter is measured and monitored to keep track of the load point of the machine and allow the operator to control the operation of the generator. MW overloading of the generator is a serious concern. MW overload means that the stator current’s limit has probably been exceeded, and this will affect the condition of the stator winding, more so if the maximum stator terminal voltage has been exceeded. Further, the mechanical limit of the generator drive train may also have been exceeded and this could affect long-term performance of the generator. It is never a good idea to exceed any mechanical or electrical limits as given by the manufacturer even for a short period of time without sanction from the manufacturer. Transient MW events from the system or internally in the machine will also show up as transients in the stator current and/or terminal voltage.

4.1.5

Terminal Voltage

The rated or nominal voltage of a three-phase generator is defined as the line-toline terminal voltage at which the generator is designed to operate continuously. The rated voltage of large hydro generators is normally in the range of 12 000–13 800 V, while for smaller machines, the voltage range is typically 2 200–6 600 V. Generators designed to the Ref. [1] standard are able to operate at 5% above or below rated voltage at rated MVA, continuously. When special requirements of a power system dictate the need for a wider terminal voltage range, the manufacturer has to account for this in the generator design and produce a larger and more expensive machine. The case where this type of variation is required depends on the location and requirements for interaction between the generator and the power system.

4.1 BASIC OPERATING PARAMETERS

183

Monitoring of the generator terminal voltage is also critical and is done on a per-phase basis. It is required to ensure that there is voltage balance at all times, to avoid negative sequence type heating effects, and it is most critical during synchronizing of the generator to the system. The terminal voltage of the generator must be matched in magnitude, phase, and frequency to that of the system voltage before closing the main generator breakers. This is to ensure minimum transient currents when closing the breakers and connecting to the system, and to deter faulty synchronization.

4.1.6

Stator Current

Stator current capability in large hydro generators depends largely on the type of machine in question. In the simplest machines (i.e. the indirectly air cooled generator), the capability of the stator winding defines the rated stator current.

4.1.7

Field Voltage

Increasing the field voltage increases the field current in proportion to the rotor winding resistance. The field voltage is monitored and alarm and trip setting are usually in place for loss of field or over excitation. It is also used to calculate rotor winding resistance and, subsequently, the rotor winding average and hotspot temperature rises during commissioning for manufacturer guarantees and when a heat run test is being performed.

4.1.8

Field Current

The capability of the rotor winding is generally determined by the field current at the rated apparent power, the rated power factor, and the rated terminal voltage. All of the capability considerations described for stator windings apply to the rotor winding as well. Increasing the field current will • Augment the MVARs exported to the system • Increase armature (stator) current if the unit is already in the boost or overexcited region • Increase the potential at the machine terminals A simplified capability curve for internal utility use is shown in Figure 4.1-3.

4.1.9

Speed

Unlike an induction machine, the synchronous generator can only generate and deliver power at one speed, which is called the synchronous speed. That unique speed is related to the system’s frequency and the number of poles of the machine, by the following Equation (4.2):

184

CHAPTER 4

OPERATION AND CONTROL

160 140 120 13.1 kV 13.8 kV 14.5 kV

Rotor current limit

100

0.65 pf

80

Stator current limit

60 0.9 pf Reactive power (MVAr)

40 20 Active power (MW) 0

0

10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

–20 –40

0.9 pf

–60 –80 Static exciter under excitation limiter –100 –120

Exciter minimum excitation

–140 –160

Loss of excitation protection

Mechanical limit 138 MW

Figure 4.1-3 Typical capability curve for internal utility use.

Synchronous speed RPM = 120 ×

4.1.10

System frequency Hz Number of poles

(4.2)

Short Circuit Ratio

The short circuit ratio (SCR) is defined as the ratio of the field current required to produce rated terminal voltage on the open circuit condition to the field current required to produce rated stator current on sustained three-phase short circuit, with

4.1 BASIC OPERATING PARAMETERS

185

the machine operating at rated speed. During operation, to maintain constant voltage for a given change in load, the change in excitation varies inversely as the SCR. This means that a generator with a lower SCR requires a greater change in excitation than a machine having a higher SCR, for the same load change. The inherent stability of a generator in a power system is partly determined by its short circuit ratio, which is a measure of the relative influence of the field winding versus the stator winding on the level of useful magnetic flux in the generator. The higher the SCR, the less influence the changes in stator current have on the flux level and the more stable the machine tends to be. But the ratio will also be larger for the same apparent power rating and less efficient. However, machines with higher SCRs are not necessarily the ones showing higher stability in a particular setting. There are other important factors such as the speed of response of the voltage regulator and excitation systems, match between the turbine and generator time constants, control functions, and the combined inertia of turbine and generator. A typical short circuit ratio range for hydro generators is 0.8–1.6 as calculated using Equation (4.3). (4.3)

SCR = 1 X ds where Xds is the saturated direct axis synchronous reactance.

4.1.11

Volts Per Hertz

The term “volts per hertz” has been borrowed from the operation of transformers. In transformers, the fundamental voltage equation is given by Equation (4.4): V = 4 44f Bmax × Area of core × Number of turns

(4.4)

where Bmax, is the vector magnitude of the flux density in the core of the transformer. By rearranging the variables, the following expression is obtained as shown in Equation (4.5): V Hz =

V = 4 44Bmax × Area of core × Number of turns f

(4.5)

Or alternatively as shown in Equation (4.6), Bmax T = constant ×

V f

(4.6)

Or in another notation as shown in Equation (4.7) Bmax

V Hz

The last equation indicates that the maximum flux density in the core of a transformer is proportional to the terminal voltage divided by the frequency of the supply voltage. This ratio is known as V/Hz.

186

CHAPTER 4

OPERATION AND CONTROL

A very similar set of equations can be written for the armature of an alternating current machine. In this case, the constant includes winding parameters such as winding pitch and distribution factors. However, the end result is the same: in the armature of an electrical alternate current machine, the maximum core flux density is proportional to the terminal voltage divided by the supply frequency (or V/Hz). In machines, as well as in transformers, the operating point of the voltage is such that for the given rated frequency, the flux density is just below the knee of the saturation point. Increasing the volts per turn in the machine (or transformer) raises the flux density above the knee of the saturation curve as shown in Figure 4.1-4. Consequently, large magnetization currents are produced, as well as increases in the core loss, due to the bigger hysteresis loop created as shown in Figure 4.1-5 [3]. Additionally, large harmonic eddy currents are developed in the core and other metallic components. All these result in substantial increases in core and copper losses, and excessive temperature rises in both core and windings. If not controlled, this condition can lead to loss of the core inter laminar insulation, as well as loss of life of the winding insulation. In fact, if a unit becomes excessively overfluxed (i.e. the maximum V/Hz has been exceeded) failure of the core may result after some time of operation, although this is very rare for a hydro generator. The manufacturer can supply the V/Hz curves upon request as shown in Figure 4.1-6. Reference [1] states that generators are normally designed to operate at rated outputs of up to 105% of rated voltage. IEEE C57 standards for transformers state the same percentage for rated loads and up to 110% of rated voltage at no load. In Bmax

ΔBmax Bmax Rated

Operating point

ΔImag

Imag Rated

Magnetizing current

Figure 4.1-4 Typical saturation curve for transformers and generators.

4.1 BASIC OPERATING PARAMETERS

187

B (T) Hysteresis loop for rated V/Hz

Hysteresis loop for increased V/Hz

H (A/m)

Area of additional hysteresis losses

Figure 4.1-5 Hysteresis losses under normal and abnormal conditions.

1.25

V/Hz (PU)

1.20

1.15

1.10

1.05

1.00

1

10

100

1000

Time (seconds)

10 000

100 000

Date: 22 February 2019 19A1079KF

Figure 4.1-6 Manufacturer issued V/Hz curve.

practice, the operator should make sure (by consulting vendor manuals and pertinent standards) that the machine remains below limits that may affect the integrity of both the generator and the unit transformer. For operation of synchronous machines at other than rated frequencies, refer to the manufacturer.

188

CHAPTER 4

OPERATION AND CONTROL

4.2 OPERATING MODES 4.2.1

Shutdown

Shutdown mode refers to the time when the generator is offline and not connected to the system. It also implies that the generator is at zero speed, with the main generator and field breakers open. Therefore, there is no energy flowing to or from the generator.

4.2.2

Field Applied Offline (Open Circuit)

The condition of the generator when the field is applied but the machine is not connected to the system is referred to as the open circuit condition. At open circuit, if the generator is spinning at its rated speed and the field current magnitude is equal to the amperes field no load (AFNL), the voltage at the generator terminals will be the nominal voltage. It is very important that the field be increased after the rotor has come to rated speed and not before. This way it minimizes the risk of over fluxing during this operation.

4.2.3

Synchronized and Loaded (Online)

Once the generator is at rated speed and rated terminal voltage, the sinusoidal waveform of the generator output must be matched to the system waveform by frequency, voltage level, and phase shift. The frequency and voltage level are achieved in the open circuit condition when the generator is brought to rated speed and the field current is raised to the AFNL value. The phase shift (or vector shift) is accomplished automatically by a “synchro scope,” which adjusts the generator output voltage to be in phase with that of the system, or manually by the operator. Once the generator is synchronized to the system, the main generator breaker is closed and the generator is connected to the system. At this point, loading the turbine will increase the generator’s MW output. Power factor and reactive power output are adjusted by changes to the rotor field current. Table 4.2-1 contains a useful method to determine the

TABLE 4.2-1 Generator operating modes

Generator breaker

Field breaker

Shutdown Run up/down

Open Open

Open Open

Speed no load field on On line and loaded

Open Closed

Rotor speed

Terminal voltage Zero Zero/Zero

Closed

Zero Zero to rated speed/rated speed to zero Rated

Closed

Synchronous

Rated or higher or lower ±5%

Rated or lower

4.2 OPERATING MODES

189

generator-operating mode, using indications of generator main circuit breaker status, field breaker status, rotor speed, and terminal voltage.

4.2.4

Start-Up Operation

Following is a nonexhaustive list of activities that should be followed before attempting to start a generator: • If any maintenance was performed, ensure all equipment and tools are vacated from the generator. • If isolation was required, make sure all the protective devices ensuring isolation are removed. • Make sure all protection is enabled and operational. In some protective schemes, a number of relays may have to have their trip function curtailed during start-up. Make sure the OEM’s operational instructions are followed exactly. • Watch the maximum open circuit terminal voltage. • Follow clear and safe synchronizing procedures as outlined by the OEM. • Do not attempt to re-energize the machine without an investigation and/or inspection after a protective relay has operated during a start-up.

4.2.5

Online Operation

Following is a nonexhaustive list of activities that should be followed during the operation of a generator: • The unit must remain within its capability curve at all times. • Voltage regulators and power system stabilizers (when applicable) should be in operation at all times. • All protection and monitoring devices must be in fully functional condition and always in operation. Typical generator trips are • Phase-to-phase fault • Stator ground fault • Reverse power (motoring) • Differential • Volts/Hz • Loss of excitation • Vibration (if unit is not closely monitored by personnel) • Other protective functions/systems might also trip the unit, according to specific unit requirements and design.

190

CHAPTER 4

4.2.6

OPERATION AND CONTROL

Shutdown Operation

Following is a nonexhaustive list of activities that should be followed during the shutdown of a generator: • The unit is unloaded to zero load. • The main circuit breaker should be tripped and then the field breaker immediately afterward (this is taken care of in the protection scheme). • If equipped with a Babbit bearing, turn oil lift system on once 30% rated speed has been reached. • After the unit has come to a complete stop and the creep detector is turned on, the oil lift system can be turned off (the creep detector should initiate the oil lift if needed).

4.3 MACHINE CURVES 4.3.1

Open Circuit Saturation Characteristic

The open circuit saturation curve for the generator provides the characteristic of the open circuit stator terminal voltage as a function of field current, with the generator operating at rated speed. At low voltage and, hence, low levels of flux, the major reluctance (magnetic resistance) of the magnetic circuit is the airgap. In the linear portion of the opencircuit curve, terminal voltage and flux are proportional to the field current. This portion of the open circuit saturation curve, which is linear, is called the “airgap line.” At higher voltages, as the flux increases, the stator and rotor iron saturate, and additional field current is required to drive magnetic flux through the iron. This is due to the apparent higher reluctance of the magnetic circuit. Hence, the upper part of the curve bends away from the airgap line at an exponential or logarithmic rate, dependent on the saturation effect in the stator and rotor. Without the presence of iron in the circuit, the airgap line would continue on linearly, meaning that the terminal voltage and machine flux would increase in linear proportion to the increase in field current. Figure 4.3-1 shows the open circuit saturation curve.

4.3.2

Short Circuit Characteristic

The short circuit characteristic curve is a plot of stator current (from zero up to rated stator current) as a function of field current, with the stator winding terminals short circuited and the generator operating at rated speed as shown in Figure 4.3-1. The short circuit characteristic is for all practical purposes linear because in this short circuit condition the flux levels in the generator are below the level of iron saturation. The short circuit curve is also called the “synchronous impedance curve” because it is the synchronous impedance of the generator that determines the level

4.3 MACHINE CURVES on

Terminal voltage

regio

n

O.C.C

Rated voltage

Ra te U n d MV it y A PF Ra ted MV A, rat e

Airlin

e

a

-lo

No

dP F

ati tur

a ds

191

Unsa tu

rated

it cu c cir sti r t teri o Sh arac ch C. C. S.

Field current IFNL

IFSC

IFNL = Field current required to produce open circuit rated voltage IFSC = Field current that produces rated armature current with short-circuited terminals The S.C.C. relates almost exclusively to the “armature reaction” of the machine

Short circuit ratio (SCR) = IFNL IFSC Figure 4.3-1 Open circuit and short circuit characteristics.

of the stator current for the machine. This can be readily seen by inspection of Figure 1.7-7. It can be seen in the figure that when Et = 0, the entire internal generated voltage (Eo) is dissipated across the synchronous impedance (Zs). The synchronous impedance is highly dependent on the armature reaction of the machine (Xa). It is convenient sometime to plot the open circuit and short circuit curves by themselves during testing, just to separate the clutter of the graph as shown in Figures 4.3-2 and 4.3-3. From these two figures, the following quantities can be derived: IFNL = Field current required to produce open circuit rated voltage of 13 800 V IFNL = 632.1 A IFSC = Field current that produces rated armature current of 2866 A with short-circuited terminals IFSC = 549.5 A The SCC relates almost exclusively to the “armature reaction” of the machine and can be calculated an alternate way as shown in Equation (4.7).

192

CHAPTER 4

OPERATION AND CONTROL

Open circuit saturation curve 20000 18000 16000

Armature voltage (kV)

14000 12000 10000 8000

Airgap field current is 587.0 (IFG)

6000 4000 Field current at 13 800 kV is 632.1 (IFNL) 2000 0

0

200

400

600

800

1000

1200

Field current (DC Amperes)

Figure 4.3-2 Typical open circuit saturation characteristic.

Short circuit ratio SCR =

I FNL = 1 15 I FSC

(4.7)

In most hydro generators, the range of SCR = 0.8 − 1.6.

4.3.3

Capability Curves

Capability curves are plots of apparent power capability (MVA), at rated voltage, using active power (MW) and reactive power (MVAR) as the two principal axis. Circumferences drawn with their centers at the origin represent curves of constant stator current. A capability curve for a 72 pole, 12 kV, 0.85 pf machine is shown in Figure 4.3-4. This curve serves to separate the region of allowed operation (inside the curve) from the region of forbidden operation (outside the curve). Note that the term “negative” in the “negative excitation limit” does not refer to actually changing the polarity of the excitation signal. It refers to the minimum limit of excitation allowed for absorbing reactive power.

4.3 MACHINE CURVES

193

Saturation curve 4000

3500

Armature current (AC Amperes)

3000

2500 Field current at Ia = 2866 Amps is 549.5 (IFSI)

2000

1500

1000

500

0

0

100

200

300

400

500

600

700

Field current (DC Amperes)

Figure 4.3-3 Typical short circuit characteristic.

On the graph, the quantities are represented in PU where 1 PU is 17.25 MVA. One axis represents MW and the other represents MVAR, and a circumference represents a constant MVA of 17.25 until other limits are encountered such as a rotor heating limit or negative excitation limit [4]. If in the case of a machines capability curve the voltage is kept constant (at rated value), then a circumference also represents a constant current trajectory. On the same graph, any line starting at the intersection of the axis moving radially outward, intersecting the circumference, represents a particular power factor, in this case it is the rated power factor of 0.85. Other information can be added (separate from what the OEM supplied) to the capability curve as shown in Figure 4.3-5 which has more information as to what limitations are present on the machine. This type of curve is normally internal

194

CHAPTER 4

OPERATION AND CONTROL

72 P 17.25 MVA 0.85 PF 12 kV 1.1 1 0.9

Field heating limit

Active power (PU)

0.8

Stator heating limit

0.7 0.6 0.5 0.4 0.3

Rated power factor line

0.2 Negative excitation limit

0.1 0 –0.7 –0.6 –0.5 –0.4 –0.3 –0.2 –0.1

Overexcited

0

0.1

0.2

0.3

Reactive power (PU)

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

Underexcited

Figure 4.3-4 Typical capability curve from the manufacturer.

to the power producer specifically for the use of the engineering and operating departments. There are many reasons why information is shown on the power producers’ capability curve such as regulatory requirements that must be satisfied, ease of operator intervention when limits are reached, ease of interpretation, etc. Table 4.3-3 compares the information on the two capability curves. Note that these two curves are for two different machines. It can be seen from Table 4.3-3 that the two capability curves are basically saying the same things where common elements are shared with the exception of excitation limit. The negative excitation limit on the OEM curve corresponds to absolutely no excitation, whereas the utility curve corresponds to an excitation limit due to the static exciter having a minimum current for firing of the thyristors, in this case, it is 10 A. More discussion on the OEM limit will be had when Figure 4.3-6 is explained shortly. The utility curve has made it a little easier for the operations staff to interpret the capability of the generator by adding numbers as descriptors as well. Figure 4.3-5 is an example of a customized capability curve specifically for that utility and that generator. Keep in mind that although units may be identical in design and may have been installed consecutively in the powerhouse, it does not mean the capability is exactly the same. For example, the rotor on one machine may reach its designed temperature rise at 1110 A and the sister machine at 1200 A, a difference of 8%. In this realworld case, this difference did not affect the guaranteed performance of the

4.3 MACHINE CURVES

195

19 17 15

Rotor current current limit limit Rotor

13 11 11.4 kV 12 kV 12.6 kV

9 7 0.9 pf

5

Reactive power (MVAr)

3 Active power (MW)

1 –1 0

2

4

6

8

10

12

14

16

18

20

–3 –5 0.9 pf –7 –9 –11

Exciter minimum –13 excitation

Stator current limit

–15 Mechanical limit 18.5 MW

–17 –19 –21

excitation Loss of Excitation protection

–23 –25

Figure 4.3-5 Basic capability curve for internal utility use.

machine, but it is an important distinction to document. There can be many reasons for this difference and technically, it should be noted. This raises a question as to whether or not to test each machine for performance guarantees that are installed in a powerhouse that may have many identical machines. The additional cost for testing of performance guarantees is an incredibly small fraction of the entire cost of installing a large generator that it should be done without a second thought.

196

CHAPTER 4

OPERATION AND CONTROL

TABLE 4.3-3 Table comparing information on capability curves

Item on curve

OEM

Internal to utility

Stator limit Rotor limit Power factor line Excitation limit Relay protection Mechanical limit

Stator heating limit Field heating limit Rated power factor line

Stator current limit Rotor current limit 0.9 pf lagging 0.9 pf leading Exciter minimum excitation Loss of excitation protection Mechanical limit 18.5 MW

Negative excitation limit (0.0 A – no excitation – top of saliency circle) Not shown Not shown

4.3.3.1 Construction of Approximate Reactive Capability Curve As stressed often in this book, operators should always maintain the generator within the capability curves of the machine. Given the importance of this, it is rare for the capabilities of a generator to be unavailable. However, for those rare occasions, one can construct an approximated capability curve and the reader is pointed to an excellent set of references [4, 6]. 4.3.3.2 Limits Imposed by the Generator The turbine and the generator are designed to operate as a unit. As was stated earlier in this book, the generator rating is almost always designed to be somewhat larger than the turbine. This fact is shown as a line inside the maximum MW output of the unit at unity PF, labeled “mechanical limit” as shown in Figure 4.3-5. In the case of this generator, 65 MVA is the 120 C total temperature limit (80 C rise), but the mechanical drive train of the generator is limited to 57.5 MW. Excitation limitations in the leading PF region also places limits on the MVAR import capability of the machine. This limitation is shown as a partial circle in the leading portion of the capability curve labeled “minimum field current” and is a characteristic of static excitation. These limits become more apparent when upgrades to the machine output are being done and studies are performed to see what thermal, mechanical, and electromagnetic limits the machine actually has. Therefore, the “working” capability curve of the entire unit represents a combination of generator, turbine, and system constraints. Pay particular attention to the fact that the orientation of Figure 4.3-4 is different than that of Figure 4.3-5. In some countries, it is common to show the MW axis on the horizontal and in other countries, it is presented on the vertical axis. Similarly then, the MVAR values would be on opposite axis as well. Lastly, notice that the values are in per unit (PU) in Figure 4.3-4 and in numbered quantities in Figure 4.3-5. Last but not least, stability considerations limit the number of MVARs the unit can import when operating in the leading PF region. This is shown as the

0.1 U/XD

Theor. stability limit Pract. stability limit

0.1 PWN

P/SN 1.4

The power chart indicates the permissible electr. load. but no mech. torque

1.2 Heating limt of stator winding

1.0 Rated load point

NN = 112 RPM

Underexcitacion limit Reserve of stability

UN = 13 800 V

XQ = 0.52

SN = 87 000 kVA

XD = 0.74

I = 1.00

FN = 60.0 Hz

U = 1.00

0.8

Angle of phase

Limit rotor temp

0.6

0.4

–2.2

–2.0

–1.8

–1.6

–1.4

–1.2

–1.0

Underexcited

–0.8

–0.6

–0.4

–0.2

0.0

0.2

0.4

0.6

0.0 0.8

PF = 0.900

0.2

Overexcited

Figure 4.3-6 Another manufacturer curve showing saliency circle and theoretical and practical stability limit.

198

CHAPTER 4

OPERATION AND CONTROL

“theoretical and practical stability limits and saliency circle” on the curve shown in Figure 4.3-6. We shall now focus on these items of the capability curve and what they represent. The machine curve depicted in Figure 4.3-6 represents a robust machine since Xd = 0.74 PU. Keeping in mind that the smaller the value of Xd, the more robust and stable the machine will be, and the more expensive it will be to purchase, so there is a tradeoff. The small semicircle at the bottom of the curve is also known as the “saliency circle.” The top of this circle is actually when there is no excitation on the machine, also represented by the PU quantity 1/Xd. In this case, the number works out to be 1/0.74 = −1.35 PU. The bottom of the semicircle is represented by 1/Xq which in this case is 1/0.52 = −1.92 PU. The value of 1/Xd is also known as the theoretical stability limit of the cylindrical rotor machine [6] and can be represented by a line drawn across the graph perpendicular to the top of the semicircle. The machines we are discussing in this book are salient pole so that this stability limit does not apply; however, it is an important point at the top of the saliency circle that should be recognized. At this point on the semicircle, with no excitation, the reluctance torque (power due to saliency) as discussed in Chapter 2, takes over. The amortisseur assembly now takes over and keeps the generator rotor in synchronism without any excitation applied to the main rotor field; this is also known as self-excitation. It should be noted that it is not possible to get to the bottom of the semicircle without losing synchronism because the theoretical stability limit will be exceeded. Further, negative excitation would be required to enter the circle, thus operation is limited to the outer portion of the circle only. It is also important to recognize as well that a finite amount of real power can be produced while on the outer edge of this circle, moving off the circle, with real power will result in loss of synchronization. The theoretical stability limit is a curve-shaped line and begins at the origin at 1/Xq on the y-axis, then passing through the 90 point of the semicircle and then becoming asymptotic to the 1/Xd line discussed previously. Since operation of the machine near the theoretical stability limit is not recommended, the OEM will place a practical stability limit on the curve which normally equates to 10% above the theoretical stability limit (dotted line on Figure 4.3-6) to give margin and warning to the user. It is recommended the reader source Ref. [6] as it has a more thorough representation of the previous discussion. 4.3.3.3 Capability Curves Adjustments for Nonrated Terminal Voltage As previously discussed, most generators allow a ±5% voltage deviation from nominal volts. The capability curves behavior must be understood when attempting to operate at ratings above and below rated voltage, in particular the rotors capability to supply MVAR output as shown in Figure 4.3-5. Excursion from the 12 kV rating for the rotor will affect the amount of rotor current available for MVAR support. In our particular case, let us assume a rotor current limit of 1140 A and subsequent heating limit of 80 C rise is shown on the curve. This means, when 1140 A flows through the rotor, the total temperature of the rotor winding is 120 C (80 + 40) based on the design standard of 40 C cold air. It takes less field current for the generator to reach a system voltage of 11.4 kV than it does

4.3 MACHINE CURVES

199

to reach 12.6 kV. Let us look at an example using simple numbers for rotor current to illustrate the example, these are not actual numbers. For operation at 12.0 kV, the rotor current required to get the generator to 12.0 kV is 300 A. This leaves 840 A left in the rotor capacity to provide the approximate 12 MVARS if required. For operation at 11.4 kV, the rotor current required to get the generator to 11.4 kV is 250 A. This leaves 890 A left in the rotor capacity to provide the approximate 13 MVARS if required. For operation at 12.6 kV, the rotor current required to get the generator to 12.6 kV is 350 A. This leaves 790 A left in the rotor capacity to provide the approximate 10.5 MVARS if required.

4.3.4

V-Curves

V-curves provide the apparent power (MVA) as a function of field current, plotted for various constant power factors, holding speed and stator voltage at the rated values as shown in Figure 4.3-7. Horizontal lines represent constant stator current. The rating of the generator is the intersection of the line for rated apparent power (1.0 PU) and the curve for rated power factor (usually 0.9 lagging) or any other power factor. All constant power factor curves converge at a common point at zero apparent power. This is at the field current for rated voltage, open circuit. It is important to remember that the V-curves shown are valid for rated voltage only, once the voltage changes from rated, the V-curves will shift.

0.50 0.00

0.75

0.90

1.00

0.90

0.75

0.50

0.0 0.25

0.8

IN = 3640

P/SN 1.00

1.0

IF0 = 972.9 A

1.2

NN = 112.5 RPM

Overexcited

UN = 13 800 V

Underexcited PF

SN = 87 000 kVA

I/In 1.4

FN = 60.0 Hz

Heating limit of stator resp. field winding Limit of stability; security 0.10 PU rated active power

0.75

0.6 0.50

0.4 0.25

PF = 0.900

0.2 0.0 0.0

0.5

1.0

1.5

2.0

2.5

3.0

Figure 4.3-7 Typical V-curve from the manufacturer.

3.5

4.0 IF/IF0

200

CHAPTER 4

OPERATION AND CONTROL

4.4 SPECIAL OPERATING CONDITIONS 4.4.1

Unexcited Operation (“Loss-of-Field” Condition)

Operation without field current is uncommon and can occur under a number of circumstances, but the two most common are loss of field during operation and inadvertent energization. 1. Loss of field during operation. If for some reason the field current goes to zero while the generator is connected to the system, the generator must immediately be removed from the grid and the wicket gates should be closed to bring the generator to a speed no load condition or a complete stop depending on the philosophy of the utility. Failure to do so in an expedient manner could result in loss of synchronism depending on the amount of real power being produced and the amount of VARs being imported. Protection is commonly provided by the loss of excitation relay to remove the generator from the system by opening the main output-circuit breaker and commanding the governor to safely stop the flow of water through the turbine. 2. Inadvertent energization. If a generator is at rest and the main generator three-phase circuit breaker is accidentally closed, connecting it to the power system, the magnetic flux rotating in the airgap of the machine at synchronous speed will induce large currents in the amortisseur bars. The stator windings will also be subjected to high electromagnetic forces at this time. The rotor then begins rotating as an induction motor as the amortisseur bars are now active in trying to counteract the field that is in relative motion to them (slipping) by setting up an opposing field to stop this. The amortisseur bars are now the main link to the rotating field in the airgap as they act like a squirrel cage winding on an induction motor. Some generator/motors are designed to be started from standstill to synchronous speed using the power system (line start); however, most synchronous machines are not designed for this. This type of inadvertent operation usually results from some sequence of events that closes the breaker between the generator and the power system while the generator is at or near standstill. The machine will attempt to accelerate to synchronous speed, and, unless there is rapid intervention, very severe damage up to and including total destruction of the turbine generator set can occur. Operation of the protective relaying is essential under these circumstances, as there is little or no opportunity for operator intervention before damage occurs. Protective relaying like the 32 reverse power relay, motoring protection will prevent prolonged operation of this event.

4.4.2

Negative Sequence Currents

A three-phase balanced supply voltage applied to a symmetrical three-phase winding generates a constant magnitude flux in the airgap of the machine, which rotates at synchronous speed around the circumference of the machine In addition, the

4.4 SPECIAL OPERATING CONDITIONS

201

slots and other asymmetries within the magnetic path of the flux create low magnitude space harmonics (i.e. fluxes that rotate in both directions) of multiple frequencies of the fundamental supply frequency. In a synchronous machine under normal operation, the rotor rotates in the same direction and speed as the main (fundamental) flux. When the supply voltage or currents are unbalanced, an additional flux of fundamental frequency appears in the airgap of the machine. However, this flux rotates in the opposite direction from the rotor. This flux induces in the rotor windings and pole body, voltages and currents with twice the fundamental frequency. These are called negative sequence currents (I2). The negative sequence terminology derives from the vector analysis method of symmetrical components. This method allows an unbalanced three-phase system to be represented by positive, negative, and zero sequences. The larger the imbalance, the higher is the negative sequence component. There are several abnormal operating conditions that give rise to large currents flowing in the field windings of synchronous machines. These conditions include unbalanced armature current (producing negative sequence currents) and inadvertent energization of a machine at rest. All large synchronous machines have (or should have) installed protective relays that remove the machine from operation under excessive negative sequence currents. To properly “set” the protective relays, the operator should obtain maximum allowable continuous negative sequence I2 values from the OEM. The values shown in Table 4.4-1 are contained in Ref. [1] as values for continuous I2 current to be withstood by a generator without injury, while exceeding neither rated MVA nor 105% of rated voltage. When unbalanced fault currents occur in the vicinity of a generator, the I2 values will probably be exceeded. In order to set the protection relays to remove the machine from the network before damage is incurred, but avoiding unnecessary relay operation, manufacturers have developed the so-called (I2)2t values. These values represent the maximum time in seconds a machine can be subjected to a negative-sequence current. In the (I2)2t expression, the current is given as PU of rated stator current. These values should be obtained from the manufacturer. Table 4.4-2 shows the typical value given in the Ref. [1] standard. TABLE 4.4-1 Values of permissible I2 current

Type of generator or generator/motor

Permissible I2 as % of stator current

Non-connected amortisseur Connected amortisseur

5 10

TABLE 4.4-2 Permissible (I2)2t in a generator

Type of machine Salient pole generator

Permissible (I2)2t 40

202

CHAPTER 4

OPERATION AND CONTROL

There are sources, in the generator and the power system, of currents at frequencies other than that of the power system, for example, current components at higher frequencies produced by transformer saturation and by incompletely filtered harmonic currents from rectifiers or inverters. Current components at frequencies lower than that of the power system have been produced by resonance between power factor compensating series capacitors (used to increase the power handling capability of long AC transmission lines) and the inductance of generators and transformers. This is commonly known as subsynchronous resonance. Off-frequency currents interact with the useful flux in the generator to produce pulsating torques felt by the combined turbine and generator shaft system. If the frequency of one component of the pulsating torque is identical to the torsional natural frequency of any mode of vibration of the complex shaft system, destructive vibration could result. The degree of damage depends on the mode shape and the level of the current damping present.

4.4.3

Load Cycling and Repetitive Starts

It is well known in the power industry that load cycling represents a long-term onerous mode of operation. Generators “like” to be in a steady-state condition, meaning where the temperatures in the machine are stable. Any situation in which load is changed significantly will result in relatively large changes in temperatures. It is the transition time between the steady states that embraces an amalgam of problems. For instance, when load is increased suddenly, the rotor and stator conductors will rise in temperature first, followed by the core and other components. As the temperature differentials increase momentarily, so do the mechanical stresses induced. Other problems that are encountered include loosening of stator wedges, loosening of the stator core, weakening of the stator endwinding support system, cracking of conductors, weakening of frame support systems, accelerated deterioration of rotor and stator insulation, rotor interpole connections developing cracks, and so forth. By far the most onerous load cycling is the complete start and stop operation. It is more commonplace these days that generators start and stop several times every day. This type of operation stresses all those elements enumerated above to the extreme. It is important to recognize that this accelerated deterioration mechanism with many starts and/or load cycling demands that the inspection intervals be significantly shorter than for units operated under base load conditions. Let us now explore in more detail what happens to the components mentioned above during a load cycle or repetitive start. There was research conducted in 1989 that showed that the copper temperature followed a load change almost immediately and that the copper temperature as measured with a fiber optic sensor was 15 C hotter than the thermocouple at the coil surface [7]. The heat generated or decreased by a load change affects the materials associated with conductors very quickly as well followed by the components associated with the copper conductors shortly thereafter. Starting with the frame, we can work our way into the machine. The stator frame sits on top of soleplates as discussed in Chapter 2 and in some

4.4 SPECIAL OPERATING CONDITIONS

203

cases is able to move radially free as it expands and contracts with temperature. Over time, this radial freedom can be compromised as the frame may not heat evenly causing the frame to bind up against the radial key. This in turn causes uneven airgaps in the machine complicating optimal operation. The tens of thousands of laminations that potentially make up a core will expand and contract due to thermal effects with load changes and start/stops. If the core is tightly stacked, laminar movement with respect the adjacent lamination should not occur. If the core is loose, laminar movement will occur to some extent with each load change or start/stop, degrading the inter-laminar insulation exposing the core to the potential of circulating currents and the development of hotspots. The stator winding, again, is made up of copper conductors bonded together with whatever substance the manufacturer has developed. The fact is that copper expands much faster and to a greater extent than anything holding it together or providing insulation between the conductors. This means there is a shear force between the copper and the surrounding materials trying to separate them. At some point in time in the life of the winding, sometimes sooner than one might think, the copper and adjoining materials are completely free from each other. This now allows the individual strands of copper within the stator winding to vibrate freely in response to the magnetic forces that surround them. This in turn causes fretting of the insulating material thus exposing the coil and bar to strand to strand shorting and the coil to “turn to turn” failures. To complicate the situation even further, the coil or bar in the slot will expand axially and radially during a load change or start/ stop. The core and coil are expanding at different rates, and if the coil or bar is not packed tightly, the relative motion between the two can remove the semiconductive coating and partial discharge in the slot can result. Extended relative motion will cause eventual groundwall insulation failure. Also, the radial expansion and contraction of the coil itself (particularly the old asphalt type) can cause the wedges to loosen. This is because the asphalt becomes soft at elevated temperatures and as the machine is loaded to rated output, the asphalt takes shape in that area (due to the wedge pressing on the coil). Then, when the asphalt cools and contracts, it leaves a flat spot in that area. This flat spot has now created a gap between the coil surface, the depth packing, and the wedge body. Thus, it is very important that a rewedge on an asphalt winding be discussed with the OEM and a procedure and/or wedge design enhancements be made during installation to alleviate this problem. This is less of a problem with modern-day coil insulation and wedging systems. If the coil was seated properly in the slot and secured with a good wedging system, the wedges should remain tight over time. The end region of the stator winding is tied to the bull ring and to adjacent coils or bars using a modern glass roving soaked in epoxy or polyester or can be the older style which is simple chord with some sort of varnish. This forms a very tightknit “basket,” which is primarily there to ensure movement is minimized during generator faults and to counteract vibrations present during steady-state operation. No matter which method is used to tie the endwinding together, load cycling and repetitive starts will cause the bar or coil to expand and contract with respect to the points that are tied off causing fretting of the insulation system. If allowed to

204

CHAPTER 4

OPERATION AND CONTROL

continue without repair, this will fail the endwinding area potentially causing extensive damage to the machine. Normally, this relative motion between coil, support ties, wedges, etc., can be noticed by fretting dust from the material being abraded away. This dust can be of different colors depending on how contaminated the airflow is in the area in question. So, any dusting that looks orange-reddish in color, or even white (not to be mistaken for partial discharge residue on the winding surface) should be addressed immediately. If the machine has excessive oil leaking and the oil is atomized in the airflow, the dust will turn to a paste. Anything on the core, frame, rotor, or winding of the generator that looks like “tomato paste or has an orange-reddish dust” is a sign that fretting is occurring and investigation as to the cause and corrective action should be priority. The OEM should be consulted if fretting is discovered and discussion of the urgency of remedial work should occur before the unit is placed back into service. The next part of the machine that suffers a great deal during repetitive starts and load cycling is the rotor. The rotor may seem like a fairly simple device, but nonetheless, without the rotor, there is no generation. This fact is often forgotten, and it is the component of the machine that often receives the least attention or maintenance. Let us examine what the consolidated copper conductors and their pole-to-pole interconnections (including the squirrel cage amortisseur, if equipped) are experiencing during a start/stop cycle and load change. Keeping in mind that all the rotor components are originally assembled at room temperature at the rotor at rest and this is the preferred position for a copper conductor or an interpole connector. Pole-to-pole connectors that are solid and short in length will potentially be most affected by the following descriptions. From a cold starting position, thus ambient powerhouse temperature, the unit is called to start. The generator reaches its normal operating speed in less than one minute after initial rotation. At this point, there is no field applied, just centrifugal forces. The strip wound copper conductors would like to pull away from the pole body and occupy the interpolar space as well as the conductors are being compressed in the radial direction. The pole-to-pole connectors are being stretched (since they are cold at startup), because the rim is also stretching due to centrifugal forces. So whatever mechanism is keeping the two connectors together (bolts, solder and bolts, braze, etc.) has to hold them together before any thermal relief arrives when the load is applied and the copper expands. These same connectors are also being pushed into the airgap due to the centrifugal forces. The nonflexible style amortisseur connections will have to endure similar conditions depending on when the amortisseur circuit is active and if this is a pump generator or not. So just by starting the machine, we have a whole bunch of things working against us. Now, the operator decides to apply field and get the unit ready for synchronizing to the system and then puts the unit on line. So what has changed in this situation? The field current is now heating up the copper conductor, and the stator winding is heating up the stator and the air inside the machine is getting hotter. With the air in the machine getting hotter and the field winding heating up the copper and pole assemblies, the copper interpole connectors begin to thermally expand and the stretch the connectors

4.4 SPECIAL OPERATING CONDITIONS

205

experienced on start-up is relieved to some degree. The rotor rim will continue to expand as it grows thermally as well. At some point when steady-state temperature is reached at a specific load, all is working as designed. The best thing to do now is to leave the generator at this load until the unit is shut down. However, this is not today’s market requirements, and this generator is going to go through several load changes per hour, sometimes large load changes that will affect the operating temperature significantly. Let us now go back to the operations center where this machine that has been running at the original load for several hours and is now thermally stable and change the load. The machine goes from full rated load to say 60% load due to changing market conditions and system requirements. What has changed now that the load has been reduced? The only thing that has changed is the thermal loading on the rotor, reduced field current means less heat and lower stator current means less overall heat as well in the stator. The air inside the machine cools down somewhat and the copper conductors now contract and move once again to a new position where they remain stable until the next time a load change is made. This small movement (thermal cycling) of the conductors, over and over again, eventually leads to separation of the copper conductors and the insulation system. Now, to make matters worse, the operations center calls for the machine to be shut down. The copper conductors within the stator winding begin to contract from their expanded condition putting shear stress once again between the copper and insulation system. The pole-to-pole connectors are put under stress once again since the copper is still hot from operation, and the rotor rim has now come back to its original position since the centrifugal forces are gone. The pole-to-pole connector is fighting the compressive forces while the copper contracts back to its cold rest position. The pole-to-pole connections must withstand these forces without failing, which usually begins with a small crack in the copper and then progresses from there. The copper conductors are now sitting back against the pole body and no longer want to occupy the inter polar space. It is easy to see that all this movement is bound to cause a problem at some point in time, so minimizing start/stops and load changes is good practice if it can be achieved.

4.4.4

Overloading

The need to remain within the capability curves of the machine at all times was previously stressed. Nonetheless, if a severe overload situation is reached, the need to schedule an inspection of the windings of the machine as soon as possible should be considered. Bear in mind that the heating developed in a conductor is proportional to the square of the current. Thus, a 10% overload condition will increase the heat generated in that conductor by about 20%. The temperature will also change in a similar fashion. However, the expected life of insulation is approximately halved for every 8–10 C increase in temperature (the Arrhenius Law, after Svante August Arrhenius, 1859–1927). Thus, long-term operation at moderate overloads or shorttime severe overloads can markedly reduce the expected life of a machine’s insulation systems.

206

CHAPTER 4

OPERATION AND CONTROL

Another important factor to consider is the real power overload in MW that the drive train is being subjected to. Overloading the machine mechanically (drive train) can cause damage in the long- and short-term depending on how much margin was built into the machine. Failure modes when it comes to mechanical items are less forgiving than a coil failure. Caution should be exercised and the OEM consulted to ensure long-term reliability.

4.4.5

Loss of Cooling

On a rare occasion, the unit may be operated inadvertently without cooling water for some period of time. This problem may result in serious overheating of the windings and, perhaps irreversible damage. After such an event, the unit should be removed from service and opened for careful inspection of the stator windings. What type of damage may occur under loss of cooling operation is largely predicated on the kind of insulation system. For example, thermoplastic systems (asphaltic) will deform under severe heating. Oozing of the asphalt from the bottom endwindings and permanent insulation migration may occur. Depending on the temperature reached and operating time, the asphalt winding may need replacing since this type of physical damage is irreversible. On the other hand, thermosetting systems such as polyester and epoxy will be more resilient when it comes to over temperature as physical insulation migration does not occur. Though a loss of expected life of the insulation might have occurred, the overall situation of the winding may still be satisfactory for long-term operation. There are a number of mechanical problems that may also result from high temperatures attained during loss of cooling operation, such as core buckling and stator frame distortion both of which are not reversible. It goes without saying that the monitoring systems put in place to prevent these types of events should be in excellent working order and maintained accordingly.

4.4.6

Over Fluxing

Over fluxing occurs when a generator is operated beyond its maximum continuously allowed V/Hz. In Section 4.1.11, an elaborate description of over fluxing was included. Once the machine is connected to the system, the probability of sudden damage due to over fluxing is very low. In any event, it is truly important that V/Hz protection is properly designed and set. Moreover, it is important to design the voltage-sensing scheme for the excitation in such a way that loss of a single potential transformer winding will not result in a V/Hz event. IEEE [8] has a good discussion on the subject and presents examples of how to design and set the protection schemes [9]. Figure 1.1-6 is an example of a V/Hz withstand curve for a 70 MW hydro generator. These curves are not typically issued unless they are asked for at the time of purchase. For your specific machine, consult the OEM when setting the protection if this curve is not available in local records.

4.4 SPECIAL OPERATING CONDITIONS

4.4.7

207

Runaway and Overspeed

The existing industry standard [1] states the unit (as a whole including turbine and shaft assembly) shall withstand without injury for five minutes the maximum speed of the combined unit, or better known as runaway speed. There have been many revisions to this particular standard over its lifetime, and thus caution must be used to determine what standard was in place when the generator was manufactured. Before the IEEE standard was in place, the utility may have specified a runaway time constraint, thus it is a prudent idea to search for original records. Each manufacturer would have a design philosophy for a runaway event (normally twice rated speed or more) and sustaining times and each manufacturer may be different. It is not recommended for a machine that has been in service for a number of years to test the runaway condition unless there extenuating circumstances that would require this test. This is a very onerous condition for the generator to sustain for any period of time and even if it is designed to sustain this without injury for a prescribed time, a full unit inspection must be completed after such an event. In some utility and private power producer practices, overspeed is referred to as the condition the generator may reach when a full load rejection is encountered due to system conditions. In this case, overspeed would be in the range of 130–145% of rated speed but only for a few seconds while the unit ramps up and then comes back down in speed to shutdown. This is not a runaway speed condition. Thus, is it always best to check with the OEM what the load rejection speed is and what the design overspeed is of the generator in question. When reading documentation on the generator, it is important to remember that runaway and overspeed are two distinctly different speeds, they are not the same number.

4.4.8

Loss of Lubricating Oil

The result of a loss of lubricating oil during operation can be catastrophic. Such events are not unheard of, but only few result in severe loss of equipment. Failure of this system can be very costly in material and lost production. It is very important not to forget to check the lubricating oil levels and temperatures when carrying out periodic inspections and operational testing of the unit’s support systems.

4.4.9 Out-of-Step Synchronization and “Near” Short Circuits Both out-of-step synchronization and short circuits occurring in or in the vicinity of the generator (in particular, between the generator’s terminals and the main step-up transformer) can result in severe damage to the unit, damaged stator coils, endwindings and endwinding supports, torsional stress and possible damage on the generator shaft and coupling. The severity of the damage depends on several factors such as the generator rating, the angle between the system and generator voltage vectors at the moment

208

CHAPTER 4

OPERATION AND CONTROL

of synchronization, and type of short circuit, phase-to-phase or three-phase. Faults within the machine itself and on the main step-up transformer, or very close to the high-voltage side of the transformer, can seriously damage the generator. Therefore, before a new attempt is made to synchronize the unit after a major out-of-step event, or in the aftermath of a strong and near short circuit, the unit should be opened for visual inspection. Out-of-step synchronization protection is discussed in Chapter 6. For additional discussion on sudden short circuits, see Section 4.5.7

4.4.10

Under and Over Frequency Operation (U/F and O/F)

U/F and/or O/F operation indicate that the generator is operating slightly under or over the rated frequency of the system. IEEE standard [1] clearly outlines the generator shall be thermally capable of continuous operation within the confines of their reactive capability curves over the ranges of ±5% in voltage and ±2% in frequency. However, the standard also states generators will also be capable of operation within the confines of their reactive capability curves within the ranges of ±5% in voltage and +3%/−5% in frequency with further reduction of insulation life. In the case of the generator, as well as the transformers connected to the generator, the main concern is that while running under frequency and at the highest allowable terminal volts, the machine may move beyond the permissible V/Hz region. This condition should result in an alarm generated by the protective 24 relay so that the operator has an opportunity to correct it either by lowering the terminal voltage manually or removing the unit from operation. In some protection philosophies, excursion from the allowed frequency fluctuations will result in the relay removing the generator from service in order to prevent damage. The relay can be coordinated with a curve from the OEM as shown in Figure 1.1-6.

4.5 BASIC OPERATION CONCEPTS 4.5.1

Steady-State Operation

A hydro generator can be seen as a nonlinear combination of magnetically coupled windings, airgap reluctance, and the electrically conductive mass of the rotor body (which acts as a distributed winding). Its electrical characteristics when it is operating in a steady-state fashion are very different from those when conditions are changing. For many conditions, a hydro generator can be represented as a reactance in series with a voltage source that reactance takes on different values for different operating conditions. In addition to the familiar concept of a reactance as it functions in an electric circuit, there are magnetic considerations that are useful in describing the operation of a synchronous machine. An inductance (which is multiplied by the angular frequency to obtain the reactance) can be defined as the flux linkages produced by one ampere of current. Thus, the reactance is a measure of how easy it is for current to produce flux in the machine.

4.5 BASIC OPERATION CONCEPTS

209

When the generator is operating in a steady load-carrying condition, it appears to the power system as a voltage source connected to the generator terminals through the generator’s synchronous impedance (Figure 1.7-7). The generator resistance is negligible, and it is common to consider only the generator’s reactance, in this case the synchronous reactance Xs. During steady-state operation, a component of flux (φA in Figure 4.5-1) is produced by the stator current and passes through the same magnetic circuit as that for the flux produced by the rotor field winding (φDC in Figure 4.5-1). The top part of the figure shows how the resulting flux from the fluxes generated by a threephase balanced winding (where three-phase balanced currents flow) is constant and of value equal to 1.5 times the maximum value of the flux produced by each

• The flux produced by the armature distorts the main flux produced by the DC rotating field • The amount of change/distortion depends on Load and Power Factor φDC A

A ia

C

N×Ia

φA = 1.5 N Im

C

W

φR

C

N×Ic Gas/Air gap

φDC

A

r to Ro

r

ato St

B

ib

Load

N×Ib

B

ic Generator

B φR

φDC

φDC

φR

φA φA φA PF = 1 Mainly distortional effects

PF lagging Demagnetizing effect

φR = Resulting flux in machine φA = Armature produced flux φDC = DC field flux

Figure 4.5-1 Armature reaction.

PF Leading Magnetizing effects

210

CHAPTER 4

OPERATION AND CONTROL

phase. This resultant flux rotates at synchronous speed. The bottom part of the figure shows how the stator produced flux affects the rotor produced flux for unity, leading, and lagging power factors. This is the “armature reaction” effect. This is an effective flux path, and a relatively high value of reactance may be expected, in the range of 1.5–2.1 PU. The PU direct axis synchronous reactance is approximately equal to the reciprocal of the short circuit ratio. The stator produced flux acts together with the rotor produced flux to create the total “useful” (meaning linking both windings) flux, called the resultant flux (φR). The way the stator produced flux affects the rotor-produced flux is called the “armature reaction” of the machine. This can be clearly seen in Figure 4.5-1, where the bottom of the figure presents how the armature reaction affects the rotor-produced flux for three power factor conditions: unity, leading, and lagging. The armature reaction of the generator affects the voltage regulation of the machine (i.e. how the terminal voltage changes as the load changes, all other things remaining the same), see Figure 4.5-2. With lagging power factors, the armature reaction tends to accentuate the voltage drop in the machine, requiring additional DC current to be supplied by the exciter for compensation. How much armature reaction exists in a machine is the result of design compromises.

4.5.2

Equivalent Circuit and Vector Diagram

Chapter 1 introduced the reader to the most basic description of synchronous machine operation. In this section, the concept will be further developed, and the use of vector analysis will be illustrated with a few very basic and simplified examples. Figure 4.5-3 represents the generator’s basic equivalent circuit that can be used by any individual to solve simple application problems. The fundamental

Terminal voltage

PF leading

Rated PF unity

PF lagging

Field current left constant

Load current

Figure 4.5-2 How the armature reaction affects the output voltage of a generator for unity, leading, and lagging power factors.

4.5 BASIC OPERATION CONCEPTS

Xs

Ra

(One-phase diagram)

Ia Load

Zs

Eo

211

Et

Machine terminals E0 = Induced electromotive force (EMF) Xs = Synchronous reactance Zs = Synchronous impedance

All phase values

Et = Terminal voltage Ia = Armature (stator) current Ra = Armature resistance

Fundamental circuit equation E0

E0 = Et + Ia (Ra + jxs) E0 = Et + Ia × Zs

Zs Et

Ia

Load (MVA)

(Oneline diagram)

Figure 4.5-3 Generator equivalent circuit.

circuit equation in this relates machine variables to the connected system’s current and voltage (at the generator terminals). Figure 4.5-4 shows the vector representation of the fundamental circuit equation in the case of a synchronous machine acting as a generator and also shows the definition of regulation as it applies to an alternator.

4.5.3 Power Transfer Equation Between the Generator and the Connected System The power transfer equation is one of the basic equations in electric power engineering. It states: “The power transmitted between two points in an AC circuit is equal to the product of the magnitude of the voltages at both ends, times the sine of the angle between the two voltages, divided by the reactance between the two points.”

OPERATION AND CONTROL

s

Eo

jIaXs

Ia Z

δ

Unity load power factor Eo >Et

Et Ia Ra

Ia

Eo

I aZ s

δ

Et Ia R a

Ø

Lagging load PF

s

CHAPTER 4

Ia X

212

Eo >> Et

Ia

Eo δ

IaZs

Ia

ø Et

ø:

Load angle

δ:

Power angle

PF = Cos ø

Leading load PF

IaXs

Eo >< Et a I aR

No-load terminal voltage On-load terminal voltage

(%R) Regulation =

Et

Eo

× 100

Et

Figure 4.5-4 Vector representation of the fundamental circuit equation.

The maximum power that a circuit can deliver between two points is, thus, when the sine of the angle between the voltages equals 1, meaning the angle between the voltages equals 90 . Figure 4.5-5 illustrates the power transfer function as it applies between two electric machines, and between an alternator and the electric power system.

4.5.4

Working with the Fundamental Circuit Equation

The following two simple circuit problems, Case 4.3 and Case 4.4, with the generator connected to the system illustrate how the fundamental circuit equation, the power transfer equation, the active power equation, and a little basic trigonometry can be used to obtain solutions (see Figure 4.5-6).

4.5 BASIC OPERATION CONCEPTS

213

Power delivered The maximum amount of power that can be transmitted between two points in the system is: Ea X

Ga

× Sin δ

X Max power =

δ Gb

E a × Eb

Power =

Eb

Ea

Eb

Ea × Eb X

Generator supplying a system

Xs G

System Et

Eo

PD =

E × Et Xs

× Sin δ

Power

Maximum power

0

90°

δ (Electrical degrees)

Figure 4.5-5 Power transfer function applied to the power transferred between to generators and between a generator and the power system.

Case 4.3 Change in Excitation A generator is supplying power to the system. Now let us assume that the excitation is changed (IF1 to IF2), but the turbine output is not changed. Additionally, the system may be assumed to be much larger than that of the generator (“infinite” system) so that the frequency of the system (hence, the generator’s speed) and the voltage at the terminals do not change. Under these circumstances, it is desired to estimate how the PF, and the armature current Ia change. The solution of this simple problem can be found by inspection of the vector diagram in Figure 4.5-7.

214

CHAPTER 4

OPERATION AND CONTROL

“Infinite” utility bus G Eo

Xs

Load

Ia Et

(Typically: Xs >> Ra) neglect Ra

Eo δ φ

Ia × Xs × cos φ = Eo × sin δ

Ia X

s

φ

Et

Ia

Ia • Xs • cos φ = Eo × sin δ

Power delivered = Et × la × cos φ =

Eo × Et Xs

× sin δ

* In an “infinite” bus, Et taken as constant * Eo assumed linear with IF for small changes of IF

Figure 4.5-6 Graphic representation of the fundamental circuit equation.

The voltage induced in the machine (E) multiplied by the terminal voltage (Et) and by the sine of the angle between them (δ) represents the power transferred from the machine to the terminals (power transfer equation neglecting generator losses) is shown in Equation (4.8), E × E t × sin δ = Power delivered = Turbine s output constant

(4.8)

However, since, as was stated earlier, the terminal voltage does not change, we have Equation (4.9) E × sin δ = constant

(4.9)

But E × sin(δ) is the vertical projection of E. Changing the field current clearly changes E. So if E multiplied by sin(δ) must remain constant, then δ must change in such a way that the vertical projection is still the same.

4.5 BASIC OPERATION CONCEPTS

Ia1 Xs

E2

E1

δ1 δ2 φ1

φ1

Et

Xs I a2

215

Constant (PD = const) = la Xs × cos φ = E × Sin δ

φ2

φ2 Ia

1

I a2

= E1 × 1. E2 ~

IF2 IF1

2. E1 sin δ1 = E2 sin δ2

φ2 = sin–1

3. E2 cos δ2 = Et + la2 × Xs × sin φ2

E1 sin δ2 E2 la2 × Xs × sin φ2 = E2 × cos δ2 –Et

4. Ia2Xs cos φ2 = Ia1Xs cos φ1 Ia2 × Xs × sin φ2 Ia2 × Xs × cos φ2

=

E2 × cos φ2 – Et Ia1 × Xs × cos φ1

E2 × cos δ2 – Et Ia1 × Xs × cos φ1 Ia1 × Xs × cos φ1 φ2 = tan–1

5. Ia2 =

Xs × cos φ2

Figure 4.5-7 Graphic solution for change of excitation from IF1 to IF2.

Finally, we know that the power delivered equals Equation (4.10): P = E t × I a × cos φ

(4.10)

By combining both equations and introducing a little trigonometry, the solution to the problem can be found. Recommended Exercise: Repeat this simple example for your generator, using MVA, volts, frequency, and field current as they apply to any given load point. After calculating the new PF and armature current, use the OEM’s V-curves of the machine to calculate the new PF and current, and compare these with the calculated values. Numerical Example A 13.8 kV hydro generator, rated 500 MVA, is delivering 250 MW at 0.8 pf lagging. If the excitation is increased by 10% what are the changes in PF and stator amps? Assume infinite bus; turbine power unchanged, and Xs = 125%. The solution is presented in Figures 4.5-8 and 4.5-9.

216

CHAPTER 4

G

OPERATION AND CONTROL

Ia = 13 074 A 250 MW at 0.8 PF lag

Xs = 125%

Et = 7 967 V (Phase value) Et = la =

13 800 3

= 7 967 V 250 × 106 = 13 074 A 3 ×13 800 × 0.8

P 3 × EL–L × PF

XBASE =

KV 2RATED

13.82 = 0.38 Ω 500

MVARATED

Xs = 1.25 × XBASE = 1.25 × 0.38 = 0.48 Ω φ1 = cos–1 0.8 = 37° Et × E1

× sin δ1

PD × Xs 3Et E1

6 E1sin δ1 = 250 × 10 × 0.48= 5 020 3 × 7 967

φ1

δ1 Et

1X s

E1sin δ1 =

Xs

φ1

Ia

PD = 3

Ia

1

Figure 4.5-8 Numerical example for Case 4.3.

Case 4.4 Change in Power A 13.8 kV hydro generator, rated 500 MVA, is delivering 250 MW at 0.8 pf lagging. If the output power is increased by 10% what are the changes in PF and stator amps? In this instance, the turbine output is changed (from PD1 to PD2) while feeding an “infinite” system. Thus, the terminal voltage and frequency are kept constant by the system. The excitation field is also kept constant. In this case, the fact that the excitation is kept constant means that E is constant. Figure 4.5-10 shows how it is obvious that δ must change with the change in power when the power transfer equation is applied to this case. This fact, and a little geometry, lead to a simple solution of the problem with the deduction of stator current and the new power factor. Figure 4.5-11 provides a simple numerical example of finding the change in stator current and power factor of a generator feeding an “infinite” power system when the excitation is kept constant and the turbines output is increased by 10%.

4.5 BASIC OPERATION CONCEPTS

217

E1 cos δ1 = Et + Ia1 × Xs × sin φ1 E1 × sin δ1 E1 × cos δ1

= tan δ1 =

5020

= 0.43

7 967 + 13 074 × 0.48 × sin 37°

tan–1(0.43) = δ1 = 23°

5020 sin 23°

E1=

E1 sin δ1 = sin–1 E2

δ2 = sin–1

φ1 = tan–1

Ia1=

= 12 771 V

E2 × cos δ2 – Et Ia1 × Xs × cos φ1

E1 1.1 E1

= tan–1

× sin 23° = 21°

1.1 × 12 771 × cos 21° – 7 967 13 074 × 0.48 × cos 37°

~ = 46°

Ia1 × Xs × cos φ1 13 074 × 0.48 × 0.8 = 15 056 A = 0.48 × cos 46° Xs × cos φ2

Conclusions By increasing field current by 10% * Power factor moved from 0.8 to 0.7 (LAG) * Armature current increased from 13 074A to 15 056 A (15% increase)

Figure 4.5-9 Continuation of Case 4.3 numerical example.

4.5.5

Parallel Operation of Generators

More often than not, hydro generators are connected directly to a common bus, and from there, to a step-up transformer. When two or more generators have their terminals connected to the same bus, a number of issues may arise. The first is the existence of circulating currents. As in the case of transformers connected in parallel, generators in parallel are affected by circulating currents if voltages and impedance do not match. In the case of generators, there is an additional degree of freedom not found in transformers; the angle of the voltage between both machines. Any mismatch will introduce significant circulating currents, resulting in an exchange of VARs between the units. This results in unwanted losses and curtailment of available output from at least one of the units. Thus, it is important that the operators control the units’

218

CHAPTER 4

OPERATION AND CONTROL

Ia2 Xs

fE so cu Lo

φ1 Ia

δ1 δ2 φ2

s

E

φ2

Et

Ia1 X

E

1

1. PD =

Et × E × sin δ Xs cnst

sin δ2 = sin δ1 ×

sin δ∞PD PD2 PD1

2. PD = √3 .EtIa cos φ

Ia2× cos φ2 =

PD2 I .cos φ a1 1 PD1

3. From the figure: E·cos δ2 = Et + la2 × Xs × sin φ2 2+ 3 4. Ia2 =

φ2 = tan–1

Ia2 × sin φ2 =

E cos δ2– Et Xs

PD1(E × cos δ2– Et) PD2 × Xs × Ia1× cos φ1

PD2 √3 Et cos φ2

Figure 4.5-10

Graphical representation of change in power.

parameters in such a way that circulating currents are kept to a minimum. Figure 4.5-12 shows how the circulating current is calculated. Interestingly, circulating currents between two or more generators tend to reduce the angle of the terminal voltages of the units. The explanation is beyond the scope of this book, but can be found in Chapter 10 of Ref. [10]. However, if there is a tendency to increase the angle and one machine is delivering more power than the other, then a “hunting” situation might be established between the generators. These types of situations can be controlled by a fine-tuned AVR and operator input.

4.5.6

Stability

One of the most fundamental concerns when operating industrial generators (and synchronous machines in general) is that they may become “unstable” and, eventually, “out-of-step” (also known as “slipping a pole or poles”). As explained in Chapter 1, the operation of a synchronous machine is predicated on the rotor and stator fluxes aligning themselves and rotating together at synchronous speed.

4.5 BASIC OPERATION CONCEPTS

219

From previous example, we know: Et/ph = 7 967 V; la1 = 13074 A E = 12 771 V Xs = 0.48 Ω

φ2 = tan–1

= tan–1

PD1 (E cos δ2 – Et) PD2 × Xs × Ia1 × cos φ1

250 (12 771 × 0.9 – 7 967) 1.1 × 250 × 0.48 × 13 074 × 0.8

= tan–1 0.64 = 32.5° From E sin φ1 = Ia Xs cos φ1 δ1 = 23° From sin δ2 = sin δ1 ×

PD2 PD1

δ2 = sin–1(sin 23° × 1.1) = 25.4° cos δ2 = 0.9 cos φ2= cos 32.5° = .0.84 PD2 250 × 106 × 1.1 = 13 696 A Ia2 = = √3 Et cos φ2 √3 × 13 800 × 0.84 • PF increased from 0.8 to 0.84 • Armature current increased from 13 074 to 13 696 A (5%)

Figure 4.5-11

Numerical example for Case 4.4.

When the machine is loaded, a torque angle appears between both fluxes. Similarly, a power angle appears between the voltage induced in the machine (Eo) and the terminal voltage (Et). Recall from Section 4.5.3 that the power transfer equation determines the power flow in the machine, which is given by Equation (4.11): P=

Eo × Et × sin δ X

(4.11)

Thus, the maximum power the machine can deliver is given by Equation (4.12): Pmax =

Eo × Et X

(4.12)

This maximum power will occur when the internal generated voltage and the terminal voltage are 90 apart. However, if additional load is applied to the unit, resulting in the voltages being pushed apart beyond 90 , the capability of

220

CHAPTER 4

OPERATION AND CONTROL

Common bus at voltage V

Ic

E1

G2

G1

E2

Ic = (E1 – E2) / (Z1 + Z2)

• In the figure, all bold letters represent vector variables •

E1 and E2 represent the back-emf in each generator, i.e. the voltage generated in the armature, before the drop across the leakage reactance.

•

Z1 and Z2 represent the synchronous impedance

•

Ic is the circulating current

Conditions for synchronization are: 1. Same phase sequence 2. Same voltage 3. Remaining within: Maximum frequency slip Maximum phase angle

Figure 4.5-12 Calculation of circulating current between two generators connected directly to the same bus.

delivering the required power (and torque) will not be satisfied, and the rotor will come out of synchronism. This phenomenon, called out-of-step or slipping poles, is extremely onerous. Generators can suffer extreme damage under this condition. Therefore, it is the practice to operate a generator with its internal angle not reaching beyond 60 electrical degrees. Figure 4.5-13 represents a simplified mechanical equivalent of slipping poles. The maximum transfer of power limit applies to any branch or element of the circuit in which a reactance separates two voltages. For a broader perspective of this issue, let us examine it first from a system’s perspective.

4.5 BASIC OPERATION CONCEPTS

221

No tension on spring at this point Safe operation angle

Spring

Torque α

ω

At this point the unit goes out-ofstep

Figure 4.5-13

Out-of-step mechanical conceptualization.

Figure 4.5-14 depicts a simple transmission system comprising two lines connecting two busses. Both lines are transferring power Po from bus A to bus B. The top of this figure shows that under this condition, the steady-state point Po is well within the maximum power transfer capability of the two lines, meaning the lines can absorb a relatively large increase in transmitted power from A to B, without any stability concern. In mathematical terms, this is indicated by the angle δ0 < 90 . Now let us assume that line 2 breaker opens following a fault on it as seen at the bottom of Figure 4.5-14. The moment line 2 opens, the maximum capability to transfer power from A to B is given by the lower curve representing the capability of line 1. However, the power being transferred is still Po. The new equilibrium point, indicated by δ1, comes very close to the maximum capability of the system. Thus, a relatively small increase in load will throw the system into disarray. The system is now denoted as being unstable or marginally stable. A similar treatment can be applied to the generator delivering power to a system. Figure 4.5-15 shows a generator feeding a power system. At normal operation, the maximum capability of the system to transfer power is denoted by the higher curve in Figure 4.5-15a. Shown there is the operating internal angle δ0, which is significantly lower than 90 . As the system experiences a fault on one of its lines, the load P2 is removed and the generator feeds only the remaining P1. Now, the turbine does not (cannot) change its output instantaneously (the turbine keeps

222

CHAPTER 4

OPERATION AND CONTROL

A

B

P1

Line 1 X1

Ga

Gb

P0 P2

Line 2 X2

Power

Line 1 and Line 2 capability P0

δ

δ0 Condition before fault A

B

X1 Breakers opened X

Ga

P0

Gb

X2 Fault

Power

Line 1 and Line 2 Line 1

δ0

δ1

Max δ

Figure 4.5-14 the fault.

After line #2 opens, systems operates close to stability limit (max.power transfer from A to B)

Power system stability case with two lines and two busses before and after

“pushing” watts into the system), so δ advances toward 90 as the system tries to find a new equilibrium. The excess power between what the turbine delivers and the output of the generator goes into accelerating the unit’s rotors and is converted into spinning energy. Depending on the power transfer capability of the remaining system and the ratio between P1 and P2, the generator may or may not remain stable. If it does not, it will slip a pole (see Figure 4.5-13) or, if the protection is adequate, it will be

4.5 BASIC OPERATION CONCEPTS

223

φ Gap δ is

ax tor tic Ro gne ma

W

X

G P1

Line

P2

Fault Unstable Pole-slip?

A2 < A1

A2 = 132 MW-deg A1 = 235

Power/Angle sinusoids Active power (MW)

70

P1 + P2

64.5

60

51.5

50

0 18

59

120

180

(b)

Prior to fault

40 After clearance

30 20

23.4

10

During fault

A2 132

235

30 60 90 120 150 180 δ0

35 A1

0 0

138

0 18

138154

(a) Load angle (electric degrees) (c) Transient stability: equal-area criterion

Figure 4.5-15

Simple case of generator stability from the generator perspective.

removed from operation. In some cases, the system may recover fully or partially (shown by the middle curve in Figure 4.5-15a). In that case, there is a greater chance that the generator will stay connected and stable. Mathematically, calculating the areas between the intersection of the power transfer curves and the output power can provide an estimation of the stability (Figure 4.5-15b, c). These areas

224

CHAPTER 4

OPERATION AND CONTROL

represent the additional spinning energy that has gone into the rotors during acceleration. This energy must return to the system once the generator is again stable (i.e. its speed is the system’s synchronous speed). For a more in-depth study of stability issues, the reader is referred to [11]. 4.5.6.1 Transients and Subtransients In the context of power system applications, a transient state occurs while a system is undergoing major changes. This may be due to, for instance, faults, switching on or off large loads, or loosing large chunks of generation. At the same time, internally, the generator is also undergoing significant changes. Under such unsteady conditions, the changing flux produced by the changing stator current in the direct axis (parallel to the pole-faces of the rotor) induces a voltage in the field winding, resulting in a field current that opposes the change in flux and, hence, the change in stator current. This makes it more difficult for the stator produced magnetic flux to pass through the rotor poles than in the steady-state condition. Under the transient condition, only the leakage flux paths of the stator and field windings are available, meaning fewer flux linkages per stator ampere. The result is that the generator looks like a reactance in the range of 0.2–0.5 PU [12], which is much smaller than the synchronous reactance. This is called transient reactance, often denoted by X d. The transient reactance is important to understanding transient stability, which, as stated above, is the ability of the power system to recover from a short circuit that has been interrupted, perhaps by circuit-breaker action. “Subtransient” is used to describe a rapidly changing condition that may last one to four cycles (0.016–0.064 seconds in a 60 Hz system). In this case, the magneto-motive force (mmf ) of the stator winding changes so rapidly that it causes currents to arise in the rotor amortisseur as well as in the field winding, all of these opposing the change in stator current. This restricts the statorproduced flux to the stator leakage paths and to the amortisseur bars and pole-face. Therefore, the generator appears as a smaller reactance, in the range of 0.13–0.32 PU with amortisseur bars and 0.2–0.5 without amortisseur bars [12]. This is called the subtransient reactance, often denoted by X d. The subtransient reactance is commonly used to calculate the maximum current following in a sudden short circuit that has occurred nearby.

4.5.7

Sudden Short Circuits

If a short circuit occurs suddenly in the power system near a hydro generator, a high-current transient ensues, which is of interest for several reasons. In the design of the hydro generator, winding forces and torques experienced by the stator and torques on the shaft and rotor assembly must be adequately accommodated. Also, external buses and circuit breakers that must carry and interrupt the current must be adequately specified.

4.6 SYSTEM CONSIDERATIONS

225

For a sudden short circuit at the stator terminals, the exciter is assumed to be a source of constant voltage; it is not controlled by the voltage regulator. In addition, the generator appears to react in a linear fashion in terms of electrical and magnetic circuits. Each winding in the generator traps the flux, linking it at the instant of short circuit. The relationship is such that the flux linking such a winding does not change instantaneously. A large direct current suddenly appears in each phase of the stator winding in proportion to the flux linking it at the instant of short circuit, in order to sustain that flux. Since there is no source of direct current in the stator winding, it decays exponentially to zero in accordance to the stator time constant Ta (0.03–0.25s with amortisseur and 0.10–0.50s without amortisseur) [12]. Large direct currents also arise in the field winding and in the rotor iron circuit to sustain the flux trapped in them at the time of the short circuit. The field current decays exponentially according to the transient time constant T d (0.5–3.3s) [12] to the steady value supplied by the exciter. The rotor iron current decays in accordance with the subtransient time constant T d (0.01–0.05s) [12] to zero, since there is no source for direct current in the rotor iron circuit. Therefore, both a decaying trapped flux in the stator and a decaying trapped flux rotating with the rotor are present. Because of relative motion, the stator flux produces a decaying alternating current of power-system frequency in all elements of the rotor, and the rotor flux produces a decaying alternating current of the same frequency in the stator winding. At the instant of short circuit, the value of the DC component of current in each phase is equal and opposite to the instantaneous value of the AC component. Thus, there is no sudden change in current.

4.6 SYSTEM CONSIDERATIONS Numerous industry standards have been developed, both nationally and internationally, that specify the required performance of a hydro generator. These standards define limiting temperatures at rating, required characteristics, and steady and transient conditions that must be successfully tolerated. Such standards are found in IEEE, IEC, BS, VDE, and other industry publications. With regard to the hydro generator, its primary requirement is to provide electric power continuously or for peak or base load periods as needed, and to do so reliably and economically. A generator is also normally required to provide voltage support to the system by supplying the needed reactive power. The rated power factor assures that the generator will have adequate ability to carry out this function. The rating normally defines the continuous duty required of the generator. A temperature class is assigned to the generator, which defines the thermal capability of the electrical insulation systems of the stator and field windings. Hydro generators are generally Class 130(B), or Class 155(F), which implies a hot-spot capability of 130 C, or 155 C respectively, which is based on an ambient air of 40 C. For convenience, table 6 of Ref. [1] (Table 4.6-1) is included here as a reference for maximum observable limits.

226

CHAPTER 4

OPERATION AND CONTROL

TABLE 4.6-1 Limiting observable temperature rises of indirectly cooled salient pole synchronous generators and generator/motors for hydraulic turbine applications

Item

Machine part

1

Stator windings Vn = 12 kV or less 12 kV < Vna < 24 kV Vna > 24 kV

2 3

4 5

Method of temperature determination

Embedded detectorb Embedded detectorb Embedded detectorb Resistance Detector or thermometer

Temperature rise ( C) at 40 C cold coolant Class 130(B)

Class 155(F)

85c

105c

85a

105a

By agreement

By agreement

Rotor windings 80 100 Cores and mechanical parts, Not detrimental to the whether or not in contact with insulation of that part or any insulation adjacent part Collector rings Thermometer 85 85 Miscellaneous parts (such as brush holders, brushes, etc.) may attain temperatures that will not injure the machine in any respect

a

For machines with rated stator winding voltage Vn (line-to-line) > 12 kV, the temperature rise of the embedded temperature detector shall be reduced according to the following relationships: −1 C for each kilovolt or part thereof 12 < Vn ≤ 24 kV By agreement Vn > 24 kV

b

Embedded detectors are located within the slot of the machine and can be either resistance elements or thermocouples. Embedded detector temperatures shall be used to demonstrate conformity with the standard for generators so equipped.

c

These values are for insulation systems with thermosetting materials. For insulation systems with thermoplastic materials, Class l30(B) and Class 155(F) shall not apply, and the equivalent temperature rises shall be 60 C for Class 130(B).

The wave shape of the stator voltage must be very nearly sinusoidal to avoid certain environmental concerns such as telephone interference. Historically it was common to specify a limiting telephone influence factor (TIF), which is calculated from the harmonic content of the voltage by using a weighting-factor curve that reflects the frequency response of older telephone systems. A deviation factor limit should still be specified. This is a measure of the maximum deviation that the stator voltage has relative to a sine wave. The reader is referenced to [13], where the TIF measurement method has been replaced with Total Harmonic Distortion (THD). Ref. [1] is currently in revision and is expected to change the TIF requirement to THD. A voltage response ratio is specified for the excitation system to be compatible with the stability needs of the power system. A hydro generator must also be able to operate successfully in a real power system where the ideal is not always achievable. Therefore, other conditions that may be experienced by a hydro

4.6 SYSTEM CONSIDERATIONS

227

generator must be accounted for in the specification of its required capability, as discussed below.

4.6.1

Voltage and Frequency Variation

The operating conditions as per Ref. [1] are outlined in Section 4.4.10. The OEM may have designed the generator to operate with a wider or narrower window than those specified here. The reader is encouraged to contact the OEM to verify the design operating parameters.

4.6.2

Negative Sequence Current

The three phases of a power system are not perfectly balanced in voltage and impedance. Accordingly, a small amount of steady negative-sequence current is produced. The standards specify a maximum steady negative-sequence current that must be tolerated. This value, which varies among the various sizes of hydro generators, is based on an economic evaluation of system needs and generator rotor heating characteristics. A disturbance may occur on one phase of the power system, which is then isolated by circuit breakers. The event may subject hydro generators in the vicinity to a large negative-sequence current for a brief period. Recognizing the economics of providing tolerance for the rotor heating that would result from such an event, the industry standards require that a hydro generator be capable of withstanding a prescribed value of I 22 t, where I 22 is the square of the PU value of the negative sequence component of current, integrated over the period of exposure (t) in seconds. 4.6.2.1 Calculation of Negative-Sequence Currents As shown in Chapter 1, any three-phase symmetrical system can be represented as a group of phasors (vectors rotating at constant speed) of currents or voltages. For example, Figure 4.6-1a represents such a balanced system of phasors. On the other hand, once a fault in the grid occurs, or the grid is unbalanced for whatever reason, current phasors will show as unbalanced in magnitude and in phase angles. Figure 4.6-1b shows such a situation (the phasors in the figure could represent either currents or voltages). To avoid the loss of simplicity introduced by the one-phase equivalent circuit approach in solving three-phase circuits, symmetrical-sequence components were added (first developed by Charles LeGeyt Fortescue, 1876–1936) to the bag of tools electrical engineers use in solving these types of problems. What the theory of symmetrical-sequence components states is that any three-phase unbalanced system of vectors (representing currents or voltages or anything else) can be replaced by three sets of three balanced phasors each: one rotating in the same direction as the original vectors, called the positive-sequence set; one rotating in the opposite direction, called the negative-sequence set; and one pulsating at the

228

CHAPTER 4

OPERATION AND CONTROL

(a)

(b) C C A

A

B B Figure 4.6-1 Sets of balanced and unbalanced three-phase phasors. (a) Balanced set of phasors rotating (by convention) in the counterclockwise direction and (b) unbalanced set of phasors rotating (by convention) in the counterclockwise direction.

same frequency as the original set, called the zero-sequence set. Figure 4.6-2 shows the original unbalanced set of vectors, together with the set of symmetrical component. Figure 4.6-2 shows nine symmetrical-component phasors. However, the same as symmetrical three-phase circuits, unsymmetrical circuits can also be solved using only one phase with the aid of the symmetrical-sequence components. Therefore, when calculating the negative-sequence components, only one phase is required. For simplicity, the phasor representing phase A is always chosen. Graphical methods exist for finding the symmetrical-sequence components, and these can be found in many books. Here, the mathematical method is introduced by way of a simple example. For example, let us assume that the following unbalanced system represents currents in a given point in a circuit: A = 10 kA 0 that is, Phase A has a magnitude of 10 kA and lies at a 0 angle from the horizontal. “A” is then the reference phasor: B = 12 kA 250 C = 7 kA 110 To find A1, A2, and A0 (i.e. the positive, negative, and zero sequence phase A phasors), the following Equations (4.13)–(4.15) are used: A1 =

1 A + a × B + a2 × C 3

(4.13)

A2 =

1 A + a2 × B + a × C 3

(4.14)

229

4.6 SYSTEM CONSIDERATIONS

C1 C A

A1

B B1 Unbalanced set of phasors rotating (by convention) in the counterclockwise direction

A2 C2

Equivalent set of symmetrical components: The upper (counterclockwise rotating) set represents the positive symmetrical components The clockwise rotating set represents the negative-sequence set The lower set represents the zero-sequence set Common practice uses the following nomenclature: - Positive set uses the “1” subscript - Negative set uses the “2” subscript - Zero set uses the “0” subscript

B2

A0 B0 C0

Figure 4.6-2 An unbalanced set of three-phase phasors and its symmetrical component equivalent.

A0 =

1 A+B+C 3

(4.15)

Take note that A1, A2, A0, A, B, and C in the above equations represent vector quantities, that is, they have a magnitude and an angle. The operator “a” in the equations has the following meaning: • a rotates the vector 120 in the counterclockwise direction (i.e. +120 ) • a2 rotates the vector 240 in the counterclockwise direction (i.e. +240 ) Now, for the solution for calculating the symmetrical sequence components, let us write all the variables required for the solution as per the equations above,

230

CHAPTER 4

OPERATION AND CONTROL

both in polar and rectangular form. The transformation from polar to rectangular form of a vector is A = angle = A cos angle + j sin angle In the present example (vector-quantities are shown in bold): A = 10 0 = 10 cos 0

+ j sin 0

= 10 + j0

B = 12 250 = 12 cos 250

+ j sin 250

= − 4 10 − j11 28

C = 7 110 = 7 cos 110

+ j sin 110

= − 2 39 + j6 58

a × B = 12 250 + 120

= 12 cos 370 + j sin 370

a2 × B = 12 250 + 240 = 12 cos 490 = − 7 71 + j9 19 a × C = 7 110 + 120 = 7 cos 230 = − 4 50 − j5 36 a2 × C = 7 110 + 240

= 7 cos 350

= 11 82 + j2 08

+ j sin 490

+ j sin 230

+ j sin 350

= 6 89 − j1 22

Now, let us plug these values in Equations (4.13)–(4.15) above: A1 =

1 1 10 + j0 + 11 82 + j2 08 = 6 89 − j1 22 = 28 71 + j0 86 3 3

A2 =

1 1 10 + j0 − 7 71 + j9 19 − 4 50 − j5 36 = − 2 21 + j3 83 3 3

A0 =

1 1 3 51 − j4 7 10 + j0 − 4 10 − j11 28 − 2 39 + j6 58 = 3 3

To convert back from rectangular to polar coordinates, the following formula is employed. If A is a vector such as A = a ± jb then A =

a2 + b2

A = tan − 1

b a

In our example: A1 =

1 3

28 712 + 0 862 tan − 1

0 86 28 71

= 9 57 1 72

231

4.6 SYSTEM CONSIDERATIONS

1 3

A2 = A0 =

1 3

− 2 212 + 3 832 tan − 1 3 512 + − 4 7

2

tan − 1

3 83 − 2 21 −4 7 3 51

= 1 47 − 60 = 1 96 − 53 24

Let us now check the result. Another key equation with symmetricalsequence components is the following Equation (4.16): A = A1 + A2 + A0

(4.16)

Equation (4.16) states that the sum of the three symmetrical-sequence components equals the original reference phasor. In the present example, the sum of the symmetrical-sequence components found by calculation must equal vector A (= 10 ∠0 ). Then, A 1 + A2 + A0 = 1 3

28 71 + j0 86 + − 2 21 + j3 83

+ 3 51 − j4 7 = 10 + j0 = A This proves our calculations are correct. Figure 4.6-3 shows the original phasor A and its associated symmetrical sequence components. The main objective of the calculation was to find out the negative sequence component. The negative sequence component is normally defined PU or as a percentage of the average of all three original vectors. In our case, A2

= 1 47

10 + 12 + 7 3 = 0 152 PU = 15 2

This type of calculation would be required while analyzing how a generator is affected by the negative sequence component during a steady-state grid unbalanced condition. In fact, things are a bit simpler, because during grid unbalanced conditions, the angles between the phasors can be always taken to be equal to 120 . Also, during unbalanced grid conditions without a fault, zero sequence components are almost nonexistent, making the search for the negative sequence component much easier, as shown in the following paragraph. Under these steady-state conditions (grid unbalance), the value found for the negative sequence component is used together with Table 4.4-1 to evaluate whether the machine remains within its safe operating region, based on the published permissive negative sequence current values.

A1

Figure 4.6-3 Symmetrical sequence components.

A2

A0

A The phasor diagram above schematically shows the original phasor A and associated symmetrical sequence components

232

CHAPTER 4

OPERATION AND CONTROL

The solution of the previous example included a zero sequence component. It is interesting to note that this component is mostly nonexistent when calculating short circuit currents in any generator with high impedance grounding (or in the grid, when a fault is not present). This applies to all hydro generators discussed in this book. Let us now look at an example of calculating the negative sequence component for the case of a short circuit on the terminals of the generator. Consider the case of a hydro generator connected to a bank of three singlephase step-up transformers. The generator is a 72 pole, 60 Hz, indirect water cooled, 13.8 kV, 300 MVA machine. When the generator is delivering rated power, a major fault occurs on Phase A of the step-up transformers. The fault causes a full short circuit across the low-voltage winding. Figure 4.6-4 shows the currents flowing prior to the fault and during the fault. It is required to calculate the negative sequence current component and to evaluate if the unit remained within the negative sequence current tolerance capability curves published in the standards. For this transient event, Table 4.4-2 must be used. 4.6.2.2 Solution Inspection of Table 4.4-2 reveals that for a salient pole generator, the maximum permissive (I2)2t is equal to 40. The most practical approach would be to read the values of the generator terminal currents during the fault, directly off the unit’s digital fault recorder (DFR). Using the three fault current values converted to PU, and using the time from DFR for the fault duration, the negative sequence current components of the fault can be found. In this case, any value exceeding 40 would be an indication that permissible levels have been exceeded and an inspection of the stator winding and rotor pole-face, pole keys, and amortisseur assembly would be warranted. One might expect all units this size to have installed DFRs in order to help diagnose this multimillion dollar asset, however, the reality is, they do not. Suppose now that the DFR values are not available. In this case, one must calculate the fault currents. Hopefully, by looking at what relays cleared the fault and their settings, one can estimate the duration of the fault. Let us assume that the fault cleared by the differential relays lasted 60 cycles (1 second). The following Equation (4.17) as found in any good book on power systems analysis (for instance [14]), gives the phase-to-phase subtransient short circuit current at the terminals of a generator: I k = 1 08

V LL Xd + X2

1 08

V LL 2 × Xd

(4.17)

In the present example, X d = 0.22 PU. Therefore, I k = 1.08 × 1/(0.44) = 2.45 PU. At this stage, a few adjustments ought to be made. For instance, it is known that at the start of the fault a DC component will be present. Depending on a number of factors, this DC offset may result in the current being at the onset of the fault,

4.6 SYSTEM CONSIDERATIONS

IA

SUT

233

Phase A

Phase B

GEN

Grid IB Phase C zN

IC IC

ZG IA Currents flowing during normal operation IB

IA

SUT

Phase A

IB Phase B

GEN IC ≈ 0

Grid

Phase C

zN ZG

IB

IA Currents flowing during the phase-to-phase fault

Figure 4.6-4 Generator subjected to a phase-to-phase short circuit on its terminals.

close to twice its AC value. Although the DC offset decays rapidly, it can be assumed to be present until the fault is cleared. A conservative approach may entail taking the DC offset current and adding it up, or some of it, to the value I k, to obtain an equivalent short circuit current. For instance, the value of 1.8 is taken by some authors [14]. In this case, the short circuit current will be I k = 1.8 × 2.45 PU = 4.42 PU.

234

CHAPTER 4

OPERATION AND CONTROL

To calculate the negative sequence current component, Equation (4.14) presented earlier in this section, is used: I2 = I2 =

1 3

4 42 0 + a2 × 4 42 180

=

1 3

4 42 1 0 + 1 420

= 2 55 30 PU that is, the magnitude of the negative sequence current is equal to 2.55 PU. The rated current of the generator is equal to 300 MVA 1 732 × 13 8 kV = 12 5 kA Therefore, the kA value of the negative sequence current is 12 5 × 2 55 = 32 kA Then, I 2 2 t = 2 55 2 PU × 1 second = 6 5

40

which is well within the permissive region of operation. However, this does not take into account the fact that the magnetic field and thus, the output current of the generator does not go down to zero immediately upon opening of the main breaker. The field current is discharged via a shunt resistor or the exciter circuit in a static exciter. The actual decay may take several seconds. Let us say we expect the field to dissipate within one second. In that case, the total (I2)2t can be approximately calculated in the following way: (I2)2t (including field discharge time) is approximated by 6 5 + 2 55 2

2

× 0 5 = 7 31

In the calculation, half of the value of the original negative sequence component is used, as an average for the decaying current, and half the time of dissipation of the field. The result is still within the permissive calculated value of 7.31. The authors want to make sure the reader understands that this is a simplified example. For a real case incident, all factors related to the fault should be investigated, and the generator manufacturer should be consulted. Different people have differing approaches for calculating these quantities, and the responsible engineers in charge of analyzing the event should take a more comprehensive look at the situation and available information. It is important to note the fact that the calculations may show that the unit has not crossed into the permissive (I2)2t region and does not indicate that the rotor did not sustain some damage. Bear in mind that the design formulas in place with each manufacturer to assure compliance with this (and other) criterion are not infallible. Conversation with the manufacturer should be initiated to ensure the generator has not suffered inadvertent damage.

4.7 GRID-INDUCED TORSIONAL VIBRATIONS

4.6.3

235

Over Current

The stator and field windings may withstand periods of over current. For example, if the system voltage drops for a brief period, the excitation system may be called upon to apply ceiling voltage to the field winding. The field current will rise according to its time constant from the initial value to a value that is higher than rated value. The higher than normal current in both windings would result in a brief excursion to higher than normal temperatures. Accordingly, industry standards require that a generator be capable of operating at specified levels of over current in the stator and field windings for a prescribed period of time (see Ref. [1]).

4.6.4

Current Transients

Current transients may occur in a power system, for example, due to a sudden short circuit or due to switching when the voltages of the circuits to be connected are unequal in magnitude or phase angle. The high currents produce high electromagnetic forces in the stator winding in the end regions and in the slots. They also result in transient torques felt by the rotor and the stator. To ensure that the hydro generator has the necessary robustness, Ref. [1] standard requires that it be capable of withstanding, without mechanical injury, a three-phase terminal, sudden short circuit while at load and at 105% voltage.

4.7 GRID-INDUCED TORSIONAL VIBRATIONS 4.7.1 Determination of Shaft Torque and Shaft Torsional Stress The calculation of these stresses is rather complicated and beyond the practical approach of this book. The expertise required almost never resides within a single utility. The OEM and specialists are the most qualified to carry out calculations on torques and stresses. Although subsynchronous analysis can be done successfully with lumped masses and spring models, an analysis for higher harmonics (supersynchronous) requires complex models of the entire train, and, often, the approach that yields the better model is the one in which the parameters of the model are taken from actual tests. Information about the torsional model of a unit can be obtained by creating short circuits at the vicinity of the generator and reading the torsional response in a number of locations of interest along the train. This type of test is like a “bump” test, eliciting responses in a wide range of frequencies. Another test is carried out by injecting different frequencies through the excitation system. In such a way, responses to specific frequency stimulus can be obtained.

236

4.7.2

CHAPTER 4

OPERATION AND CONTROL

Material Changes Due to Torsional Vibrations

Metallic shafts may fracture due to a sudden application of torque beyond their maximum capacity to carry torque, or may fail due to torsional fatigue due to numerous twisting cycles. There are a number of mechanisms driving this type of failure, such as high- and low-cycle fatigue, and specific crack initiators such as shaft fillet areas, or fretting fatigue areas such as shaft keyways. Also crucial is the issue of critical crack and crack growth. A most important fact is that torsional fatigue is lifetime limited. The graph in Figure 4.7-1 provides a view of loss of life of a shaft due to cyclic torsional oscillation. The graph shows, for each strain level applied to a shaft, the statistical expected number of cycles before crack initiation. This graph will be different for different metal alloys and geometries.

4.7.3

Types of Grid-Induced Events

There are a number of grid events that may result in torsional oscillations of the generators. Depending on a number of factors, such as the intensity of the event as measured at the generator terminals and the type of generator prime driver characteristics, the grid-induced events may or may not produce deleterious results (i.e. damage to some of the unit’s components). Some “events” are really not events, but conditions that may be permanent fixtures on parts of the power system. The following subsections list some of the grid events and conditions that may induce torsional vibrations in generators.

10–1

High cycle fatigue

Strain

Low cycle fatigue

10–4 102

Fatigue cycles

107

Figure 4.7-1 Typical life endurance of a shaft under periodic torsional strain.

4.8 EXCITATION AND VOLTAGE REGULATION

237

4.7.3.1 Short Circuits The strength of the torsional effect in a given generator emanating from a specific grid short circuit depends on a number of factors, such as electrical distance to the plant, voltage level of the affected system, speed of fault clearing, whether or not there is automatic reclosing, type of short circuit (one phase to ground, phase to phase, or three phase), and whether or not there are series and/or shunt capacitors on that system. Auto reclosing has long been identified as having a major potential magnifying effect on the oscillations of a generator. 4.7.3.2 Negative-Sequence Occurrences These normally arise during a short circuit or another grid imbalance. Large short circuits at the terminal of the generator (or on one of the transformers connected to the generator) will give rise to large negative sequence currents to a degree that may require removal of the rotor for inspection. However, most negative sequence events will be within the capability of the generators. 4.7.3.3 Out-of-Phase Synchronization This is one of the most onerous events a unit can experience. Out-of-phase synchronization is a step load applied to the generator when the main breaker is closed. From the point of view of torsional vibration, it is like applying a “bump” stimulus to the generator. This is a wide-band stimulus that will elicit response in a number of resonance frequencies. Some may reach damaging amplitude if the step load is large enough. Induced, electromagnetic and mechanical transients can last up one second or so. A severe out-of-phase synchronization incident can render a unit inoperable for a long time. Thus, proper protection against this type of occurrence is essential. This topic is covered in Chapter 6. 4.7.3.4 Out-of-Step Event Also called “slipping poles,” this is another very serious event that will cause oscillations among the unit’s components. Protection against this type of operating condition is also covered in Chapter 6. 4.7.3.5 Load Rejection When a large amount of load is lost (for instance, if the plant becomes an island due to some grid event), a negative-load step is applied to the generator. This event will also induce torsional oscillations but, for the most part, these events can be handled successfully by the generators.

4.8 EXCITATION AND VOLTAGE REGULATION 4.8.1

The Exciter

The exciter supplies direct current to the field winding of the generator, at whatever voltage is required to overcome the resistance of the winding. The rating of the

238

CHAPTER 4

OPERATION AND CONTROL

exciter is specified as its output power, current, and voltage corresponding to the rating of the generator, taking into account the temperature limits of the generator’s field winding. The exciter rating generally has some margin over this requirement, as defined when the generator is designed. The most common type of exciter used in early years was the commutatortype DC generator and is a fascinating piece of engineering and construction. This is very rarely used for new generators today. Any of the following systems usually supply the newer hydro generators: • A shaft-driven alternator with solid-state diode rectifiers • A solid-state, thyristor-based rectifier supplied by a transformer, deriving its power from the power system or from the generator’s output • A shaft-driven alternator with its output winding on the rotor, its output rectified by rotating solid-state rectifiers (commonly called a “brushless exciter”) The normal function of the exciter is to provide the proper level of direct current to the generator field winding, as required for the apparent power being supplied to the system, the terminal voltage, and the power factor of the generator load. In addition, the exciter must also be able to produce a ceiling voltage (which is the maximum exciter voltage) and to operate at that condition for a specified brief period, as required by the voltage response ratio, which is specified in excitation system’s specification. The voltage response ratio is a measure of the change of exciter output voltage in 0.5 seconds when a change in this voltage is suddenly demanded [15]. When the exciter is a rotating machine driven by the generator shaft, it becomes part of the hydro generator shaft system. It must be designed to accommodate axial movement due to thermal expansion and vertical motions of the generator shaft due to thrust being applied to the turbine.

4.8.2

Excitation Control

4.8.2.1 Steady State With the governor fixed and, therefore, the active power output of the generator fixed, and with the configuration of the power system fixed, an increase in exciter output, that is, in generator field current, causes the stator voltage to try to rise. This changes the power factor and causes the reactive power delivered by the generator to increase. While the governor responds to provide the power needed by the system, the exciter enables the generator to provide the needed reactive power and, thus, to help provide the needed voltage support in the system. The control system includes a voltage regulator that causes the generator’s field current to be at whatever level is required to maintain the stator terminal voltage at a selected value. The control system also can be instructed to hold the generator field current at a desired value when voltage regulator is not needed. This is done by “manual control.”

4.9 REFERENCES

239

A lower limit is provided so that the field current is not reduced to the point where stability margins are compromised. An upper limit is provided so that the capability of the exciter and that of the generator field winding are not exceeded. Volts/Hertz protection is commonly provided to prevent the level of the magnetic flux in the generator and in the unit step-up transformer from exceeding safe levels. A Volts/Hertz control is occasionally specified to adjust generator excitation so as to avoid over fluxing. 4.8.2.2 Transient The ability of the excitation system to change the generator field voltage rapidly may be important to system stability. Stability may be difficult to achieve when the system supplied has relatively high reactance; for example, when a long transmission line separates a generator from its load. In such a situation, providing an excitation system with a high voltage response ratio may help in the system’s design. It can help reduce major expenses in additional transmission line construction. A relatively new concept made possible in part by the use of thyristor power rectifiers is the high-initial-response excitation system. In such a system, the output voltage of the exciter changes almost instantly on command, enhancing system stability. Another concept in excitation control function is the power system stabilizer (PSS). It operates to enhance stability in situations where one power system may swing at low frequency relative to another (i.e. subsynchronous resonance conditions).

4.9 REFERENCES 1. IEEE C50.12 (2005). IEEE Standard for Salient-Pole 50 Hz and 60 Hz Synchronous Generators and Generator/Motors for Hydraulic Turbine Applications Rated 5 MVA and Above, IEEE. 2. IEEE 492 (2011). Operation and Maintenance of Hydro Generators, IEEE. 3. Beckley, P. (2002). Electrical steels for rotating machines. In Electrical Steels for Rotating Machines, IEE. 4. Milanicz, D. P. (1994). Reactive Capability Limitation of Synchronous Machines. IEEE Transactions on Power Systems, Vol 9 (No.1) 29–40. 5. Adibi, M. A. (1994). Reactive capability limitation of synchronous machines. IEEE Transactions on Power Systems 9(1), 29–40. 6. Walker, J. H. (1953). Operating characteristics of salient pole machines. Proceedings of the IEE – Part II: Power Engineering 100(73), 13–24. IEE. 7. Research done at Ontario Hydro by G.C.Stone and H. Sedding in early 1990s—no literature reference—only verbal confirmation. 8. IEEE (2003). C37.106-2003: IEEE Guide for Abnormal Frequency Protection for Power Generating Plants, IEEE. 9. Farnham S. B. and Swarthout, R. W. (1953). Field excitation in relation to machine and system operation. Transactions of the American Institute of Electrical Engineers. Part III: Power Apparatus and Systems 72, 1215–1223. AIEE. 10. Say, M. G. (1978). Alternating Current Machines, Pitman.

240

CHAPTER 4

OPERATION AND CONTROL

11. Anderson, P. and Fouad, A. (2003). Power System Control and Stability, IEEE Press. 12. Westinghouse, E. (1964). Electrical Transmission and Distribution, Westinghouse Electric Corporation. 13. IEEE-115 (2019). Guide for Test Procedures for Synchronous Machines, IEEE. 14. Roeper, R. (1972). Short-Circuit Currents in Three Phase Networks, Siemens Pitman. 15. IEEE (2014). 421.2-2014: IEEE Guide for Identification, Testing, and Evaluation of the Dynamic Performance of Excitation Control Systems. New York, IEEE. 16. IEEE (2015). C57.12-2015: IEEE Standard for General Requirements for Liquid Immersed Distribution, Power, and Regulating Transformers.

4.10 FURTHER READING Dorsey, S. F. and Smedley G. P. (1956). The influence of the fillet radius on the fatigue strength of large steel shafts. Proceedings of the IME-ASME Internal Conference on Fatigue of Metals.

CHAPTER

5

MONITORING AND DIAGNOSTICS

Generator operation should be kept within design limits for optimum performance and to maintain reliability and longevity of the equipment. Monitoring of the online performance of the generator is done by using installed sensors and instrumentation; therefore, good sensor information is critical in making a correct diagnosis. Good sensor information includes ensuring that the right sensors are installed in the right places, and that they are in good working order. In addition, the information from the individual sensors is often used in conjunction with other monitoring information, to make a more detailed and useful diagnosis. For instance, if all stator winding temperatures are hotter than normal and the stator current is above maximum allowable, then one would conclude that there is an overload situation that may be correctable by nothing more than reducing load. However, if the same sensors are in alarm when the machine is perhaps only at three-quarters load, then one would conclude that some other problem is present, such as low cooling water flow, plugging of the cooler tubes, plugging of the stator ventilation ducts, or stator winding water ducts. At this point, it is the additional sensor information that would be used in trying to diagnose the problem. But the initial diagnosis was actually done by a combination of information from two sensors: the stator winding temperatures and the stator current. It is easy to see how monitoring can help avoid major failures before they happen, by early warning of problems. It is also easy to see that the more extensive the monitoring is, the more that can be determined during operation. This allows more flexibility in operation by knowing more about the performance of the machine. It may even be possible to extend the generator life by adjusting the operation to avoid known operating regimes or ranges that cause some generator parameters to exceed their limits. It is important that, for whatever sensors are installed in any particular generator, the most efficient use is made of the information from each sensor. There are numerous

Handbook of Large Hydro Generators: Operation and Maintenance, First Edition. Glenn Mottershead, Stefano Bomben, Isidor Kerszenbaum, and Geoff Klempner. © 2021 The Institute of Electrical and Electronics Engineers, Inc. Published 2021 by John Wiley & Sons, Inc.

241

242

CHAPTER 5

MONITORING AND DIAGNOSTICS

types of generator monitoring systems and/or approaches to generator monitoring, but they are not always equal in their effectiveness. Some machines are minimally equipped with installed sensors, whereas others have as much installed instrumentation as possible, including specialized monitoring devices. Generally, the highly equipped machines are large units for which outage time is extremely costly. This chapter covers some of the different approaches to generator monitoring and the level of monitoring that can be accommodated. Included is a description of the various monitoring devices presently available, along with how they are generally used to determine abnormal operation, and how they could be further used in various types of monitoring systems. An example to which the reader is referred to is Ref. [1]. Readers are required to review, understand, and apply the safety procedures found in Section 7.2.1.

5.1 GENERATOR MONITORING PHILOSOPHIES Generator online monitoring and diagnostics covers a wide range of approaches, from minimal monitoring with few sensors and simple alarms up to elaborate expert systems with extensive diagnostic capability. The level of system sophistication for the most basic monitoring allows the operator simply to keep track of generator operation by periodically checking various operating parameters on the gauges and indicators provided. Some sensors that are connected to an alarm system can also give warning when a static high limit is reached. As hardware and software capability in data acquisition systems and computers has progressed, the ability to provide better monitoring has increased dramatically. Modern systems can handle dozens of sensors at one time and scan them for information far more frequently. Combining sensor inputs provides “intelligent” indications of problems that would otherwise not be foreseeable. This is done by computer modeling that predicts how various generator components will react during load changes and operating events. The use of such techniques allows closer tracking of sensors and the ability to diagnose problems at a much earlier stage in their progression. The information gathered from sensor readings by the monitoring systems can now more readily be stored as an archived history of the performance of a machine. This can be used for long-term trending and maintenance management of the equipment. In addition, the use of graphical user interfaces has allowed more meaningful presentation of the data collected so that operators can interpret the information faster and more accurately. Readings are presented in both numeric and graphical form. Short-term trends are used to compare various operating parameters as operators attempt to diagnose problems based on such things as temperature rise with increase in load.

5.2 SIMPLE MONITORING WITH STATIC HIGH-LEVEL ALARM LIMITS

243

Since generator monitoring has become very sophisticated, so-called expert systems are available but are also expensive. In some cases, high-level sophistication is not necessary. A utility must assess its needs based on the equipment under consideration. A large 700 MVA unit may warrant the installation of an expert system, whereas a 100 MVA machine may only need a more basic level of monitoring to suit the particular needs and philosophy of the user. When deciding on the correct approach for generator monitoring of any machine, it is always a good idea to first understand what the needs are. Then, one must understand the monitoring options that may be employed and know the cost to provide them. The sections that follow in this chapter provide descriptions of some of these types of monitoring approaches that may be used.

5.2 SIMPLE MONITORING WITH STATIC HIGH-LEVEL ALARM LIMITS Simple monitoring implies that the generator itself has very few sensors installed, and that only the most necessary and basic operating parameters are selected for permanent monitoring. Alarms are generally set at some warning level before the static high limits are reached. This is in an attempt to warn the user that something is progressing in the wrong direction and that remedial action may be necessary shortly. Generally, all generators have their main electrical parameters connected to a computer data acquisition system so that the operators are aware of the load point of the machine and where they are operating in relation to the limits of the generator (e.g. Supervisory Control and Data Acquisition SCADA or Distributed Control System DCS). The main electrical parameters include megawatts, megavars, stator current, terminal voltage, frequency, field current, and field voltage. All of these have operating limits that, if exceeded, can cause damage to one or more of the generator components. In addition to the electrical parameters, there are operating values that must be monitored to ensure that the generator operating limits are followed. Some of the more critical parameters include speed, temperature, airgap, bearing oil temperatures, stator cooling water temperature, pressure, and conductivity (for water cooled windings), bearing vibration, and raw service water temperature. These critical parameters tell the operator something about the condition of the generator or its components. In addition, all have specific operating limits that, when they are exceeded, have certain consequences. Not all parameters have the same level of priority. For example, exceeding the temperature limit of a component by 10%, while at steady-state operation, may be less of a concern than exceeding a vibration limit by 10% on another component. It really depends on the machine design and operating history. It is always a good idea to make sure the operations staff is well informed of all maintenance issues on the machine. In addition not all parameters have to be monitored to guarantee proper operation of the generator. For example, for a large indirectly cooled stator winding,

244

CHAPTER 5

MONITORING AND DIAGNOSTICS

if all of the stator winding temperatures are normal during the on load condition, it tells an operator that there must be water flow to the surface air coolers. Therefore, it is not absolutely necessary to monitor the surface air raw cooling water flow itself. Monitoring the stator cooling water outlet temperature is sufficient to safely operate the generator and detect when problems are occurring. There is also a wide range in the number of sensors installed by the various manufacturers. There are some machines with no core thermocouples and only a handful of stator winding thermocouples (TCs) or resistance temperature detectors (RTDs) installed, and some that have as many as a dozen core TCs and several dozen RTDs or TCs in the stator winding. The variation is extensive and is generally dependent on the manufacturer. However, when purchasing a new machine, the user can specify how many sensors and what type to be included in the machine when delivered. Although there is a cost associated with these extra sensors, it is a negligible amount of money when considering the total cost of the generator and the valuable information provided.

5.3 DYNAMIC MONITORING WITH LOAD VARYING ALARM LIMITS Regardless of what type and number of sensors are installed and connected to the monitoring system, it is always a good idea to get the most out of what is monitored. A more effective way to get the best use of installed sensors is by dynamic monitoring. Dynamic monitoring simply means having alarm limits that change in relation to the generator load point, rather than waiting for a high-level alarm limit to be reached. Relying on static high limits can sometimes mean that a problem has progressed too far to allow any meaningful corrective action by the time the alarm limit is exceeded and the operator is notified by the alarm. The premise behind dynamic monitoring is to mathematically predict what a particular sensor or group of sensors should be reading at any operating point and compare it with the actual sensor reading. The difference between the two can be closely monitored, and if the deviation is more than a previously determined limit, it can be alarmed or brought to the attention of the operator. The obvious advantage is that much earlier warning can be obtained. Also, it makes it possible to look for long-term problem trending, as well as more immediate failure modes. To do this, a mathematical model of the generator parameter being monitored must be available. Then it has to be customized to the particular machine being monitored. Subsequently, this mathematical model becomes an artificial sensor or an indicator of a problem. The following is a brief description of how this type of sensor is constructed. One of the best examples of an artificial indicator built from sensor readings into a mathematical model is that of water cooled stator winding hose outlet temperature measurement. To build this indicator in its simplest form, one must look at what affects the stator winding temperature during all modes of operation, but specifically when the generator is connected to the system and loaded.

5.3 DYNAMIC MONITORING WITH LOAD VARYING ALARM LIMITS

245

In the case of the stator winding hose outlet sensor, we are not concerned with stator winding temperature when the generator is offline since no current flows in the winding. Fault current could flow when the generator is on open circuit and a failure of the groundwall insulation occurs. However, this is a case in which generator ground fault relay protection comes into play, and stator winding temperature monitoring is a secondary issue. The main concern is the temperature of the stator winding when the machine is online and stator current is flowing in the winding. To begin, the stator winding hose outlet temperature will be at least that of the water inlet temperature. Therefore, the first component of a stator bar/coil hose outlet temperature, Tout, will be the temperature of the cooling water in, Tin. Stator bar/coil temperature will increase as electrical current flows in the copper of the winding. The relationship of temperature to electrical current is well known as T α I2. Therefore, if the generator is at full load while the stator current is theoretically at its maximum (Iref), then the temperature of the stator bar/coil hose outlets will be some temperature above the cooling water inlet temperature. The difference between the cooling water inlet and outlet temperatures will be the temperature rise, dTref, at this reference load, due to the heat input from the stator bar/coil I2R losses. The temperature difference between Tout, and Tin will obviously change as the generator loading (operating stator current, Is) is increased and decreased. Applying the relationship T α I2, we can use Is and Iref in the form (Is/Iref)2 to account for generator load changes. Therefore, the basic formula to calculate stator winding hose outlet temperatures can be written as shown in Equation (5.1) T out = T in + dT ref

Is I ref

2

(5.1)

In the relationship above, we can see that the portion of the function (Is/Iref)2 is equal to one, as it should be, when fingerprinting of the stator winding temperatures is done at the reference load. As Is becomes lower, at lower loads, the temperature calculated for Tout will decrease proportionally [2]. Using the formula, the difference between the measured reading and the calculated value can be closely monitored. An alarm value (e.g. 5 C) can then be added to the calculated value to produce the dynamic alarm limit as shown in Equation (5.2): T alarm = T out + 5 C

(5.2)

If the deviation is more than the calculated alarm limit, Talarm, it is then brought to the attention of the operator (see Figures 5.3-1 and 5.3-2). It should be noted that the preceding algorithms are in their simplest form. Other factors must be included for complete accuracy. The manufacturer generally knows these factors. For example, the stator bar/coil expected temperature calculation can also be enhanced, to include a factor allowing for variable coolant flow in water cooled stator windings. When implementing these types of models, the utility should consult the manufacturer before implementation. The distinct advantage of using this type of indicator in conjunction with direct sensor readings is the capability to predict expected values over the entire

246

CHAPTER 5

MONITORING AND DIAGNOSTICS

46 44

47

48

45

1

90

2

3

°C

80

4

5 6

70

43

7

60

42

8

50

41

9

40

40

10

30 39

11

20

38 SCW outlet # 37

10

12

0

13

36

14

35

15

34

16

33

17 32

18 19

31 30 29 Static high temperature limit Dynamic temperature alarm Measured values

28

21 27 26

25

24

23

20

22

Figure 5.3-1 Polar graph of instantaneous temperature magnitude for all hose outlets monitored.

100.0

High limit Dynamic alarm level Measured SCW outlet temp SCW inlet temperature Average stator current(kA)

90.0

Temperature (°C)

80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 0.0 0

10

20

40 30 Time (minutes)

50

60

Figure 5.3-2 One hose-outlet sensor indicating plugging as temperature in relation to time.

5.4 ARTIFICIAL INTELLIGENCE (AI) DIAGNOSTIC SYSTEMS

247

load and power factor range of the generator, and compare them to the actual readings. This allows a much improved and closer degree of monitoring on specific generator components, rather than simply relying on a maximum limit before an alarm is incurred. Using this dynamic monitoring method, one can look for deviations of only a few degrees above normal (for temperature relationships), at any load, and be provided with much faster warning of impending problems in the generator, long before measured parameters get anywhere near their absolute limits. This technique can also be used in indirect cooled generators to minimize thermal cycling by keeping the machines at a more constant temperature by monitoring the surface air cooler water flow and subsequent air temperature exiting the cooler. It can also be used to reduce the amount of cooling water used during operation which for some installations is important as this water is often taken from the penstock and is consequently not used for power generation. Change in water flow rates in the surface air cooler should only occur within the design limits of the cooler.

5.4 ARTIFICIAL INTELLIGENCE (AI) DIAGNOSTIC SYSTEMS Monitoring systems with diagnostic capability using artificial intelligence software are more commonly referred to as expert systems. The purpose of an expert system for monitoring and diagnostics of large generators is to collect, analyze, and interpret generator and auxiliary system sensor information, and provide early diagnosis of developing problems in the generator and its associated systems. The expert system should provide an easily understandable description of the suspected problem and recommendations to correct it, or bring the unit to a safe operating condition in a timely and appropriate manner. The main advantage of an expert system is its ability to look at all available sensor data in real time, correlate it as an “expert” would, and continuously update the diagnosis based on changing sensor readings. This allows operators to react quickly at the onset of the majority of generator problems that may be experienced during operation and avert major failures. In additin, an expert system gives station maintenance engineers a tool to closely monitor and log the performance of the generator, and make better maintenance decisions from the data collected during operation. There are a variety of types of expert systems in use today on large rotating machines. Their deterministic capabilities rely on such methods as rule-based systems, pattern recognition, neural networks, and Bayesian belief networks. Regardless of the type of expert system implemented, all have a number of common elements. These include a knowledge base containing the equipment facts, component relationships, and mathematical models; data acquisition hardware and software for sensor inputs; AI (artificial intelligence) software, more commonly referred to as an inference engine, to perform the reasoning function between the knowledge base and sensor inputs; a graphical user interface to allow the

248

CHAPTER 5

MONITORING AND DIAGNOSTICS

operator to interact with the system; installation software to allow changes and updates to be made to the system by the experts; and in some cases a simulator for offline testing and training. Within the various elements of the expert system, some interesting and novel techniques have been developed to provide accurate analysis of impending problems. These include mathematical modeling techniques used for logical and probabilistic determination of large generator problems. Also, methods of combining sensor inputs to create mathematical indicators of problems and techniques for dynamic tracking of problem indicators over the full power factor and load range of the generator have been developed. An expert system consists of a computer for monitoring and processing of data, external data acquisition hardware for collection of the generator and auxiliary systems sensor inputs, and the software that forms the basis of the expert system. The data acquisition system is used to collect raw sensor data from the generator and auxiliary systems. The number of sensors monitored will vary, depending on the particular generator and how extensively it is instrumented. The expected readings of the monitored sensors are determined during installation and configuration of the expert system, by “heat run and vibration signature” tests on the generator at various loads and power factors to fingerprint the expected machine’s behavior. The fingerprint data is used to produce scaling factors for specific formulas developed to track sensor inputs over the entire load and power factor range (i.e. dynamic tracking), and to set maximum sensor limits according to insulation class, machine rating, and other machine specific parameters such as shaft, frame, and core vibration. Such formulas are proprietary to the company’s expert system and can vary from company to company. Thus, it is important for the user to research which expert system will best suit the fleet of generators that need to be monitored. Examples of sensors are those providing direct temperature readings from thermocouples and RTDs, pressure readings, voltage measurements, current measurements, and equipment status (breaker open/closed, pump on/off, tank level high/low, etc.). These are generally the instruments that are hard wired directly to the data acquisition system. The readings are used in their raw form both in terms of the measurement value and units. Within the software there is the artificial intelligence (AI), the knowledge base, third-party software for such things as the graphical user interface, installation and simulation software. Within the knowledge base are the general and specific generator information on problems and indicators of problems, which the AI software must process. The knowledge base is generally a refined database that attempts to incorporate OEM specific designs and emerging or known issues as well as industry wide commonality in generator operation and troubleshooting. The knowledge base consists of the possible generator and auxiliary system problems. Attached to the problem network or table are the indicators that consist of as many sensors and problem indicators as available from the installed generator instrumentation. The problem indicators are, in effect, the sensor inputs or combinations of sensor inputs that convey the information to the problem set that some

5.4 ARTIFICIAL INTELLIGENCE (AI) DIAGNOSTIC SYSTEMS

249

operating parameter or limit has been exceeded, and that a real problem is occurring. The AI software determines this, as it processes the information from deviations between actual sensor readings and expected readings. The expert system software uses the sensor inputs to look for deviations in readings that indicate generator problems and reports the relevant information to the monitor for operator interaction. Dynamic monitoring lends itself to the expert system application extremely well and takes even greater advantage of this technique. For example, some stator winding problems are not simply related to only one stator bar/coil, but to a particular parallel, phase, or the whole winding. We would, therefore, expect to see all the bar/coils associated with the overheated parallel in high temperature alarm, and the other bar/coils to go down in temperature since no stator current is flowing. This event, however rare, illustrates the power of the expert system as there are countless other scenarios that could occur during operation. Let us say, for the same winding configuration, in a water cooled stator winding, that one phase loses all cooling water flow due to plugging of the coolant path, which is sometimes made possible by the configuration of the water delivery system to the winding. We would then expect to see all the bar/coils associated with this phase in high temperature alarm since the phase is still carrying current but is not being cooled. Finally, consider the whole winding in the case in which the surface air cooling water flow is greatly reduced but is still flowing. The temperature monitoring indicators for the stator bar/coils will see normal stator current and inlet water temperature, but reduced water flow. Therefore, all the surface air cooler air outlet temperatures should be in alarm since all will be reading higher than the calculated expected value and the stator winding temperature will be trending upward at this point. The point of the last three examples is that the root problem is not related to the bar/coils themselves. Therefore, a further method is required to establish that the problem is not with the stator bar/coils but is rather rooted in the water connectors, brazed connection on the circuit ring, or the cooling water delivery system. To do this, we can use the stator winding diagram to form a sensor network and map out which bar/coils belong to each of the three phases, and, subsequently, which of these are in each parallel [2]. Using the stator bar/coil temperature models, the expert system can then reason that the problem is related to, for example, the red phase only because the winding mapping tells it that only the bar/coils in the red phase are overheating. Therefore, the expert system would report simply that the red phase is overheating, rather than all the stator bar/coils in the machine. One can then use graphical abilities of the computer to track the temperature of the affected phase in comparison to the other phases and to load changes. It should now be clear that the more monitoring points that are available, the more accurate and comprehensive the expert monitoring system can be. This was discussed previously in this chapter when ordering a machine, spend the extra money and outfit the generator with sufficient monitoring points.

250

CHAPTER 5

MONITORING AND DIAGNOSTICS

The degree of overheating can be determined by further using the individual bar/coil temperature models to give the average temperature of the bar/coils in the affected phase, or the hottest bar/coil in the affected phase, and so on. Handling of the temperature reporting is discretionary and is simply a matter of choice in this case. Hence, there is additional flexibility built in with this type of approach.

5.5 MONITORED PARAMETERS All generator monitoring systems require sensors connected to the generator and its auxiliary systems to provide a diagnosis of problems that may be occurring. These sensors are various types of instrumentation that are installed to measure some parameter that is important to the safe operation of the machine. A sensor can come in many forms such as a thermocouple, pressure gauge, level switch, accelerometer, proximity probe, and so forth. In this section, we discuss the basic parameters monitored in generators and their auxiliary systems, and the effect of operation outside the limits of the parameter. A brief description of the types of sensors or instruments used for each parameter is included, as well as what each detects and how they are used. In addition to the basic individual parameter or sensor information, some sensors can be related to another sensor to form an indication of a problem or existing generator condition. Here we describe how these sensors can be used in conjunction with others, to provide some additional diagnosis of the condition of the generator. In discussions of generators in general, the machine is broken down by components and subsystems that have common elements. The sensor information in this chapter will be organized by subsystems of the generator.

5.5.1

Generator Electrical Parameters

5.5.1.1 Generator Output Power Generator output power is the real MW power output from the generator. It is a function of the stator terminal voltage, current, and the generator power factor as shown in Equation (5.3): MW = MVA × PF = √3 × V L × I L × PF

(5.3)

where, VL is the line voltage and IL is the line current. The power outputs, both active power (MW) and reactive power (MVAR), are monitored by voltage and current signals taken from the generator potential and current transformers. The signals are processed to provide the MW and MVAR

5.5 MONITORED PARAMETERS

251

information and displayed and recorded in the main control room, to keep track of the load point of the machine and allow operator control of the generator. MW overload on the generator is the main concern, and often this means that the stator current limit has been exceeded and worse, that the mechanical limit of the drive train may have been exceeded. High stator current will affect the condition of the stator winding from a thermal aging point of view. Further the mechanical stress may be beyond design limits for the drive train. Excessive stator terminal voltage can also result in an overload condition and stresses the groundwall insulation of the stator winding, but this is generally alarmed and relay protected to ensure that an overvoltage situation does not occur. 5.5.1.2 Generator Reactive Power Generator reactive power is the megavolt-amperes reactive (MVAR) output from the generator. As with the MW of the generator, the MVARs are also a function of the stator terminal voltage and current. However, the MVARs are principally determined by the field current input to the rotor. Therefore, when MW is held constant, varying the field current will change the MVAR. The relationship is as shown in Equation (5.4): MVAR =

MVA2 − MW2

(5.4)

The MVAR loading of the machine must also be monitored, since it also has operating restrictions as shown in Chapter 4 with the capability curves. Exceeding the maximum MVAR loading means that the field current limit on the rotor has likely been exceeded in the lagging power factor range which will cause the rotor winding to overheat, if the rotor is the limiting factor on that part of the capability curve. In addition, the stator terminal voltage can also be exceeded during excessive MVAR loading and cause stator core over fluxing and, hence, the volts per hertz (V/Hz) curve as discussed in Chapter 4 then applies. From this discussion, it can easily be seen that it is very important to heed the capability diagram at all times during operation. Exceeding the minimum MVAR loading means that the field current on the rotor has been reduced to a very low level, such that the generator is operating near the bottom of the leading power factor range (this has also been discussed in Chapter 4). When the MVARs are reduced beyond design limits, the possible problems that can occur are exceeding the minimum terminal voltage limit thus affecting the static excitation system, and loss of stability from slipped poles if the reluctance torque of the rotor is exceeded. 5.5.1.3 Three Stator Phase Currents The three stator phase currents are monitored via the generator current transformers. Different classes of current transformers may be employed for metering and relaying requirements. They are located close to the generator winding on the main output terminals (bus or bar). The analog output from the current transformers, approximately 5 A at rated generator output current, may be displayed and/or recorded on the main control panel or computer display.

252

CHAPTER 5

MONITORING AND DIAGNOSTICS

The three phase currents flowing in the stator winding produce I2R losses, which directly affect the temperature of the winding. Excessive current will cause temperature increases proportional to the square of the current. In addition, vibration and bar/coil bouncing forces are induced in the stator windings in proportion to the square of the current flowing. Temperature and vibration affect the electrical and mechanical integrity of both the strand, turn (where applicable), and groundwall insulations, and the stator bar/coil semiconductive surface coatings in the form of tape or paint. The mechanical bond integrity between the copper strands and the groundwall insulation is also affected by temperature and thermal cycling due to the shear forces that exist. The stator currents are monitored and used to provide indication of an overload condition and phase current unbalance. The unbalanced currents in the stator winding can be used to calculate the negative-sequence currents flowing in the rotor amortisseur. 5.5.1.4 Stator Terminal Voltage The stator voltage is monitored by the generator potential transformers located at the main output terminals of the stator winding. Different classes of potential transformers may be employed for metering and relaying requirements. The stator voltage is generally displayed and or recorded in the main control room or computer display. Terminal voltage is a function of magnetic flux, rotor speed, and the stator winding configuration. Excessively high voltage on the stator winding can prematurely age the groundwall insulation and deteriorate the stator bar/coil surface coatings due to electrical breakdown and subsequent degradation. This degradation is normally seen at the semiconducting and grading tape/paint interface in the form of a thin white line encompassing some or all of the circumference of the coil or bar. This degradation, depending on how severe or how long it has been occurring, can erode into the mica layers of the stator winding. Once again, if the voltage becomes too high, the V/Hz curves apply due to the over fluxing condition. The stator terminal voltage is monitored to look for anomalies, either too high or too low, and to monitor the degree of phase voltage unbalance. Monitoring of the generator terminal voltage is critical during synchronizing of the generator to the system. The terminal voltage of the generator must be matched in magnitude, phase, and frequency to that of the system voltage before closing the main generator breakers. This is to ensure smooth connection to the system, with no malsynchronization occurring. 5.5.1.5 Field Current The field current measurement is brought back to the control room for monitoring and can also be used with the field voltage to calculate the rotor winding average temperature. In the case where the temperature is being calculated, an extremely accurate rotor winding resistance is required. This value is normally found in the original commissioning data or can be measured when the machine is out of service at some specific temperature. The direct current flowing in the rotor

5.5 MONITORED PARAMETERS

253

winding produces I2R losses, which directly affect the temperature of the copper winding. The temperature of the rotor winding is measured by using the measured field current and the measured field voltage to calculate the resistance of the rotor winding. It is important to keep in mind that the voltage and current readings need to be calibrated and checked for accuracy if a meaningful temperature is desired. Design data should not be used as the calculated value and the measured value are never identical. Excessive field current will cause the field winding to overheat and result in high terminal voltages. A field current that is too low can cause decrease of the terminal voltage to below its minimum allowable value and once again the V/Hz curve applies. In addition, if the field current is reduced too far, the static exciter system could shut down due to minimum current for the silicon controlled rectifiers to operate. In this case, the machine would become self-excited as discussed in Chapter 4. If the prime mover is still actively pushing out more MW than the reluctance torque can accommodate, and the operator does not attend to this situation immediately, pole slipping will occur and the unit will become unstable. 5.5.1.6 Field Voltage Field voltage is measured at the excitation system supply to the rotor and is generally taken as the voltage across the sliprings. The measurement is brought back to the control room for monitoring and is also used with the field current to calculate the rotor winding average temperature. Once again, this voltage must be calibrated and checked to ensure accuracy. One option is to use an unloaded brush to pick up the voltage, i.e. one that is not carrying any current and is specifically for this reason. Otherwise, the voltage drop across the brush must be taken into account somewhere in the circuit. Increasing the field voltage increases the field current proportional to the rotor winding resistance. It is used to calculate rotor winding resistance and, subsequently, the rotor winding average temperature. AVR problems can cause the field voltage to become too high and, in turn, cause the excitation to increase beyond design limits. Field voltage may be increased in times of system events to boost generator voltage by “field forcing.” This term is used to describe increasing the field voltage to approximately double its normal value, on a short-term basis. The capability of any rotor during field forcing depends on the field winding design such as cooling efficiency, how much overcapacity was designed into the field winding to begin with, how close to that over capacity is the machine operating at just before field forcing begins, and the insulating materials used. 5.5.1.7 Frequency The frequency of the generator output voltage is monitored at the generator potential transformers. The frequency is usually displayed on the main control panel. Frequency is measured in cycles per second or Hz and refers to the electrical frequency of the generator. It is monitored for abnormal deviation from the

254

CHAPTER 5

MONITORING AND DIAGNOSTICS

system frequency, which is 60 or 50 Hz, depending on location in the world. Over frequency is most often the result of an instantaneous load rejection. Under frequency is generally caused by a system event rather than the generator itself. The effect on the generator, however, is almost always an attempt by the system to extract excessive current from the stator and to drag the rotor speed down. This also has the effect of depressing the stator terminal voltage. To offset this, the excitation system for the generator will generally go into “field forcing” to try to maintain rated terminal voltage. Therefore, there is a possibility of sustaining overheating in both the stator and rotor windings during this type of event. Protection against under frequency events should be provided. 5.5.1.8 Volts per Hertz V/Hz is the ratio of terminal voltage to generator electrical frequency. It is put in place to protect the generator from over fluxing during open circuit operation and while on load. The V/Hz curve as issued by the manufacturer should be used to set the protections. 5.5.1.9 Negative Sequence Negative sequence generally refers to negative sequence currents in the stator winding which then induce currents in the rotor amortisseur circuit. These induced currents that flow in the amortisseur winding occur at twice rated frequency component. The unbalanced condition may be due to a machine or system problem but is more often system related rather than due to a problem in the generator itself. There is always a small natural degree of imbalance in these three-phase currents, but they are not harmful below the continuous I2 value. There are two components of negative sequence to consider. The first as just mentioned is the continuous I2 component, which refers to the amount of phase unbalance the generator can tolerate for an infinite operating period. The second is the transient component called I 22 t which refers to the degree of short-term phase unbalance that the generator can withstand. For large hydro generators, a typical continuous I2 value of 5 for nonconnected amortisseurs and 10 for connected amortisseurs would be normal. This means that the generator could carry a continuous phase imbalance in the stator winding of 5% or 0.05 pu and 10% or 0.010 pu of the rated stator current without damaging any of the rotor components. During faults, the I 22 t value in the stator winding should not exceed 40 according to [3]. Values that are between 40 and 80 suggest the generator may have sustained varying degrees of damage and a full unit inspection is highly recommended. Serious damage should be expected if the I 22 t value is over 80. When the degree of imbalance becomes significant, it appears as 120 (100) Hz currents flowing in the amortisseur circuit which can overheat the amortisseur bars themselves as well as the rotor pole-face. Relay protection should be provided to detect the level of negative-sequence currents and initiate a generator trip.

5.5 MONITORED PARAMETERS

5.5.2

255

Stator Core and Frame

5.5.2.1 Core Temperatures Stator core temperatures are monitored by thermocouples (TC) or resistance temperature devices (RTDs) embedded between the stator core laminations at strategic locations (see Figure 5.5-1). In the radial direction, these locations are usually in the core yoke and their exact location will vary from manufacturer to manufacturer. In the axial direction, they are positioned to give the best coverage from top to bottom as possible depending on how many are installed. The coverage provided will be such that the various core heating modes may be fully monitored. These include global heating, core-end heating, and localized heating. Global overheating of the stator core may occur from the core being loose and the subsequent vibration deteriorates the inter-laminar insulation. Other causes of global core overheating are low flow or loss of flow entirely of the cooling water, ventilation issues (original design or re-engineered), and restricted ventilation from plugged core air ducts. Core-end overheating from fringing flux is a condition that is not common in hydro generators since core-end heating may not reach objectionable levels when operating under excited since only limited amounts of fringing flux is typically produced. The heating that could occur would definitely affect the end of the core laminations and clamping assemblies. Depending on how hot these areas become

Figure 5.5-1 Embedded core thermocouple on lamination. Note that this figure is an example of where the thermocouple is located, the lamination or laminations above and below would be notched accordingly to accommodate the wire.

Volts/Hz (Both in per unit of rated values)

256

CHAPTER 5

MONITORING AND DIAGNOSTICS

1.3

25% Limit for short time based on high core flux

1.2

Decreasing current capability for longer times based on limiting additional temperature rise

1.1

1 1

10

100

1000

10 000

100 000

Time (seconds)

Figure 5.5-2 V/Hz curve.

would then cause concern for the insulation system of the core and the stator winding. This core-end overheating described here is not a hotter core-end due to poor ventilation patterns in the machine, this is a separate issue. A certain amount of over fluxing may be tolerated for a short period of time from a few seconds to tens of seconds as shown in Figure 5.5-2. This is related to a number of factors such as the electric loading of the machine, size of airgap, end core pack design, stray flux, and magnetic saturation of the core, which are all affected to some degree by leading power factor operation. The amount of effect in the leading power factor range, for any given machine, will depend on the core-end design in terms of electric and magnetic loading characteristics. In this overheating mode, both ends of the core are affected while center is not affected. During leading power factor operation, the interaction of the magnetic fields in the core-ends is such that there is a higher degree of axial flux impingement on the core, which tends to enter the iron at the core-ends as a fringing flux. As the power factor or field current is reduced, fringing increases and, subsequently, the core-end temperatures rise. The level of temperature rise in the core-ends is dependent on a variety of generator design features, thus, some machines will see this affect more than others. Having said this, the core clamping assemblies will also see a temperature rise due to the higher degree of flux fringing on the core-ends. It is once again important to remember that if the machine is operated within the capability curve issued by the manufacturer this core-end heating will not be an issue. This point cannot be stressed enough. Local core overheating is a condition generally related to a small area of the core and is usually due to a localized defect in the stator laminations. These defects can stem from foreign material that worked its way into the airgap, hammer strikes during re-wedging operation, loose vent spacers, broken welds on the core-end

5.5 MONITORED PARAMETERS

257

fingers causing pressure loss as the finger vibrates, poorly manufactured core finger plates where the fingers are not level and cannot provide uniform pressure on the core laminations, or a wave in the core that has loosened the local core area and has degraded the inter-laminar insulation. This breakdown of the inter-laminar insulation or shorting of the core laminations causes eddy currents to flow and thus the heating. It is important to recognize that some of these mechanisms may take decades to develop, such as the broken welds on the fingers or the loose vent spacers, thus the need for visual inspections on an ongoing basis. 5.5.2.2 Core Vibration Vibration in the stator core is naturally produced by the action of passing poles at synchronous speed pulling and releasing the core at twice rated operating frequency. The magnetic field itself also influences the shape of the core known as a magnetostriction effect, which produce the typical transformer type noise which is more objectionable if the stator core is loose. The pole or direct axis carries the main flux, while the winding or quadrature axis carries only the leakage and stray fluxes. Therefore, a large difference in magnetic force is inherent between the two axis. A large magnetic force is generated in the pole axis, and a weak magnetic force is present in the winding or quadrature axis. Since each pole has a north and a south associated with it (with a quadrature axis in between each) and there is rotation, the stronger and weaker magnetic pulls generate vibration at twice the line frequency. As mentioned many times in this book, the core must be maintained tight or fretting will occur between the laminations. Minor fretting will tend to deteriorate the inter laminar insulation, but if the core becomes too loose, the laminations and or the space blocks may even fatigue, with the result being pieces of loose core material breaking off and causing damage. Another very important variable for core vibration is the core splits if the machine was not continuously piled. The core splits are originally assembled with a packing paper or otherwise popularly known as “fish” paper. This packing would typically have been in the order of 0.030–0.050 , for example and the purpose was twofold. The reason is to connect the core sections together so the core behaves as one continuous solid ring, not separate pieces when excited by the flux. A properly packed core will not allow axial vibration of the laminations at the split. As discussed, vibration will cause fretting of the laminations as well as shorting in this local area. When inspecting the machine, pay particular attention to the split area to ensure no fretting or greasing is taking place due to relative movement of the coreends. Relative movement of the core-ends deteriorates the packing material over time and is an indication that the packing paper is no longer performing its intended function which is to keep the two sections of core in mechanical contact. Consideration should be given to reinstall the packing paper at the splits provided the core to frame attachments have not deteriorated. Depending on the degree of deterioration of core to frame attachment, repacking the split may not provide a long-term repair. If this is the case, core replacement should be considered.

258

CHAPTER 5

MONITORING AND DIAGNOSTICS

A sure sign the split-packing is no longer performing its intended function is twice the line frequency component that is dominating the noise/vibration spectrum. This noise can be quite noticeable as a loud and constant hum while the machine is excited. Experience has also shown that machines can hum for some time while they warm up and then this humming noise subsides. This is an indication that the splitpacking is no longer performing its intended function since when the core heats up the split sections come together enough to put the resonant frequency off the twice the line frequency. Consideration for repacking is also suggested in these situations. It is also likely fretting dust will be present at the splits, so a visual inspection is highly recommended. The split-packing is like any other component it does not last forever and should be given some attention particularly with machines that are over 50 years old. Monitoring of core vibrations can be done with special accelerometers mounted on the back of core in multiple locations to determine the magnitude and phase of radial vibration, see Figure 5.5-3. The suggested radial vibration limits for curve D (38 μm p-p for 100 Hz and 35 μm p-p for 120 Hz) as shown in Figure 5.5-4 is from a set of 700 MVA machines built by a major OEM but can be used for most sizes of machines [4]. These limits were agreed as part of the performance guarantees for 20 generators at a station in Brazil [4]. The suggested radial vibration limits, as mentioned above, are not recommended for the areas near or at core splits. For these areas, it is suggested that curve C (21 μm p-p for 100 Hz and 18 μm p-p for 120 Hz) as shown in Figure 5.5-4 should be used.

Figure 5.5-3 Core accelerometer placed on back of core between keybar and core bolt.

5.5 MONITORED PARAMETERS

259

1000 600 400

Vibration amplitude – peak to peak (μm)

D 200 C 100 80 60

B A

40

20

10 5

2 1

1

2

3 4

6 9 10

20

40 60 100 120

Vibration frequency (Hz)

Figure 5.5-4 Limits for vibrational displacements [4].

5.5.2.3 Frame Vibration Frame vibration is also excited by the same magnetic pull influences on the stator core and, consequently, the resulting vibration produced in the frame. There are known cases of vibration resonance occurring on the frame as a result of the frame having a resonant frequency near to twice the line frequency. If the machine is new or has had this issue from the original day, it may be necessary to stiffen the frame, change the stator winding group connection pattern or the number of stator slots. The generator core and frame are typically designed to be an interference fit while in operation, thus, the frame will tend to be locked onto the core and form part of the resonant system. This may not be the case when the generator is cold at initial startup as previously discussed. Some machines may be designed so that the core and frame do not to lock together radially when in service. Severe damage to the frame can occur by initiating cracks in the frame welds or in the frame members themselves particularly the keybar assemblies that secure the core to the frame. Residual damage from the high vibrations associated with

260

CHAPTER 5

MONITORING AND DIAGNOSTICS

frame vibration is likely to be transmitted to other components of the generator if the situation becomes severe. Good core to frame coupling as previously mentioned is required to ensure that the core and the frame move together if designed to do so. There is evidence of numerous cases in which core frames became “uncoupled” from the core and impacting damage is found at the core-to-keybar interface. The stator frame has the same suggested vibration limits as shown in Figure 5.5-4.

5.5.3

Stator Winding

The stator winding is a high cost component of the generator. Most serious stator problems are statistically found in the stator winding, due to the nature of the construction of the component and the operating duty it must endure with the relatively “soft” materials employed, when compared to the stator core or frame. To judge the condition of the machine, operating load conditions can normally be correlated to temperature, vibration, partial discharges [5], and other parameters being monitored. Temperature is the main parameter monitored, and the winding may be monitored for temperature using embedded detectors between top and bottom bars or front and back legs of the winding in the slot. The endwinding may be monitored for radial, axial, and tangential vibration at multiples of the operating frequencies, and phase circuit rings and terminal bushings may be monitored for both temperature and vibration. Correlations normally are made of identical quantities at fixed operating conditions, held constant for an adequate time period to establish steady-state conditions. However, some occurrences relate to changing conditions and in these cases, an exact record and methodical variation of conditions may be required for a proper diagnosis. 5.5.3.1 Conductor Bar/Coil Slot Temperatures Conductor bar/coil slot temperatures are monitored by a thermocouple or RTD embedded between the top and bottom bars/coils in the slot. Well-instrumented generator stator windings employ an RTD or thermocouple embedded in the conductive middle slot strip or spacer as shown in Figure 5.5-5. The more temperature sensors in the machine the better known the temperature profile will be. It is well worth the investment when putting the winding in to install as many RTD or TCs as practical per phase. A minimum of 48 sensors is recommended as this will give a good cross section of coverage. These sensors can be distributed evenly along the top, middle, and bottom of the winding, for example. During commissioning, the hottest sensors are monitored and remain connected to the data acquisition system if the data acquisition system cannot handle all of the sensors. The remaining sensors are terminated in a terminal box and remain as spares to be checked later on in the life of the winding to determine if they are now part of the hottest group. Today’s acquisition systems can monitor a good portion of the minimum required with little extra cost. It is important to remember that these sensors can fail over time so when installing the minimum recommended here, 30 years

5.5 MONITORED PARAMETERS

261

Slot wedge

Under wedge ripple spring Conductive slot side spring

Laminated stator core

Copper strands (20 per turn)

RTD three wire lead

Turn to turn insulation Ground insulation

Conductive bottom slot strip Conductive middle slot strip Integrally moulded corona conductive layer

Figure 5.5-5 Illustration of a slot RTD installed in separator pad and located in the slot between the top and bottom stator bar. Source: Courtesy of Dr. Michael Znidarich.

later it could be that only 70% remain operational. There is nothing worse than a machine that had only eight sensors installed and only four are operational now and there has been a winding failure and coils need to be cut out. Based on real-world experience, monitoring the temperature now becomes an issue. Slot temperature monitoring allows detection of overload and reduction/loss of cooling, so trending of these temperatures is highly recommended as it will be useful one day. 5.5.3.2 Stator Winding Differential Temperature Stator winding differential temperature refers to the monitoring of the hottest to coldest operational stator winding temperatures. It further refers to a condition of temperature imbalance between individual stator bar/coils or phases. The condition may be caused by phase imbalance due to a system problem, localized bar/ coil-to-bar/coil temperature differences, plugging, other flow restrictions, and high resistance or broken electrical joints. An attempt is generally made to keep the temperature differential to the minimum possible, but it is not uncommon for a machine to have an inherent 10 C difference from hottest to coldest bar/coil temperature. This can be due to differences in the cooling circuit from such things as

262

CHAPTER 5

MONITORING AND DIAGNOSTICS

some bar/coils having the phase connectors in series with the stator bar/coil cooling circuit. Slot RTDs and/or stator cooling water outlet temperature sensors indicate the condition. Stator winding differential temperature can also refer to the difference in inlet-to outlet temperature of the stator winding cooling water. This should also be within a characteristic range for any particular machine. The differential temperature across the water cooled stator winding, from the stator cooling water inlet to the outlet of the generator, can be monitored to ascertain that the design temperature difference across the whole stator winding is at the correct level. Correct differential temperature is an indication that the stator cooling water flow rate is adequate for cooling of the stator winding as a whole. This type of monitoring is not indicative of problems in individual stator conductor bar/coils. When the temperature differential is higher than normal, or higher than recommended by the manufacturer, this may indicate a partial blockage somewhere in the stator winding or stator cooling water system. To determine the source of partial blockage, other testing and monitoring is required to identify the plugged location. 5.5.3.3 Stator Surface Air Cooling Water Inlet and Outlet Temperature and Water Flow This raw water inlet and outlet temperature is monitored to ensure the generator surface air coolers are operating efficiently. The amount of raw cooling water into the cooler is also monitored. These temperatures and water flow will give an indication if cooler tubes are blocked internally, or there is restricted airflow through the cooler. The original cooler manufacturer curves should be consulted as a base line for the cooler as designed. 5.5.3.4 Stator Surface Air Cooler: Hot and Cold Air Temperatures The generator hot air (air going into the surface air cooler) and generator cold air (air exiting the surface air cooler) are also monitored to ensure the surface air coolers are operating efficiently. 5.5.3.5 Stator Winding Differential Pressure There should be a normal differential pressure that exists across the water cooled stator winding, during operation, based on design factors. A higher than normal differential cooling water pressure across the stator winding can indicate that there is a cooling water flow problem. This may be due to some form of plugging in the stator cooling water system. The differential pressure across the stator winding from inlet water manifold to outlet water manifold is usually monitored to ensure proper cooling water flow, and maintain the correct operating temperature in the stator winding. If the pressure differential is much higher than expected, a large general obstruction to the flow of the stator cooling water may be present. If the pressure drop is very low across the generator, this may indicate a different problem such as a pump problem, blockage

5.5 MONITORED PARAMETERS

263

in one of the external system components, or a large leak of stator cooling water out of the external system piping before the generator inlet. 5.5.3.6 Stator Endwinding Vibration Monitoring Endwinding vibration can damage the mechanical integrity of the stator conductor bar/coils and the other stator winding components that make up the electrical and cooling water delivery portions of the total stator winding. Endwinding vibration is usually a symptom of loose endwinding support structures. Checking and retightening of these support structures is required to avoid damage to the stator winding and associated hardware. Vibration monitoring may be employed to indicate increasing vibration levels and plan outages to make repairs. Damage from vibration can initiate leak problems from cracked strands, initiate cracks in the brazed joints, fretting of the endwinding support system, and loosening of fittings. There are many different endwinding support system designs in operation. Some are designed with the intent that they do not need retightening, but loosening may still occur naturally over time due to vibration from the forces induced by the high AC electromagnetic fields inside the machine. Other designs are provided with mechanisms for periodic retightening, and allow for natural loosening. Again, the degree of natural loosening and the time it takes depend on the design of the endwinding support system and the dampening effect of the support system. Premature loosening may be enhanced by oil ingress, which will also create a greasing effect which is a tell-tale sign things are loose. A dark brown or black (depending on the oil color) grease-like substance will be seen at ties and blocking interfaces where fretting has occurred due to the oil ingress. If the machine is clean, it will be a whitish powder where the fretting is occurring. System faults will again affect the integrity of the endwinding support system due to the high forces induced during such events. In some instances, vibration transducers are installed to monitor the online vibration characteristics of the stator endwinding. Vibration transducers are becoming more popular as a method of diagnostics for machines in service. Figure 5.5-6 shows such a transducer on the endwinding of a 1.95 m (76.7 ) high by 4.35 m (171.2 ) diameter, 276 slot machine. Measurements are generally done in the tangential and radial modes. The circuit rings that are located outside of the winding are in fact an extension of the stator endwinding in the sense that in almost all hydro machines, all connections from the winding jump onto a circuit ring to connect to the main and neutral buses. In machines that are free of oil, looseness caused by vibration may be indicated by paint cracks where ties and blocking interface with the winding. The circuit rings may also have paint cracks at the blocking interface holding the rings in place or a white powder at this interface may be present if vibration is more severe. In these cases, there may be some white or light gray powder which indicates some looseness not to be confused with corona discharge. The white or gray powder is the micro fine abrasion of the insulation system (the epoxy and tape) against the object it is rubbing against. A good close inspection will reveal that the white or gray power is in fact worn insulation. Repair may entail carefully

264

CHAPTER 5

MONITORING AND DIAGNOSTICS

Figure 5.5-6 Fiber-optic stator endwinding vibration transducer. Source: Courtesy of Qualitrol-Iris Power.

cutting the ties that support the winding and blocks or loosening the block assembly if secured with a bolting arrangement. It will be very evident once the tying or blocking is removed if fretting of the insulation is occurring. Repair of this eroded insulation will require re-applying the insulation that has been removed. This may be accomplished by applying a half lap layer of mica and epoxy to the affected area or inserting felt to fill the gap that has formed if just superficial damage has been done. If in doubt about how critical the damage is, consult with the OEM. Reinstallation of the tying or blocking is very simple and straightforward and one should not worry about removing it to begin with in order to get to the root of the damaged area. Figure 5.5-7 shows a flat bar circuit ring arrangement that had circuit ring vibrating while in service and the resulting minor damage that occurred. For large diameter machines, abrasion of circuit rings may be reduced or eliminated by providing features in the support blocks that allow the rings to expand and contract more freely, contact the OEM to see if this solution is applicable. 5.5.3.7 Stator Winding Ground Alarm/Trip Stator winding grounds occur upon failure of the groundwall insulation on the coil or bar. In the majority of cases, this occurs in the stator slots next to the stator iron. In addition to the stator winding conductor bar/coils, the phase connectors and terminals are covered by this protection. In fact, this is more a protection scheme than a monitoring parameter. It is generally used to trip the generator on such an occurrence. A detailed description of this topic is covered fully in Chapter 6. A ground alarm does not necessarily mean that a winding insulation failure has occurred. It can be caused by instantaneous overvoltage from either machine

5.5 MONITORED PARAMETERS

265

Figure 5.5-7 Loose circuit ring vibrating and the white powder produced against the support block.

transients or system spikes, and it only indicates that an event has occurred. Some basic electrical testing as well as a thorough visual inspection is usually carried out to prove that the ground insulation is still viable. The generator may then be returned to service if the tests show that no failure has occurred. 5.5.3.8 Stator Winding Partial Discharge Monitoring Partial discharge monitoring is a very popular diagnostic tool used in monitoring and trending the stator winding for partial discharge activity while the unit is in service. The location of the partial discharge couplers on one phase bus of the stator winding is shown in Figure 5.5-8 and the instrument is shown in Figure 5.5-9. The coaxial cables from the partial discharge couplers are terminated inside an electrical box. The partial discharge analyzer as shown in Figure 5.5-9 is connected to these terminations and processes the signals. Partial discharge (PD) in generators is mostly associated with the highvoltage stator conductor bar/coils and associated circuit rings and main output bus. Partial discharge is the generic term for small electric current pulses flowing from one surface of the generator stator winding to another, or to another component of the generator, usually the stator core. These are not necessarily harmful and are quite normal. Partial discharges are also constantly occurring during machine operation. Partial discharge generally occurs across the small airgaps or voids within the main stator winding insulation, and this is referred to as an “internal” type of discharge. Partial discharge can also take place on the outer surface of the stator bar/coil insulation protruding from the end of the stator core slot into the endwinding region, and this is generally referred to as “endwinding” discharge.

266

CHAPTER 5

MONITORING AND DIAGNOSTICS

B2 Incoming phase bus B1

L2 L1 + B1 = L2 + B2 (All measurements in ns)

L1

Coaxial cables

Figure 5.5-8 Shows typical location of the partial discharge couplers on the line end of the winding. Source: Courtesy of Qualitrol-Iris Power.

Figure 5.5-9 The Iris Power PDA-IVRP instrument. Source: Courtesy of Qualitrol-Iris Power.

Similarly, this can occur on the circuit rings and main output bus given the ideal conditions of insufficient spacing and/or contamination. The above terms are general in nature and can be broken down even further to describe the different types of discharge within these designations. Partial discharge that takes place between the surface of the insulation in the stator slot and the slot wall and/or the semiconducting coating (corona-protection layer) is generally referred to as slot discharge. This type of discharge can be very destructive to the stator winding groundwall insulation and is highly dependent on design, manufacturing and assembly issues, in order that the insulation system will maintain its integrity for reliability and long service life. The copper conductors within the stator bar/coils must be completely electrically insulated from the stator

5.5 MONITORED PARAMETERS

267

core to prevent destructive discharge currents flowing to ground. Therefore, the bar/coils should be installed in the slots such that each bar/coil cannot move in its slot and maintains good and continuous electrical contact with all grounded surfaces of the slot. On a high-voltage stator bar/coil, if there is a gap of a critical size between the stator bar/coil and the slot wall, a charge will build up and eventually discharge across the gap to the stator core. This is extremely damaging to the groundwall insulation of the bar/coil. To avoid such gaps, good surface contact between the stator bar/coils and the stator core must be maintained, and a semiconducting coating is used for corona or discharge suppression. It is usually made of graphite or carbon-loaded varnish, and applied in paint form or within the stator bar/coil armor tape. Also, the semiconducting or corona protection layer must be of the proper resistance value (typically 500–15 000 ohms per square) and applied correctly. If done correctly, this semiconducting layer will establish the necessary low resistance contact between the surface of the stator bar/coil and the core. Restricting bar/coil movement in the slot and proper application of the semiconducting material is critical for good contact and minimizing any gaps that may occur. When discharge does occur, it may be recognized by the typical white residue produced. In the capacitive type of slot discharge, as the semiconductive coating is eroded away, large amounts of ozone are produced and the bar/coil surface resistance increases and the effect simply multiplies. Eventually, the groundwall insulation is electrically pitted and itself erodes away until there is not enough groundwall insulation left to hold the stator voltage, and a ground fault will occur, see Figure 5.5-10. This type of failure mechanism is highly unlikely, but there have been documented cases in private investigations. This semiconducting layer is extremely important in controlling the capacitive effects of slot discharge by ensuring that the resistance between the bar/coil and the stator core is low enough not to allow voltage build up that will promote discharge. If the semiconducting layer is removed, for whatever reason, the resistance between the bar/coil and core surface can get too high and capacitive effects will overtake. Sometimes, however, there is slot discharge that originates internally in the stator bar/coil groundwall insulation by delamination of the insulation material or by separation of the groundwall from the copper conductors. Even though the bar/

PD - slot discharge

Stator slot wall Gap

+++++

Charge on insulation surface

Semiconducting corona protection Groundwall insulation Copper conductors

Figure 5.5-10 Slot discharge from loss of the semiconducting coating or corona protection layer, capacitive type discharge. Source: Courtesy of Alstom Power Inc.

268

CHAPTER 5

MONITORING AND DIAGNOSTICS

Stator bar

Slot discharges Voids and delaminations in groudwall Delamination copper surface

Iron core

Figure 5.5-11 Internal and slot discharges. Source: Courtesy of Alstom Power Inc.

coils may be tight in the slots and the semiconducting coating is intact, voltage breakdown occurs inside the bar/coil insulation and damaging discharge currents result. Therefore, capacitive slot discharge can be broken down even further to surface discharge, internal voids and de-lamination, and groundwall separation from the copper, (see Figure 5.5-11), and each of these shows up as a different characteristic of the partial discharge measurements. This will be discussed a little further on in this section under online measurement techniques. 5.5.3.9 Vibration Sparking There is another mechanism in the slot section of the stator winding that causes erosion of both the semiconductive layer of the bar or coil and the stator core itself. This is due to loose bars or coils in the slot and the semiconductive coating having too low a resistance. It is called vibration sparking and the most troubling attribute of this mechanism is that it can occur in any slot in the machine, independent of the voltage on the bar or coil. In the worst cases, with a loose winding, this mechanism can be widespread all over the machine. Figure 5.5-12 shows parasitic currents that would flow based on a sample core geometry. Currents can flow in both top, and bottom bars or coils, however, for clarity of illustration, only the bottom bar or coil leg is shown as carrying the current in Figures 5.5-12 and 5.5-13. The bigger outer current loop goes from the keybar and around the top and of the core, then via the semiconductive coating of the stator bar goes to other depending on the contact of the bar to the core. As can be seen, there are big current loops and smaller ones as well. Just as voltage is induced within the copper of the bar, the same voltage is also induced within the corona layer [6]. If the stator bar is touching the core with a large number of contact spots then this will allow the parasitic currents to flow without harm at low densities. If the bar is vibrating in the slot and the contact is lost between the semiconducting layer and the core it is possible for an electrical arc to occur as shown in Figure 5.5-13. This electrical arcing is quite destructive and can fail the groundwall insulation in a very short period of time depending on the amount of movement the bar has in the slot. If there is sufficient clearance in a slot to allow significant movement of approximately 0.1 mm

5.5 MONITORED PARAMETERS

Bottom bar Keybar

Top bar

X X

Current loop

X Yolk flux

X

Stator core

Figure 5.5-12

Shows stator core and parasitic currents.

Electrical arc Bottom bar Keybar

X

Current loop

Top bar Lost contact

X X

Yolk flux

X

Stator core

Figure 5.5-13

Shows the electrical arc as bar is vibrating and contact is lost.

269

270

CHAPTER 5

MONITORING AND DIAGNOSTICS

(0.00393 ) failure may occur very quickly because vibration sparking is much more aggressive than partial discharge [7]. It is important to remember that partial discharge relies on the potential difference across an airgap to exceed the breakdown strength of the air (inception voltage). However, for vibration sparking, there is no inception that occurs because the current flowing already existed prior to the contact being broken between the stator core and the semiconductive coating [6]. Reference [6] determined that the resistance of the semiconductive layer on the bar should be no lower than 5000 ohms per square to prevent this vibration sparking from occurring. There are machines however that are known to have loose bars with lower resistance than discussed that do not suffer from vibration sparking. The minimum resistivity depends on the axial length of the semiconductive coating along the slot section of the bar that is isolated from the core during vibration, which in turn depends on how loose the bar is in the slot [7]. So far, we have discussed slot discharge and vibration sparking in some detail, but there is also endwinding discharge to consider as it can also cause damage to the insulation system. One of the main locations of endwinding partial discharge is at the stresscontrol voltage grading coating at the slot exit. Some of the reasons for endwinding discharge are defects such as electric stress concentrations at the interface between the semiconducting slot coating and the stress control coating, localized mechanical damage on the bar/coil surface, or improper application of the stress control coating. Endwinding discharge may also occur further out in the stator endwinding, past the stator slot exit, due to chemical contamination, loose conductive particles, vibration, mechanical damage, relative movement of endwinding components, and insufficient spacing between conductor bar/coil involutes in the endwinding (see Figures 5.5-14–5.5-16). When the discharge is even further out into the phase connections or stator terminals, those components are also susceptible to the above mechanisms. All of the abovementioned discharge mechanisms in the endwinding require a voltage difference from one location to the other in order for discharge currents to flow along the surface of the bar/coils and across endwinding blockings and ties. Normally, the voltage differences are controlled by the stress grading, groundwall insulation, and adherence to proper clearances and cleanliness. When there are problems with any of the above issues, severe discharge currents can flow and cause insulation burning and eventual failure. One of the worst cases is when there is discharge between phases, since the voltage difference is large and the fault current is severe when a full breakdown occurs. In summary, all the mechanisms that promote partial discharge create areas of voltage stress where electrical charges build up and discharge. The result is possible damage to the voltage grading systems on the bar/coils, the interstrand insulation, or the groundwall insulation. A failure of the stator winding insulation is costly to fix, in terms of both the capital cost to repair or replace a stator bar/coil and the outage time required to

5.5 MONITORED PARAMETERS

271

Main insulation Slot semiconducting corona protection Voltage grading layer

Iron core Damaged endwinding corona protection

Spacer

Endwinding discharges across phases and separations in overhang support system

Figure 5.5-14 Shows the various locations for surface discharges in the endwinding. Source: Courtesy of Alstom Power Inc.

Figure 5.5-15 slot exit.

Damage at the semiconducting/stress control interface just outside the

272

CHAPTER 5

MONITORING AND DIAGNOSTICS

Figure 5.5-16 Endwinding discharges where conductive debris is extending the ground potential past the gradient.

complete the work. Therefore, much effort has been invested over the years in developing techniques to identify the occurrence of PD in the stator winding. Discharge activity in the stator winding can be measured during an overhaul by energizing the winding with an AC test transformer and checking for the presence of the high-frequency currents induced by partial discharge activity. However, although offline testing shows the relative magnitude of partial discharge activity, it is often difficult to identify the cause of any increase in measured levels. Further, offline tests will not readily detect inductive slot discharge since there are no significant electromagnetic or mechanical forces to drive vibration of the bar/ coil in the slot. Offline tests are useful however in detecting capacitive slot discharge where there is damage to the corona-protection coating, and in detecting some forms of endwinding discharge activity. Basically, the methods for offline detection of PD do not cover the operating effects that also promote PD activity, such as those due to thermal and vibration effects. To provide the best PD detection, an online method of monitoring is required. Discharge activity in the stator winding can be measured during normal operation (online monitoring) by detecting the high-frequency currents and/or voltages at the connections to the stator terminals, as well as by other means. However, since the generator is ultimately connected to the power system through an isolated phase bus and by excitation systems, and so on, partial discharge measurements are subject to interference from these types of sources external to the stator winding. Various analytical methods and procedures have been developed to isolate partial discharge arising in the stator winding and to identify the cause(s) of changes in the levels of activity, as well as to overcome interference signals from external sources. These procedures can be successful depending on the design of the stator winding,

5.6 RADIO FREQUENCY MONITORING

273

the magnitude of partial discharge activity, and the relative level of external interference, the method of detection, and software used to interpret the data and filter out the unwanted interference. Great strides have been made in recent years in developed technology that can interpret the wide variety of partial discharge patterns and make a definitive diagnosis of the exact location of the discharge and the root cause. Some types of discharge do produce easily recognizable patterns, but often there are a number of mechanisms in play that cause multiple patterns to be observed simultaneously. This type of situation makes interpretation complicated and not always 100% reliable. Regardless, there are a number of approaches to online PD monitoring, and all are considered viable methods. The following methods are some of those more commonly used on generators today.

5.6 RADIO FREQUENCY MONITORING Radio frequency (RF) monitoring is a technique for detecting electrical sparking and arcing or stator winding PD inside the generator [8]. It operates on the premise that arcing in the stator winding will cause radio-frequency currents to flow in the neutral of the winding. The types of stator winding problems or failure mechanisms that will cause these RF currents to flow are conductor bar/coil strand cracking, electrical joint failure, and partial discharges due to insulation problems. To monitor these currents, a high-frequency current transformer (CT) is placed around the neutral grounding lead, before the neutral grounding transformer, as shown in Figure 5.6-1. The output of the CT is fed to a radio-frequency monitor for signal processing and analysis. The signals from the CT are filtered to examine those that are in the correct frequency range for radio-frequency arcing. The monitor generally has a set-point or an alarm limit that is adjusted to a predetermined level for which the RF activity is known to be excessive. When the level of RF activity increases to where the set point is reached, the operator is notified of an RF problem in the generator.

Generator

Main transformer

System

RF-CT

RF monitor

Neutral grounding transformer

Figure 5.6-1 Radio-frequency monitoring.

274

CHAPTER 5

MONITORING AND DIAGNOSTICS

It can be difficult however, to distinguish between the sources of RF arcing, and it is not always possible to identify the root cause. Sometimes, one can only say that arcing is occurring and at a certain level. In addition, filtering does not completely eliminate noise therefore creating a problem in signal discrimination. This is especially true for noise generated from the sliprings/brush gear and the shaft grounding brushes. At the time of this writing, advances in computer and software analysis have allowed better discrimination between RF signals that are actually PD and those that are noise or from another source.

5.7 CAPACITIVE COUPLING Capacitive coupling has been in use since the early 1950s and was developed as an alternative to RF monitoring [8]. In contrast to RF monitoring, in which detection of PD is at the generator neutral, capacitive coupling is done at the line ends of the generator winding, meaning at the output of each of the phases. The improvement is that PD can be detected on a per phase basis. To measure the PD activity on the generator phases, a tuned capacitor or “capacitive coupler” is connected to each of the phases as close to the generator line end as possible if using a permanent installation (normally done during a rewind). The installation simply consists of a capacitor mounted on the generator frame and connected to the line end of the winding using a shielded cable with the shield grounded to the termination box. See Figure 5.7-1 for a simple schematic and Figure 5.7-2 for a typical partial discharge coupler installation on a winding. Depending on how many legs there are in each of the phases will govern how many PD couplers can practically or economically be installed. For example, if your generator is a two-circuit Y connection, then two couplers per phase yields six (2 × 3 phases) in total and that is a typical number to monitor. However, if your machine Line end

Parallel #1 Parallel #2

Permanent capacitive couplers installed at line end of all parallels

Neutral end To PDA

Figure 5.7-1 A simple schematic for PD coupler locations on the generator winding.

5.7 CAPACITIVE COUPLING

275

Figure 5.7-2 Typical partial discharge coupler installation.

is a four- or six- or eight-circuit Y connected machine, you would need to determine how many legs makes sense from a practical and economical point of view to monitor. More PD information on the winding is a good idea, but there is extra time and effort in collecting and interpreting the data. The couplers collect the partial discharge readings as they become available and are connected to a “smart box” which interpret and display the PD activity in a variety of ways according to the user preference. The “smart box” contains all the hardware that then feeds into the software package that displays on a computer screen. This is a huge advancement from manually taking readings and needing a skilled technical expert to interpret the data off an oscilloscope. Once the couplers are installed, PD readings can be taken at any time and during any load condition without invasive procedures on the winding, it is all done while the machine is online during commercial operation. The loading on the generator is varied to provide distinction between the types of PD activity present. Since temperature, humidity, and vibration have an effect on the level of PD activity, changing the loading conditions will cause variations in terminal voltage and stator current. Hence, there will be variation in winding temperature and stator bar/coil bounce forces or bar/coil vibration. Since positive PD is associated with surface discharges, one could conclude, for instance, that if the magnitude of the measured positive PD pulses increased with load and there were no increases in the magnitude of the negative PD pulses, then there may be slot discharge activity caused by loose stator wedges. Similarly, if there is equal positive and negative PD activity one could conclude that there is discharge in the bulk of the groundwall insulation, and so forth. In order to verify the source of PD pulses external to the winding, most hydro generators with circuit rings use two capacitive couplers physically on opposite sides of the winding usually about the same distance from the main terminals to

276

CHAPTER 5

MONITORING AND DIAGNOSTICS

distinguish the source. This is accomplished by comparing the pulse arrival times, if a pulse is seen at both couples at or near to the same time it must have come from the main terminals and thus from outside the generator. Where circuit rings are not available it is possible to use two capacitive couplers per phase installed on the isolated phase bus, where they can be separated by a number of feet. When a PD pulse is measured on both couplers, its direction may be determined by which coupler is the first to see the pulse. Therefore, if the coupler farthest from the generator picks up the PD pulse first, then its direction is toward the generator, and vice versa. If the coupler closest to the generator is the first to see the pulse, then it had to come from the generator. There has been much progress made in the directional capability of capacitive coupling, but again, noise is a problem and often masks the true PD being measured and requires complex algorithms within the “smart box” to isolate. There is a wealth of partial discharge literature available from all kinds of sources, the reader is encouraged to seek out this literature if a more in depth discussion is of interest.

5.8 STATOR SLOT COUPLER Although rarely used on hydro generators because of cost, another method for detecting PD is the stator slot coupler (SSC). This coupler is basically a tuned antenna with two ports. The antenna is approximately 18 long and is embedded in an epoxy/glass laminate with no conducting surfaces exposed. SSCs are installed under the stator wedges at the line ends of the stator winding, such that the highest voltage bar/coils are monitored for best PD detection. This SSC now takes most if not all of the space where the bar/coil depth packing would reside. Ensure there is sufficient space to install the SSC before purchasing. Since the SSC is also installed lengthwise in the slot at the core-end, its two-port characteristic gives it inherent directional capability. The problem of noise is virtually eliminated in the SSC. Although the SSC has a very wide frequency response characteristic that allows it to see almost any signal present in the slot in which it is installed, it also has the characteristic of showing the true pulse shape of these signals. This gives it a distinct advantage over other methods that cannot capture the actual nature of the PD pulses. Since PD pulses occur in the 1–5 ns range, and are very distinguishable with the SSC, the level of PD activity can be more closely defined. In addition, dedicated monitoring devices have been devised to measure the PD activity detected in the SSC. The capability for PD detection using the SSC and its associated monitoring interface is enhanced to include measurement in terms of the positive and negative characteristic of the pulses, the number of the pulses, the magnitude of the pulses, the phase relation of the pulses, and the direction of the pulses (i.e. from the slot, from the endwinding, or actually under the SSC itself at the end of the slot). This option to measure PD can be very expensive depending on how many slots are monitored. See Figures 5.8-1–5.8-3 for a typical arrangement for an SSC.

5.8 STATOR SLOT COUPLER

SSC under stator wedges

Stator core

Stator bar under SSC and wedges Endwinding Stator wedges

SSC output cables

Figure 5.8-1 Plan view of the SSC installation.

Wedge

Depth packing

Wedge slider SSC

Top pad Stator core Top bar

Side packing

Slot separator

Bottom bar

Bottom pad

Figure 5.8-2 Stator slot coupler end view.

277

278

CHAPTER 5

MONITORING AND DIAGNOSTICS

Figure 5.8-3 Installed stator slot coupler.

5.9 ROTOR The rotor has minimal instrumentation due to its dynamic nature. However, technology has progressed to the point that wireless devices such as a data collection box attached to the rotor shaft that collects temperature readings from mounted temperature sensors and then transmits to a receiver outside of the generator are available. It should be recognized that even though the technology exists, the rotor is a spinning mass and can render a poorly attached data collection box useless not to mention the damage the can be done if this box gets loose inside the rotor. Extreme caution should be exercised when using this type of technology. Regardless, there are numerous ways to monitor the performance of the rotor indirectly. Rotor unbalance may show up as increased rotor vibration (amplitude or orbit

5.9 ROTOR

279

changes). Parameters that may be monitored include vibration, field current, field voltage, and vibration of the rotating exciter. 5.9.1.1 Rotor Winding Temperature Rotor winding temperature is generally measured as a function of the winding resistance during the original heat run before in service. This produces the average temperature of the winding but does not indicate the temperature of the hottest part of the winding. Advancements in technology now allows for temperature monitoring of the complete rotor pole body assembly as well as the field winding while the machine is in service. One such technology uses special patented probes that are inserted through the back of the stator core until the probe is flush with the stator bore. The probe measures the temperature and outputs a 4–20 mA signal that is processed through proprietary software that displays the temperature on a computer screen [9]. As many probes as required are installed axially along the stator core to monitor the entire height of the pole body and field winding. This monitoring will also give inter polar temperatures so the whole spectrum of temperature of the rotating mass can be analyzed. Another method to measure temperature on the rotor is using optical Fiber Bragg Gratings technology. This technology is completely immune to interference from electromagnetic, electrostatic, or radio frequency sources. Further, the local sensors that are part of this optical fiber are nonconductive and can be used anywhere on the rotor field winding to give a hotspot temperature where desired [10, 11]. 5.9.1.2 Pole-face Heating Pole-face heating can result from reduced airgaps due to the tooth ripple flux inducing relatively high-frequency eddy currents in the pole-face laminations. Poleface heating can also occur due to unacceptable high negative sequence currents circulating in the stator core. Continuous negative sequence in the stator results from unbalanced loads which must be limited to levels indicated in Ref. [3]. Short-term negative sequence limits must be defined by the OEM. Asynchronous operation or operating the machine with cut out coils or bars can also cause this issue. 5.9.1.3 Rotor Winding Ground Alarm Rotor winding grounds are actually a leakage of field current to the referenced ground of the rotor itself. In addition, grounds can occur via the slipring and field lead insulating systems. Finally, grounds can also occur in the excitation system external to the rotor. Rotor ground protection is used to detect a field ground. The rotor winding ground indication is normally connected as an alarm and not an actual generator trip. It is left up to the operator or suitable authority to decide whether the unit should be taken offline. Although it may be common for utilities to operate with a single rotor ground for short periods of time until a convenient outage, this practice has a high degree of inherent risk. Should a second ground occur anywhere on

280

CHAPTER 5

MONITORING AND DIAGNOSTICS

the rotor, current will circulate through the two ground points, creating overheating in the affected rotor components, and shorting out the poles in between the two grounds. This could cause significant unbalanced magnetic pull on the rotor if sufficient number of poles are removed from the circuit. This can cause guide bearing damage and possibly a rotor/stator collision. The first ground alarm may be caused by a more serious matter such as melting of the interpole connections. Conventional relaying protection is unable to detect the open circuit, and when left undetected the melting copper connections will eventually develop a ground fault potentially causing serious damage to the machine, more about this topic in Chapter 6. 5.9.1.4 Rotor Winding Shorted Turns Detection Rotor winding shorted turns, or also known as “turn shorts,” can occur from an electrical breakdown of the turn insulation, foreign material trapped between turns, or mechanical damage of the turn insulation allowing adjacent turns to come into contact with each other. When shorted turns occur, the total ampere turns produced by the rotor are reduced since the number of useful turns has been reduced by the number of turns shorted. The result is typically a small increase in required field current input to maintain the same number of ampere turns at the same load point. This in turn may result in an increase in the rotor winding temperature. At the location of the shorted turn, there is a possibility of localized heating of the copper winding and arcing damage to the insulation between the turns. Depending on the fault type and location this damage can propagate and worsen such that more turns are affected, or the groundwall insulation becomes damaged and a rotor winding ground occurs. Offline methods for detecting shorted turns include winding impedance measurements as the rotor speed is varied from zero to rated speed. Recurrent surge oscillograph (RSO) based on the principle of time domain reflectometry and impedance testing (VIW) can be used when the rotor is stationary (see Chapter 11 for more information). In addition, if there are many poles suspected of having a significant number of shorts, this may be identified by producing an open circuit saturation curve and comparing it to the original open circuit saturation curve. If the field current producing the required rated terminal voltage has measurably increased from the original design curve, then a significant number of shorted turns may be present. To identify shorted turns while the machine is online, rotor flux monitors installed on the stator face looking at the airgap can be used as shown in Figure 5.9-1. This flux probe is very small and does not interfere with machine operation. The data taken from the flux probe is processed through proprietary computer algorithms that analyze the flux patterns as shown in Figure 5.9-2. Then the software calculates where the shorted turns may be located as shown in Figure 5.9-3 [12]. It is clear from the figure that pole number 72 may have shorted turns when referenced to pole 66 that should not have any. This system can also measure the rotor shape, and changes in the airgap between rotor and stator [13].

5.9 ROTOR

281

Figure 5.9-1 Shows a flux probe installation in progress on the stator core. Source: Courtesy of Qualitrol-Iris Power.

Flux signal

Flux density

Relative flux amplitude (%)

80 60

Raw flux

40 20 0 –20 –40

Integrated flux

–60 –80 5

10

15

20

25

30

Time (ms)

Figure 5.9-2 Iris Power.

Software output for flux probe data collected. Source: Courtesy of Qualitrol-

5.9.1.5 Shaft Speed The speed is generally measured by a probe mounted below or above the rotor, looking at a toothed wheel or key phasor on the rotor shaft. The speed signal is normally logged and used for various protections on the generator such as overspeed trip, creep detection, runaway detection, etc. The frequency is usually taken from the output bus of the generator.

282

CHAPTER 5

MONITORING AND DIAGNOSTICS

Figure 5.9-3 Polar representation of flux probe data collected showing shorted turns on pole 72. Source: Courtesy of Qualitrol-Iris Power.

5.9.1.6 Rotor Vibration Rotor vibration refers to vibration monitoring on the shaft and the bearings [14]. Shaft vibration is the movement of the shaft in relation to the bearing mounts or generator footing, where the vibration probe is mounted. Bearing vibration is the movement of the bearings relative to the generator footing. Normally, vibration is measured in units of displacement (i.e. thousandths of an inch or micrometers, peak to peak on the displacement signal obtained). Vibration is usually measured as velocity and then converted to displacement using Equation (5.5): D=

V 2πF

(5.5)

where, D is the displacement (for example) in mm peak, V is the velocity in mm/s, and F is the frequency in Hz. This is a handy formula as you can isolate frequency components of interest. One such component for a 60 Hz line frequency would be the 120 Hz component. For example, if the velocity measured was 9 mm/s, then: D=

9 = 0 011 93 mm 2π120

5.9 ROTOR

283

Multiplying the result by 1000 yields a peak value of 11.93 μm and by multiplying by 2 gives the peak to peak value of 23.86 μm. All generators operate under strict bearing and shaft vibration guidelines and limits usually set down by the manufacturer, or IEEE, IEC, etc., but these are often modified by the experienced operator. High vibration is caused by mechanical issues (turbine and runner) and by electrical issues via stator/rotor eccentricity and circularity excursions from design values. Rotor mechanical imbalance is caused by conditions such as rotor circularity or concentricity, bearing loading and coupling alignment issues such as a bend (dog leg) in the shaft. The level of severity is usually determined by the magnitude of vibration present and may require an outage to correct the source of the vibration or to apply balancing weights to offset the imbalance. For example, if a rim must be reshrunk to correct circularity or concentricity deviations from allowable design parameters, the new shape or concentric position may necessitate an addition/removal of the existing weight or a new weight installed in a different location. For mechanical vibration problems, vibration levels generally remain constant regardless of field current changes but will vary with the shaft rotational speed and, possibly wicket gate position. For example, a hydro generator typically has a “rough” zone where the runner just does not perform favorably at certain gate positions. In this case, the generator should not be operated in these “rough” zones. The “rough” zone is well below the efficient point in the runner operating range anyway, so, the moral is, do not spend too much time in this zone. When the generator is constructed rotors are typically circular and concentric about the rotating axis. Similarly, the stator is typically circular as well. The rotor is typically designed to operate indefinitely with only a small eccentric error with respect to stator center without adverse effect to machine components. This error means that one side of the airgap will be less than the minimum design putting extra stress on components due to the magnetic interaction between the rotor and stator. Each OEM has their own design criteria, for example one OEM may design for a 0.79 mm (1/32 ) offset from perfect center of the rotor and stator. This design criteria is not for fatigue life but for magnetic stability so the generator will be mechanically stiff enough to not allow the rotor to hit the stator. There are inherent built in design limits for rotor and stator offsets, these limits should be confirmed with the OEM. In the real world, no rotor is perfectly circular or concentric. Each rotor in service will have some circularity deviation whether it is a “bump” where a few poles are sticking out more than others from the perfect circle coupled with a nonconcentric rim on the spider. Since generator designs vary considerably, it is important to identify specific design limits for each machine. Operating outside these limits may produce unfavorable vibration, reduced component fatigue life, and/or forced outages. Consultation with the OEM on the measured operating limits is advised. The rotor may be off center and noncircular and not cause any vibrations of significance. It is important to recognize that design variations in concentricity, in other words, airgap reduction, causes extra stress in that area for the stator frame

284

CHAPTER 5

MONITORING AND DIAGNOSTICS

and may result in cracked keybar to frame welds and can also induce vibrations that cause objectionable noise. Vibratory noise of this nature is a complex subject and the reader is referred to the OEM for consultation. Further, unbalanced current flow in parallels of the stator winding caused by the rotor not being centered in the stator can activate the split phase protection if so equipped. Bearing and shaft vibration on both ends of the generator may be monitored to detect any or all of the above abnormalities in terms of the magnitude, phase, and frequency of the vibration at variable load conditions. A frequency analysis can also be performed to provide more detail of the vibration pattern. 5.9.1.7 Torsional Vibration Monitoring Although rarely occurring on a hydro generator, occurrence of oscillating shaft torques, particularly those of a severe nature that can affect the remaining life of the generator shaft may need to be considered. By measuring the pertinent parameters and entering all the event information into a model of the generator shaft, the loss of life of the shaft can be calculated. Torsional events are generally caused by severe system disturbances or power system resonant frequencies that are inadvertently stimulated. These cause the generator shaft to respond by oscillating sub synchronously on top of the shaft operating speed. The effect causes excessive oscillating torque in the shaft. If not dampened, the oscillation will eventually run away and cause failure of the shaft. As oscillations are dampened, they decay to zero. Excess torsional stress on the generator shaft can lead to loss of mechanical life, so it needs to be accounted for in the original design. 5.9.1.8 Shaft Voltage and Grounding Brush All bearings (on the side of a hydro generator rotor opposite the turbine) have insulation fitted in the bearing base or entire upper bracket to prevent the circulation of possibly damaging current through the bearings. A principal source of such currents is induced shaft voltage. Shaft voltages are most often a result of some asymmetry in the magnetic circuit of the machine, which causes some net flux linkage with the rotor shaft that induces a voltage in it. These voltages are alternating at a multiple of system frequency and are low, usually less than 15 V. If there is a continuous low impedance path from the shaft ends through the bearings and base, the current produced by the shaft voltage may be 100 A to several thousand amperes and enough to rapidly damage the bearing. Terminals for checking the installed bearing insulation resistance are sometimes provided. Generators with no bearings above the rotor (opposite the turbine) do not require bearing insulation. However, any fittings or apparatus at this end of the machine that might contact the shaft in any way or during high vibration must be insulated. On units with a static excitation supply, a shaft grounding brush may be supplied immediately below the rotor. Static excitation can supply a small current to the rotor through capacitive coupling between the field coils and the poles. While

5.9 ROTOR

285

Figure 5.9-3 Shaft grounding brush.

the shaft is usually grounded through the water in the turbine, the brush is additional insurance when operating unwatered. Such brushes are not normally sized for large currents from faults. The grounding device consists of a carbon brush(es) or copper braid(s), with one end riding on the rotor shaft and the other connected to ground, see Figure 5.9-3. 5.9.1.9 Bearing and Oil Temperatures Bearing temperatures are affected by vibration, alignment, oil condition, bearing preloads, cooler efficiency, and the amount of oil fill in the bearing. Excessive bearing temperatures can result in an overheating and a subsequent wipe. It may not take very long for a bearing to wipe due to temperature increases that occur very rapidly, so monitoring must be continuous, especially after an overhaul or an alignment. The temperatures are monitored by thermocouples embedded in the bearing segments. The oil that is circulating inside the enclosure surrounding the thrust bearing can be cooled using water and cooling coils. Inlet and outlet temperatures should be monitored to ensure the design water temperature increase is achieved thus removing the proper amount of heat and if possible should also monitor for water leakage. Also, thermocouples inside the oil pot should be monitored to ensure the oil is the proper temperature. As stratification of oil can occur in oil pots, causing uneven or increased bearing temperatures if an increase in oil cooling occurs due to cooling water flow or replacement cooler design changes. Always consult the OEM if a change of cooling coil design, location or temperature monitoring is desired.

286

CHAPTER 5

MONITORING AND DIAGNOSTICS 90° 45°

Modbus TCP

Plant control system

Sensor signals

45°

45°

0°

45° 45°

180° 45°

45°

270°

Stator PhRef

Admin access

SCADA

Protection relay outputs

AGTracll

Data transfer

Air gap continuous monitoring processing unit

Conditioners

Air gap signals

Rotor

Modbus communication

Turbine Air gap system diagnostic software

Figure 5.9-4 Airgap system architecture. Source: Courtesy of Qualitrol-Iris Power.

5.9.1.10 Online Airgap Monitoring Online airgap monitoring systems have come to the forefront in the last decade and continue to be one of the number choices of parameters to monitor while the generator is in service. These systems are very useful as they allow the user to capture in real time the shapes of the rotor, stator, and the airgap dimension during any operating condition chosen. Information during startup, shutdown, field on, field off, full load, condense mode and anything in between can be saved, analyzed and then used during the planning of generator outages if remedial work needs to be completed. This system is very useful in power plants that have alkali aggregate reaction (AAR) to monitor the growth and movement of the concrete as this can have a profound effect on the airgap and the generator stator and related components if left unchecked (see Chapter 8 for more information). An example of the system architecture is shown in Figure 5.9-4. The system allows the information to be integrated with the plant control system or better known as SCADA which is very beneficial.

5.10 EXCITATION SYSTEM 5.10.1.1 AC Power into Exciter The AC power into the exciter is the power consumed by the excitation system. In static excitation systems, it is the power delivered by the AC supply transformer connected to the generator output leads which then feeds the excitation cubicles. In rotating excitation systems, it is the input to the main rotating exciter from a

5.10 EXCITATION SYSTEM

287

rotating or static pilot exciter. For any particular load, the AC power should be at a certain level and a check can be made of the consumption to see if it is in the correct range for the load output. 5.10.1.2 DC Power Out of Exciter The DC power out of the exciter is the power delivered to the rotor winding after rectification of the AC signal. For a static exciter, this is accomplished by thyristor bridges and for a rotating exciter it is accomplished by the commutator circuit, and by diodes or thyrisors for brushless exciters. The difference between the AC power in and DC power out should only be the normal losses of the particular excitation equipment. A large differential between the two is an indicator of a possible excitation system fault. 5.10.1.3 Main Exciter Cooling Air Inlet Temperature The cooling air temperature into the main excitation system, whether a rotating or a static exciter, is generally the ambient air of the powerhouse. Cooling air is sometimes filtered and directed over the excitation equipment for cooling and is then discharged back into the powerhouse. The cooling air temperature will vary with the seasons as the powerhouse general atmosphere temperature changes. Therefore, it is more likely to have temperature issues relating to the cooling air supply for the excitation equipment in the summer. It is important to remember that brushes for a rotating exciter have an optimum temperature range. Deviations from this range can result in excessive brush wear and/or poor commutation. Remember that the brush current loading accounts for the majority of the total temperature of the brush, so it is important that all brushes are sharing the current as equally as possible. 5.10.1.4 Main Exciter, Cooling Air Outlet Temperature The cooling air temperature out of the main exciter (i.e. static or rotating) over the inlet temperature is an indicator of losses of the main excitation components. Therefore, the temperature rise of the outlet cooling air above the inlet cooling air of the main exciter should be monitored to ensure proper operation. 5.10.1.5 Sliprings and Brush Gear, Cooling Air Inlet Temperature The slipring and brush gear air is generally the air from the powerhouse at its ambient value and sometimes forced over the slipring and brushes from the rotor assembly. In some closed cooling designs, generator cold air is pumped into the slipring enclosure by the rotor and then sent into the powerhouse. In this case, a duct assembly would be placed inside the pressurized rotor cold air compartment to acquire air and may or may not be filtered. Care must be taken to ensure clogging of the filters does not occur as the airflow will be reduced and overheating of the brushes may occur. This may lead to excessive brush wear and increased maintenance costs.

288

CHAPTER 5

MONITORING AND DIAGNOSTICS

5.11 REFERENCES 1. Roland, S. and Bahadoorsingh, S. (2008). A framework linking insulation ageing and power network asset management. ISEI 2008 IEEE International Symposium on Electrical Insulation, Vancouver, Canada (9–12 June 2008). 2. Klempner, G. S. (1995). Expert system techniques for monitoring and diagnostics of large steam turbine driven generators. IEEE/PES Winter Power Meeting/Panel Discussion Paper, New York (February 1995). 3. IEEE (2015). IEEE C50.12-2005, IEEE Standard for Salient-Pole 50 Hz and 60 Hz Synchronous Generators and Generator/Motors for Hydraulic Turbine Applications Rated 5 MVA and Above, New York, IEEE. 4. Moraes, J. D., Rodriguez Villalba, J., and Salatko, V. (1979). Selection of design features for 737 and 823 MVA hydrogenerators for ITAIPU project. IEEE Transactions on Power Apparatus and Systems Pas-98(6), 2329–2337. 5. IEEE (2000). IEEE Std 1434-2000 IEEE Trial-Use Guide to Measurement of Partial Discharges in Rotating Machinery, New York, IEEE. 6. Liese, M. and Brown, M. (2008). Design dependent slot discharge and vibration sparking on high voltage windings. IEEE Transactions on Dielectrics and Electrical Insulation 15(4), 927–932. 7. Stone, G. C., Maughn, C. V., Nelson, D., and Schultz, R. P. (2008). Impact of slot discharges and vibration sparking on stator winding life in large generators. IEEE Electrical Insulation Magazine 24(5), 14–21. 8. Electric Power Research Institute (1988). Handbook to Assess Rotating Machine Insulation Condition, EPRI Power Plant Series, Vol. 16. EPRI. 9. Vibrosystm (2018). Rotor pole temperature monitor. http://vibrosystm.com/energy/ hydro-complete-monitoring-solutions/hydro-sensors/, http://vibrosystm.com/wp-content/uploads/9621-25D1A-100.pdf (accessed 5 July 2018). 10. Hudon, C., Lévesque, M., Essalihi, M., and Millet, C. (2017). Investigation of rotor hotspot temperature using Fiber Bragg Gratings. Proceedings of the 2017 IEEE Electrical Insulation Conference, Baltimore, MD (11–14 June 2017). IEEE. 11. Hudon, C., Guddemi, C., Gingras, S., Leite, R. C., and Mydlarski, L. (2016). Rotor temperature monitoring using fiber Bragg gratings. Electrical Insulation Conference 2016, Montreal, Canada (19–22 June 2016). IEEE. 12. Sasic, M., Stone, G., Stein, J., and Stinson, C. (2013). Detecting turn shorts in rotor windings. IEEE Industry Applications Magazine, March/April, pp. 63–69. 13. IRIS Power (2018). RFAII-S: rotor flux monitoring (motors and generators). https://irispower.com/products/rfaii-s (accessed 15 April 2018). 14. ISO 7919-5 (2005). Mechanical Vibration: Evaluation of Machine Vibration by Measurements on Rotating Shafts. Part 5, Machine Sets in Hydraulic Power Generating and Pumping Plants, Geneva, ISO.

5.12 FURTHER READING Culbert, I. M., Dhirani, H., and Stone, G. C. (1988). Handbook to Assess Rotating Machine Insulation Condition, EPRI Power Plant Series, Vol. 16. EPRI.

5.12 FURTHER READING

289

IEEE (2000). IEEE Std 1-2000 IEEE Recommended Practice: General Principles for Temperature Limits in the Rating of Electrical Equipment and for the Evaluation of Electrical Insulation, New York, IEEE. IEEE (2011). IEEE Std 492-2011 IEEE Guide for Operation and Maintenance of Hydro Generators, New York, IEEE. IEEE (2014). IEEE Std 1434-2014 Guide to the Measurement of Partial Discharges in AC Electric Machinery, New York, IEEE.

CHAPTER

6

GENERATOR PROTECTION

6.1 BASIC PROTECTION PHILOSOPHY Protection provides the most crucial function to maintain the reliability of the power system. Protection is often called a protection system, since it comprises many discrete components – primary and auxiliary relays, current transformers (CTs), potential transformers (PTs), circuit breakers, tele-protection system, etc. The key function of any protection system is to minimize the physical damage to the power equipment due to an electrical fault in the system or abnormal condition of the power equipment (overspeed, loss of synchronism, etc.). Another critical function of any protection system is to safeguard the personnel, who operate and maintain the power equipment, as well as the public. It is also important that the protection systems be coordinated in order to minimize the extent and duration of the outage. The proper coordination of the protection systems is achieved through the implementation of correct relay settings supported by extensive protection system studies. Protection systems are inherently different from other electrical systems that exist in a power plant. They are less likely to operate, but if and when they do, they must operate rapidly and accurately. A failure to operate, when they need to, could lead to the catastrophic failure of equipment, subsequent production loss, as well as putting personnel safety at risk. Thus, their periodic and thorough maintenance is of paramount importance and must not be overlooked. Purchasing, installing, designing, and maintaining a generator protection system is not an insignificant cost. Nonetheless, electric power generators must be adequately protected as they are the critical and most expensive electrical apparatus in a power plant. Given the physical proximity between the generator and the main output transformer, these two apparatuses can be protected together in the same zone of protection sharing some of the protective functions such as a unit differential protection. Handbook of Large Hydro Generators: Operation and Maintenance, First Edition. Glenn Mottershead, Stefano Bomben, Isidor Kerszenbaum, and Geoff Klempner. © 2021 The Institute of Electrical and Electronics Engineers, Inc. Published 2021 by John Wiley & Sons, Inc.

291

292

CHAPTER 6

GENERATOR PROTECTION

It is impossible to describe or prescribe a protection system that may be fit for all generators. Therefore, the description we attempt here is of the most commonly encountered protection arrangements and functions for a generator. The basic protective components are well known, and one can find relay setting calculation methods readily available from a number of protection standards, guides, and technical writings. In addition to electrical protections in the power plant, there are mechanical protections provided for the monitoring of the cooling water temperature and pressure for the surface air coolers and bearings. There are a multitude of other systems that monitor everything from partial discharge and vibration to flux levels in the airgap. Readers are required to review, understand, and apply the safety procedures found in Section 7.2.1.

6.1.1

Generator Protection System

Generator protection relays are designed to protect the generator and its associated power equipment against short circuits and abnormal conditions. The functional capability of detecting the specified condition in a generator is represented by a protective function. Thus, there is a dedicated protection relay for every protective function. If a relay only monitors and protects against a single condition, it is said that the relay is a “single function device” as shown in Figure 6.1-1. In the past, most relays were single function devices. With the advent of microprocessor technology, however, relay manufacturers have combined several functions in one unit or device. These are called “multifunction” relays, as shown in Figures 6.1-2 and 6.1-3. According to Ref. [2], the “multifunction” relays are defined as: “A device that performs three or more comparatively important functions that could only be designated by combining several device function numbers” Some multifunction relays are dedicated for the protection of transformers, while others protect transmission lines, motors, or generators. Advances in technology have led to significant cost savings. Today, a multifunction relay built with five protective functions is most probably less expensive than five single function relays. A multifunction relay containing all the protective functions required for the protection of a generator can be installed in combination with a few discrete relays providing backup protection for critical functions. Alternatively, two or more multifunctional relays can be applied, providing partial or comprehensive redundancy. There are many combinations of these discrete and multifunctional relays that can be adopted, depending on when the power plant was built, the size of the generators, system conditions, the philosophy of the protection engineer, and many other factors. Figure 6.1-4 shows a typical generator and transformer protected by a multifunctional digital protection system. Relays or protection devices are divided into two categories according to how they process data. The first category is that of analog relays; the second is that of numerical (also called digital) relays. Bear in mind that a relay can be electronic but still process the data in an analog manner. The advantages of numerical relay

6.1 BASIC PROTECTION PHILOSOPHY

293

Figure 6.1-1 Singlefunction protective relay.

Figure 6.1-2 Schweitzer Engineering Laboratories SEL 300G multifunction relay.

are accuracy and flexibility in use. For instance, a numerical relay offers user-shaped protection widows such that the user can change the shape of the operation/nonoperation areas for a specific function of the relay. Furthermore, the shape of the region of operation may change according to system conditions (adaptive function). Finally, there is another approach for protecting large generating units using so-called expert protection systems. The idea is to protect the unit based not only on the basic protective functions given in Section 6.2 but also built-in-expert system in

294

CHAPTER 6

GENERATOR PROTECTION

Figure 6.1-3 Beckwith Electric Co. Inc. M-3425A multifunction relay. Source: Courtesy of Beckwith Electric Co.

These functions are available in the comprehensive package. A subset of these functions are also available in a base package.

M-3425A typical connection diagram

Utility system

52 unit

This function is available as a optional protective function. This function provides control for the function to which it points.

M-3425A Targets (optional)

50 BFPh

CT

50 DT

VT (note 1)

Integral HMI (optional)

CT (residual) (note 4)

Metering 87

Waveform capture

52 Gen

25 VT

IRIG-B Front RS232 communication 81R

Rear RS232 communication

81A

81

27

59

24

M (Metering)

VT (Note 1)

Rear ethernet port (optional) Rear RS-485 communication

(Note 3)

M-3921

59X

Multiple setting groups Programmable I/O

64F

64B

27

Self diagnostics Dual power supply (optional)

78

60 FL

51V

50/27

40

32

21

50

49

46

M

CT

(Metering)

Breaker monitoring

67N operating current (software select) I 50 BFN 67N

3V (Calculated) V

Trip circuit monitoring

67N Polarization (software select)

V

50N

51N

3I

Event log

(Note 5) 3V (Calculated) 59D line side voltage (software select)

V

59D

27 32

27 TN

64S

59N

High-impedance grounding with third harmonic 100% ground fault protection

R

87 GD

50 BFN

50N

CT (neutral) (notes 2 and 5)

51N R

Low-impedance grounding with ground differential and overcurrent stator ground fault protectin

Figure 6.1-4 Beckwith M-3425A multifunction generator protection outline (typical). Source: Courtesy of Beckwith Electric Co.

6.2 IEEE DEVICE NUMBER

295

the form of diagnostic prescriptions. Invariably, building the expert system consists of defining probable causes for a particular combination of symptoms, expressed as a probabilistic tree. There are a limited number of OEM’s that have this type of expert protection system available for installation on existing and new machines. It is recommended that the original OEM be consulted if an expert system for your machine is being considered.

6.2 IEEE DEVICE NUMBER A numbering system, according to a worldwide accepted nomenclature, identifies protective functions. The functions shown in Table 6.2-1 are typical of generator protection [1, 2]. A number of the functions included in are so important that they will always find their way into the protection scheme of any generator (e.g. device number 40, 59G, 87). Others may be omitted in some applications (e.g. 60). The larger and more expensive the generator, the more comprehensive is the protection applied to the machine. As explained earlier, for most large machines, some of the applied protective functions are overlapped by more than one relay or protective device.

TABLE 6.2-1 Generator protective device function numbers

Device #

Function

15 21

Speed/Frequency matching Distance relay – generator system phase backup protection for unclear system phase faults with a time delay for coordination V/Hz protection Synchronizing or synchronism-check device Undervoltage relay Undervoltage relay tuned to the third harmonic frequency (180 Hz) Reverse/Low forward power protection Inadvertent energizing protection using voltage supervised overcurrent relaying Loss of field protection Stator unbalanced current protection – negative sequence relay Breaker failure protection Voltage controlled time overcurrent relay phase backup protection Overvoltage protection relay – generator phase overvoltage protection Stator to ground fault protection Voltage balance relay – detection of blown transformer fuses or loss of potential Rotor ground fault protection Loss of synchronism or out-of-step protection Over- and under-frequency relay Differential relay. Primary phase-fault protection for the generator

24 25 27 27TH 32R/32F 50/27 40 46 50BF 51VC 59 59G 60 64F 78 81O/U 87

296

CHAPTER 6

GENERATOR PROTECTION

6.3 BRIEF DESCRIPTION OF PROTECTIVE FUNCTIONS It is beyond the scope and purpose of this book to go into a detailed description of each protective function and the various schemes that are used for a generator protection. Instead, a basic description of some of the protective functions and their application will be provided. For those readers interested in an in-depth discussion on the topic, Refs. [1, 3, 4] should be consulted. For the same reason, no specific values are recommended for setting protective relays. These values often depend on the particular machine and system to which it is connected. There are numerous sources for information on the setting of protective relays. The manufacturers’ manuals will provide a good place to start. Others are provided in the References section at the end of this chapter.

6.3.1 Synchronizer and Sync-Check Relays (Functions 15 and 25) The combination of a synchronizer (15) with a synch-check relay (25) provides the means by which the unit can be brought up to speed automatically and synchronized to the system. Before doing so, the amplitude of the voltages of the system and generator terminals must be within a narrow margin so that the breaker can be closed, as must be the angle of the terminal and system voltages. The slip, which is the frequency difference between the machine and the system, must be lower than a prescribed value. Almost always, two relays are provided: the synchronizer and the sync-check. This division of labor is based on the need to avoid the destructive results of synchronizing a unit out-of-step due to the failure of a single protective device. In older installations, closing the breaker is done manually while the synccheck relay monitors all voltages, vector angles, and frequencies, making sure they are within their prescribed values. Infrequently, some operators close the breaker by keeping the “close” button depressed when the unit is brought to the right speed and voltages, allowing the angle to be taken care of by the sync-check relay. This practice has resulted in more than one unit synchronizing out-of-step due to a failure of the relay (function 25). The failure can cause serious and catastrophic damage. Thus, it is imperative that during manual operation the actual breaker closing signal be sent when the conditions for synchronization are met, leaving the sync-check system as a backup device, as it is supposed to be.

6.3.2

Stator Ground Protection (Functions 59G and 27TH)

The vast majority of generators are generally grounded through a method called the high impedance grounding. As shown in Figure 6.3-1, the high impedance grounding is accomplished by connecting the grounding distribution transformer at the generator neutral with a resistor at the transformer secondary. This arrangement generally limits a ground fault current from 5 to 20 A. If there is a phase to ground

6.3 BRIEF DESCRIPTION OF PROTECTIVE FUNCTIONS

297

52

G

R

59N

27TN

Figure 6.3-1 High impedance ground protection and grounding transformer.

fault at the terminal, the full generator phase to neutral voltage will be imposed across the primary winding of the grounding transformer. An overvoltage relay (59G), which is connected across the secondary resistor for the stator ground protection, will measure the equivalent secondary voltage for the relay operation. But the overvoltage relay is only sensitive enough to detect a ground fault for approximately 95% of the stator winding from the generator terminal, since there is a small amount of zero sequence current flowing through the ground under normal operation. For the protection of the remainder of the stator winding, which is 5% from the neutral, an additional ground protection may be required. There are two most widely used techniques for the protection of the bottom 5% coverage: Thirdharmonic voltage-based techniques and neutral subharmonic voltage injection. Third harmonic voltage-based technique utilizes the third harmonic voltage present in the machine during normal operation. During a ground fault, however, these third harmonic voltages are considerably reduced. Under voltage relay (27TH), tuned to 180 Hz, is connected across the secondary of the grounding distribution transformer to detect the ground fault for the bottom 5% of the stator winding. With the reduction of the third harmonic voltage resulting from the ground fault, the under voltage relay will drop out to close its contact to trip the generator or alarm the operator. The second technique, which is based on neutral subharmonic voltage injection, is provided to detect the ground fault throughout the stator winding, as shown in Figure 6.3-2. The system works by injecting a subharmonic voltage at the neutral, typically 15 or 20 Hz, and measuring the resultant 15/20 Hz current. If there is

298

CHAPTER 6

GENERATOR PROTECTION

GSU

Generator G

RL

15/20 Hz

Relay

Figure 6.3-2 Subharmonic voltage injection scheme.

a ground fault, the magnitude of 15 Hz current increases to trip the generator. The detailed setup and explanations of the scheme is beyond the scope of this book.

6.3.3

Phase Backup Protection (Functions 21 and 51VC)

Unlike the other protective functions provided for the generator, the phase backup protection is implemented to detect faults outside the generator zone. This backup protection is necessary in case that system faults external to generator zone are not cleared due to failures of line or system protection. Two types of relays are commonly used for the system phase backup protections: voltage controlled (51VC)/restrained (51VR) time overcurrent relay and distance relay (21). The voltage controlled/restrained time overcurrent relay is typically found in an industrial plant where the generator is directly connected to a bus which serves distribution and utilization equipment using overcurrent devices [5]. The time overcurrent relay should be supervised by a voltage element in order to discern a fault from an overload condition. If the current and time settings are set too low, the relay will trip the generator unnecessarily on normal overloads. If the settings are set too high to allow coordination with downstream devices, the relay may not respond at all in the event of a system fault. This is due to the generator fault current decrement characteristic varying in subtransient, transient, and synchronous regions.

6.3 BRIEF DESCRIPTION OF PROTECTIVE FUNCTIONS

299

The distance type relay is predominantly used industry-wide according to an IEEE survey conducted in 1990. It is usually applied where the output of the generator is stepped-up via the generator step-up transformer to a transmission voltage-level, which is also protected with distance relay(s). Backup phase distance protection relays are usually time-delayed to coordinate with system or line protections. Current source for phase fault backup protections is normally provided from current transformers connected to the neutral side of the generator and voltage source for the relay comes from voltage transformers on the generator bus.

6.3.4

Volts/Hertz Protection (Function 24)

As explained in Chapter 4, core damage due to overfluxing is a rare event. However, when overfluxing occurs, the damage can occur within several minutes, possibly leading to partial or complete destruction of the core’s insulation. Therefore, it is important that V/Hz protection be applied and properly set. Consult your OEM for a V/Hz curve for the generator in question. When the generator is subjected to extreme levels of over-excitation, excess magnetic flux will saturate the core steel at the ends of the core and overflow into structural steel parts and into the air. Structural steel parts such as finger plates, which are not laminated and thus not designed to carry magnetic flux, will heat rapidly [6]. Flux flowing in unplanned air paths may link conducting loops in the windings, leads, or structural parts, and resulting circulating currents in these loops can cause dangerous temperature rise [7]. The damage due to excessive V/Hz is most likely to occur while the generating unit is being manually synchronized by operators solely relying on control panel indicators. If there is an open circuit in the PTs or blown fuses in the excitation system, which are not monitored by the loss of potential protection devices, operators are prone to increasing the field continuously without realizing the loss of voltage on the PTs. For V/Hz protection, two relay characteristics are typically used: definite time and inverse time relays. For definite time relay, dual level definite time curves are typically used. The disadvantage of this scheme is that there is unprotected area, since the generator V/Hz capability curve is of inverse nature. For this reason, inverse-time relays provide the optimal protection and operational flexibility because they coordinate better with the operational limits of the equipment [1]. The OEM will be able to provide a V/Hz curve for the generator and the protection can be set accordingly.

6.3.5 Reverse/Forward Power Protection (Functions 32R and 32F) The use of directional power or reverse power relays is not common on hydro units since their turbines are not normally damaged by sustained reverse power operation. That is one reason why synchronous condenser mode of operation is often employed [8].

300

CHAPTER 6

GENERATOR PROTECTION

In some cases, there may be mechanical issues associated with complete closure of the wicket gates and motoring of the unit. In these cases, the mechanical limitation must be clearly documented since operation of this relay element could separate units from over-generated isolated grids or prevent them from participating in restoration activities. If hydro generators are operated in electrical islands, normal governing operation could unload them following a load loss. Many hydro generators are designed for synchronous condenser operation, with or without tail-water depression to spin the turbine in air. Based on the above discussion, these elements should not be set to trip unless the turbine manufacturer or the local mechanical engineering department identifies a mechanical issue with sustained reverse power operation [8]. Low forward power elements may be used to supervise orderly shutdown based on mechanical trips. For example a shutdown from a mechanical problem (e.g. bearings, oil supply) will close the wicket gates to unload the unit. A low-forward-power element can then be used to complete the circuit for tripping the generator breaker, preventing a significant speed rise that could aggravate to damage the bearings [8].

6.3.6

Over/Undervoltage Protection (Functions 59 and 27)

One of the main causes of overvoltage is load rejection followed by generator overspeed. In this case, over excitation (V/Hz) is not excessive, but the sustained overvoltage may be above permissible limits. Speed control and voltage regulation of hydro generators may be relatively slow, especially for old machines. Overvoltage relays are also used as backup to function 24 during normal operation of the machine. The under voltage relays are mainly installed for the purpose of identifying loss of PT voltage or to identify dead bus conditions for certain alignments.

6.3.7

Loss of Field Protection (Function 40)

Loss of excitation can occur due to a number of reasons: the main exciter failure, accidental tripping of the field breaker, open and short circuits in the field winding, loss of AC supply to the excitation system, etc. The loss can cause severe system voltage depression (depending on the system and conditions at the time), as the generator can become a huge VAR drain. Consequently, the unit will over-speed, operating as an induction generator if water to the prime mover is not closed off. The high stator currents during this event will cause rapid overheating in the stator and rotor windings, resulting in rapid unit failure if the unit is not tripped immediately. In addition, if the generator which has sustained a loss of filed is not isolated, transmission lines can trip due to power swings or excessive reactive power flow to the faulty generator. The most widely used technique for the loss of field protection is the use of a single-phase distance relay, which is supplied with terminal voltages and currents.

6.3 BRIEF DESCRIPTION OF PROTECTIVE FUNCTIONS

–8

–6

–4

–2

0

2

4

6

301

8 R(sec Ω)

Offset = ½X′d 2

Diameter 1 = 1.0 PU

4

6

8

12

X(sec Ω)

10

Diameter 2 = Xd

14

16

Figure 6.3-3 Protection using two loss of excitation relays.

The relay measures the apparent impedance as viewed from the machine terminals and operates when the impedance falls inside the relay’s circular characteristic. The relay circle is offset from the origin of R–X plane by one half of the direct axis transient reactance, X d 2, to prevent misoperation during system disturbances and other fault conditions. Depending on the protection requirements, two relays are sometimes used as shown in Figure 6.3-3. The diameter of the first relay circle is adjusted to be equal to 1.0 PU. value of the generator, while the second one to the direct-axis synchronous reactance (Xd). Both relays are applied with a short time delay, which is to provide security against stable power swings that are recoverable.

6.3.8

Stator Unbalanced Current Protection (Function 46)

There are a number of system conditions that may result in unbalanced three-phase currents in a generator. For instance, they can come from unbalanced loads, single pole opening of a breaker, asymmetrical transmission systems, and open circuits, which will greatly increase the negative phase sequence currents. The system negative sequence currents can induce a double frequency current in the surface of the rotor which can cause high temperatures if left for an extended period of time [1].

302

CHAPTER 6

GENERATOR PROTECTION

Generators must meet minimal requirements for sustaining unbalanced currents without injury. These requirements are specified in IEEE Standards [9] for the generator continuous I2 capability and its short time capability in terms of I 22 t. The salient pole generator with amortisseur windings is capable of withstanding unbalanced currents (I2) up to 10% of the rated stator current, while the generator without those windings is less capable of doing so, typically 5%. The short time capability, I 22 t, of the generator is permissible up to 40. The protection against unbalanced currents was implemented in the past by using electromechanical overcurrent relays that measure negative sequence components. These relays provided basic protection due to its lack of sensitivity against unbalanced system faults. However, modern digital relays allow setting the protected region of operation in such a way that closely matches the withstand capability of the protected generator. This allows a more sensitive and discriminatory approach.

6.3.9

Voltage Balance Protection (Function 60)

The main function of the voltage balance relay is to avoid false tripping of other protection relays due to a loss of secondary voltage feed, for instance, by a blown voltage transformer (PT or VT) fuse. Voltage balance schemes are possible in most large generators because such units have at least two PTs feeding the protection and monitoring systems. The voltage-balance relay senses and compares the secondary voltage of different PTs, and when it determines that a “blown fuse” situation arises, it blocks the operation of certain voltage controlled relays and alarms. Figure 6.3-4 shows two PTs being monitored by a voltage balance relay.

GSU Generator G

PTs

RL

60

Figure 6.3-4 Example of a voltage balance relay circuit.

6.3 BRIEF DESCRIPTION OF PROTECTIVE FUNCTIONS

303

GSU Generator CT G

PT

RL

60 NS

Figure 6.3-5 Example of a generator with one PT feeding its protection and excitation system.

In those older machines in which only one PT feeds the protective and excitation systems, it is still possible to sense and alarm for a blown fuse condition. This is attained by using a scheme that compares negative sequence voltages in the secondary of the PT (which will arise as a consequence of a primary fault or a blown fuse condition), with negative currents in the secondary of the current transformer (CT). If negative sequence currents are not present, it indicates that a fault in the primary system did not occur, and thus it must be a blown fuse condition. This voltage/current negative sequence comparative function can be found in certain modern digital protective packages, as shown in Figure 6.3-5.

6.3.10

Breaker Failure Protection (Function 50BF)

Faults involving the generator require tripping of the generator breaker first, and then the next or alternate circuit breaker upstream. Failure of any such circuit breaker to operate properly results in loss of protection for the machine and other abnormal conditions, such as motoring, or continued fault conditions where the machine feeds a fault until another circuit breaker can clear the event. Adverse conditions also arise if only one or two poles of a circuit breaker operate, resulting in single-phase operation with the accompanying negative sequence currents as a consequence and time of the essence. Activation of a breaker failure scheme is carried out by a combination of triggering signals from the generator protective relays, overcurrent relays, and circuit breaker auxiliary switches, via a timer. Some modified schemes also include in their triggering circuit the trip signal from the neutral of the main step-up transformer’s overcurrent relay. This change is to protect against circuit breaker head flashover, which is when arcing occurs across the circuit breaker contacts due

304

CHAPTER 6

GENERATOR PROTECTION

62

BFI Generator Protective trip Breaker fail initiate

Timer

52a Breaker closed

50BF

BF timer

BFT Breaker failure trip

Current detector

Figure 6.3-6 Functional diagram of a generator zone breaker failure scheme. Source: From IEEE [1].

to high voltages in older style air magnetic breakers. This type of activity does not exist with the modern-day SF6 or vacuum style circuit breakers. The protection is designed to operate against the flashover of two poles. Figure 6.3-6 shows the functional diagram of a simple breaker failure protective scheme.

6.3.11

Rotor Ground Fault Protection (Function 64F)

Rotor field windings are designed to operate ungrounded. As a result, a single short to ground, in theory, should not be reason for concern, because it will not interfere with the normal operation of the machine. However, the appearance of a second ground can be very detrimental to the operation of the generator, as well as to its integrity. The second ground may result in the following: • Unbalanced airgap fluxes with increased rotor vibrations if sufficient poles are affected. • Unbalanced magnetic pull acting on the rotor pulling it toward the stator. • Rotor and stator collision if sufficient poles are affected. • Fluctuating VARs and output voltage. There are a number of methods in existence for the detection, alarming, and/or tripping of generators due to field ground faults. Some methods use a voltage source, and others use a passive unbalanced bridge. In addition to Ref. [1], another good reference is the chapter on protection from the Westinghouse manual on applied protective relaying [10].

6.3.12

Inadvertent Energization Protection (50/27)

In many instances, not enough attention is paid to the protection of the generator when the unit is offline. For example, protective devices that are crucial in avoiding unwanted and often damaging occurrences are left nonoperational. It is important that when the unit is offline, the protective relaying systems are kept operational. If work is being carried out on the relays, any risk should be ascertained and mitigating action taken.

6.3 BRIEF DESCRIPTION OF PROTECTIVE FUNCTIONS

X2G

X1T

305

X1S

EG

ET

ES

Figure 6.3-7 Approximate equivalent circuit.

One very onerous event is the inadvertent energization of the generator while offline and at rest. When a generator is energized without field while at rest, it behaves as an induction motor and can be damaged very quickly if the amortisseur circuit is unable to accommodate the high currents [11]. The maximum fault current a generator may experience during the inadvertent energization will be determined by the magnitude of the total impedances “seen” by the power system as shown in Figure 6.3-7 and in Equation (6.1) I=

X 1S

Es + X 1T + X 2G

(6.1)

where, X1S = System positive impedance X1T = Unit transformer positive impedance X2G = Generator negative impedance. Es = System voltage ET = Transformer HV voltage EG = Generator terminal voltage There are several ways that a generator can be protected against this occurrence. The loss of field (40) protection can provide protection against the inadvertent energization as long as the voltage source is available to the relay. While the unit is at rest, however, it is most likely that the voltage source is removed from service, disabling the relay to operate [11]. The fault currents during the inadvertent energization are within the pickup range of the reverse power (32) relay, though one problem with the reverse power protection is a long time delay associated with the protection. The negative sequence (46) protection, based on a time overcurrent relay, can respond to negative sequence present during the energization. As such, there is a severe limitation of conventional generator relaying to detect the inadvertent energization. Dedicated protection, which is solely designed for the detection of inadvertent energization, may be required. The details of such protection are beyond the scope of this book, but the schemes relying on directional overcurrent relays, frequency supervised overcurrent relays, distance relays, and system backup relays can be considered. Reference [11] provides a good discussion of the subject.

306

CHAPTER 6

6.3.13

GENERATOR PROTECTION

Out-of-Step Operation (Function 78)

Overload conditions, electrical faults, and system disturbances can cause the generator to lose synchronism with the system. If the generator goes out-of-step with the system, the resulting high peak currents and off-frequency operation may cause winding stresses, pulsating torques, and mechanical resonances that are potentially damaging to the generator. The generator must be separated from the system immediately, preferably during the first half slip cycle before it is damaged or before a widespread outage can occur. Protection against the out-of-step condition is based on the fact that the apparent impedance, as seen at the generator’s terminals, changes in a predicted manner during an unstable condition. This is similar to the loss-of-excitation condition. However, the loss-of-excitation relay will not pick up every occurrence of an out-of-step condition because the apparent impedance behavior is different for both conditions. Therefore, to fully protect against out-of-step conditions, a dedicated relay (or protective function within a multifunctional device) must be included in the protection package. Tripping the unit within the first slip cycle has major advantages in the case of an out-of-step event. This fast protective action tends to considerably reduce the very large oscillating shaft torque that can otherwise occur. For those readers interested in an elaborated discussion of this topic, Refs. [1, 12] provide a good starting point.

6.3.14 Over-/Under-Frequency Protection (Function 81O/U) Chapter 4 contains a discussion about the consequences to the integrity of the generator while operating for short time periods at frequencies higher or lower than nominal. Over- and under-frequency operation generally results from full or partial load rejection and overloading conditions, respectively. Load rejection can be caused by a fault in the system or load shedding. Overload conditions may arise from tripping a large generator or a transmission line. The frequency the machine will attain following load rejection or overload is a function of how much load has changed and the governor droop characteristics. For instance, a governor with a 5% droop characteristic will cause a 1.5% speed increase for a 30% load rejection (speed change in percentage equals droop in percentage times load change per unit). As discussed in Chapter 4, the manufacturers can provide V/Hz withstand curves that should be used in setting the function 81 relay.

6.3.15

Generator Differential Protection (Function 87)

Phase-phase and three-phase faults in a generator stator winding can destroy the machine in a very short time, due to their extremely high fault current magnitudes. Differential protection relays are usually implemented in order to provide high

Operating current

6.4 TRIPPING AND ALARMING METHODS

307

Slope 2

Slope 1

Minimum pickup id Restraint current

Figure 6.3-8 Generator percentage differential relay slope characteristic.

speed protection against such multiphase faults. There exist various differential relay schemes, among which the differential relay scheme based on the various slope percentages is most widely used. The preference of its use is due to its greater ability to discern an internal fault from an external one. The percentage slope characteristic can have multiple slopes, varying from about 5% at low through-fault current values to 50% or more at high through-fault current values as illustrated in Figure 6.3-8. This varying slope percentage scheme should provide extremely sensitive protection to internal faults, while it should become insensitive to external faults, which are usually accompanied by a much greater CT error currents.

6.4 TRIPPING AND ALARMING METHODS Earlier in this chapter, we discussed protective functions without reference to how an actual trip or alarm is implemented. There are many different schemes used to trip a unit as the need arises. There are also many different choices as to when to alarm, when to trip, when to alarm-and-trip, and so forth. The choices are unique as they depend on the country, age of the plant and equipment, and so forth. This book is not geared toward an exhaustive discussion of protection systems and philosophies. Therefore, only an impartial listing of some of the elements that go into the decision-making will be enumerated. Factors that could or should be taken into account when determining what type of trip should be implemented are as follows: • How onerous the fault can be to the generator • How onerous the fault can be to the power system • Over-speed risk post-trip • Fault spreading to other equipment

308

CHAPTER 6

GENERATOR PROTECTION

• Need for maintaining auxiliary loads post-trip • Need to trip the excitation Depending on the answers to those questions, the following tripping decisions may be implemented: • Generator Trip: The generator main breaker is tripped together with the excitation. • Circuit Breaker Trip: Trips the generator main breakers, but not excitation. This allows the unit to be reconnected to the system in a shorter time. Overspeed risks must be taken into account as the unit speed will rise during this trip time and then come back to normal rated speed relatively quickly. • Manual Trip: The generator is tripped manually by the operator. • Rundown and Trip: The operator unloads the generator manually or automatically to zero or near zero power and then trips the generator circuit breaker isolating the unit from the system. This strip is normally employed when manual tripping of the generator is needed. Table 6.4-1 is a sample of some of the more common protective actions that may be found for the various types of faults. These protective actions are not all inclusive and may not be what is used at any given generating station; they are merely what are considered common practices. TABLE 6.4-1 Example of faults and protective actions taken

Stator electrical faults Stator ground fault Stator phase to phase fault V/Hz Stator overvoltage Rotor overheating Rotor ground Loss of excitation Mechanical and thermal faults Stator core local overheating Bearing vibration excessive Synchronizing error Motoring Loss of SCW flow System faults Un-cleared system faults Unbalanced stator currents Loss of synchronism Abnormal frequency operation

Protective action take Generator trip, shutdown Generator trip, shutdown Generator trip, shutdown Voltage restoration by filed current runback Reduction in field current Alarm and if sustained generator trip Generator trip, shutdown Load reduction and if sustained generator trip Generator trip Synchro-check relay operation Generator trip Load reduction and if sustained manual or auto runback and generator trip Generator Generator Generator Generator

trip trip trip trip

6.6 FURTHER READING

309

6.5 REFERENCES 1. IEEE (2006). C37.102-2006: IEEE Guide for AC Generator Protection, IEEE. 2. IEEE (1996). C37.2-1996: IEEE Standard Electrical Power System Device Function Numbers, 1996-R2001, IEEE. 3. Berdy, J. (1975). Loss of excitation protection for modern synchronous generators. IEEE Transactions on Power Apparatus and Systems 94, 1457. 4. GEC (1975). Protective Relay Application Guide Book, GEC Measurements. 5. Baker, D. S. (1982). Generator backup overcurrent protection. IEEE Transactions on Industry Applications IA-18(6), 632–640. 6. Traxler-Samek, G. and Schmidt, E. (2002). Analytic determination of eddy current losses in the stator clamping plates of salient pole synchronous machines. Proceedings of the 15th, International Conference on Electrical Machines; 2002; Bruges, Belgium. 7. Alexander, G. W., Corbin, S. L., and McNutt, W. J. (1966). Influence of design and operating practices on excitation of generator step-up transformers. IEEE Transactions on Power Apparatus and Systems PAS-85, 901–909. 8. Roger Bérubé, G., Hiusser, P., Kim, S., and Mark Wilkinson, G. (2013). Draft: Generator Protection: Requirements for Coordination with Equipment Capability and Excitation Limiters in the Ontario Electricity Market, Hydro Engineering Division, Ontario Power Generation. 9. IEEE (2005). C50.12-2005: IEEE Standard for Salient-Pole 50 Hz and 60 Hz Synchronous Generators and Generator/Motors for Hydraulic Turbine Applications Rated 5 MVA and Above, IEEE. 10. Westinghouse Electric Corporation, Relay and Telecommunications Division (1982). Applied Protective Relaying, Chapter 6: Generator Protection. Westinghouse Electric Corporation, Relay and Telecommunications Division. 11. IEEE, PSRC (2011). IEEE Tutorial on the Protection of Synchronous Generators, 2, IEEE, Piscataway, NJ. 12. Baldwin, M. S. and Daugherty, R. H. (1978). Tripping of generators during asynchronous operating conditions. Presented at 40th Annual American Power Conference, Chicago (24–26 April 1978).

6.6 FURTHER READING Gibson, W. P. and Wagner, C. L. (1969). Application of loss-of-field relays on cross compound generators. Presented at PEA Relay Committee Meeting, Harrisburg, PA (24 October 1969). IEEE (1993). C37.101-1993: IEEE Guide for Generator Ground Protection, 1993R2000, IEEE. IEEE (1997). C50.14-1997: Requirements for Combustion Gas Turbine Driven CylindricalRotor for Synchronous Generators, IEEE. IEEE (2001). Std. 242-2001: IEEE Recommended Practice for Protection and Coordination of Industrial and Commercial Power Systems, IEEE. IEEE (2003). C37.106-2003: IEEE Guide for Abnormal Frequency Protection for Power Generating Plants, IEEE. IEEE (2005). Std 67-2005: IEEE Guide for Operation and Maintenance of Turbine Generators, IEEE.

310

CHAPTER 6

GENERATOR PROTECTION

IEEE (2014). IEEE Std. C50.13: IEEE Standard for Cylindrical-Rotor 50 and 60 Hz Synchronous Generators Rated 10 MVA and Above, IEEE. Lewis Blackburn, J. (1987). Protective Relaying, Principles and Applications, Marcel Dekker Inc. Power System Relaying Committee (1975). Loss-of-field relay operation during system disturbances. IEEE Transactions on Power Apparatus and Systems, September–October 1975. 94, 1464–1472.

CHAPTER

7

INSPECTION PRACTICES AND METHODOLOGY

7.1 SITE PREPARATION Site preparation is the first significant activity that should be carried out before inspecting a generator. Every inspection of a large machine scheduled or not, long or short requires a sensible effort toward site preparation. The goal is to minimize the risks of contaminating the machine with any foreign material or object, as well as to ensure a safe environment in which to perform the inspection. Site preparation should be planned ahead of time, and it should be maintained from the moment the generator is opened for inspection until it is reassembled and readied for operation. Neglecting to prepare and maintain a proper working environment in and around the generator has the potential for resulting in undue risks to personnel safety and machine integrity.

7.1.1

Foreign Material Exclusion

Foreign material exclusion (FME), a term that originated in the nuclear industry, is the set of procedures geared to minimize the possibility of intrusion into the machine of foreign material before, during, and after the inspection. In principle, the definition of foreign material is anything not normally present during the operation of the generator that may adversely affect its constituent components if left there. For instance, although ambient air is not necessarily considered a foreign material, the water content of the air is. Water definitely is an extraneous element that should be kept from condensing on the machine’s windings and contributing to electric breakdown of the insulation during testing or in service. Most generators are provided with heaters to keep the generator

Handbook of Large Hydro Generators: Operation and Maintenance, First Edition. Glenn Mottershead, Stefano Bomben, Isidor Kerszenbaum, and Geoff Klempner. © 2021 The Institute of Electrical and Electronics Engineers, Inc. Published 2021 by John Wiley & Sons, Inc.

311

312

CHAPTER 7

INSPECTION PRACTICES AND METHODOLOGY

temperature above the surrounding ambient to avoid condensation. Particular attention has to be paid to this feature when extensive shutdown is planned. Heaters are usually turned off and locked out during crawl through inspections with the doors and covers open therefore some moisture contamination can occur. Before return to service it may be necessary to dry out the generator to obtain acceptable electrical test readings prior to performing high-voltage testing or returning these machines to service. For appropriate dry-out procedures, consult with the equipment manufacturer. It is also important not to inadvertently contaminate the generator with corrosive liquids such as solvents, cleaners, or with oils, and so forth. The worst threat to the windings are any foreign metallic objects. They can become airborne due to the high speed of the cooling air or become dislodged from the rotor and pierce the insulation when striking it. Magnetic particles have been known to cause failures in generator stator windings by penetrating the insulation over long periods of time, due to the electromagnetic forces acting on the particle. It is unfortunate when magnetic particles (also called magnetic “termites” or “worms”) manifest themselves inside the generator, resulting in damage to the windings. It is truly shameful when these particles are introduced by negligent practices. It is quite common during an overhaul to grind and weld near the open generator and have contaminants settle onto the winding. Magnetic, as well as nonmagnetic metallic objects may be subject to eddy current heating depending where they are located in the generator, detrimentally affecting the insulation with which they come into contact. Foreign metallic objects such as nails, welding rods, or scrap steel used in the construction phase inadvertently left in the bore, can short circuit the laminations of the core and cause eddy currents to flow as shown in Figure 7.1-1; subsequent damage can result, as illustrated in Figure 7.1-2. Continued operation under this condition may result in a winding failure due to localized temperature rise of the core and/or fusing of the laminations. Precautions should be taken to eliminate the possibility of metallic parts or other foreign objects entering the machine. Metallic objects not required for the examination, or for performing work inside the machine, should be left outside the stator bore. This entails removing any coins and other objects, such as medallions, cell phones, unnecessary pens, pencils, from pockets, prior to entering the machine. Inspection tools should be carried into the machine on an “as needed” basis. When using mirrors or flashlights in otherwise inaccessible areas, these should be secured by strapping them to one’s wrist or wearing them as a necklace. Of course, the user should use caution whenever an object is tied onto the body. This object can get caught in various areas of the machine and can cause injury to the person performing the inspection if pulled out inadvertently while moving through the machine. Wherever possible, invest in a magnetic-based inspection tool so it can be placed anywhere on the generator and it will not move. However, in such a case, it is crucial to keep track of such a tool, so it is removed from the machine at the end of the inspection or maintenance activity. Taking an inventory of all tools is strongly recommended both before entering the

7.1 SITE PREPARATION

313

Stator winding Keybar

Eddy current

Metallic object

Temperature rise of laminations on packets and/or fusing

Figure 7.1-1 Shows the effects of a metallic object on the surface of a stator core.

Figure 7.1-2 Foreign material caused serious core damage.

machine and after exiting it. This is a time-consuming practice but recommended for all large generator inspections. A lack of adequate access control can cause serious damage to the stator or rotor of a large generator, in particular the stator core as shown in Figure 7.1-2. This type of occurrence is, unfortunately, not as rare as it should be. Thus, it is critical to account for all tools and parts coming in and out of a generator during an outage.

314

CHAPTER 7

7.1.2

INSPECTION PRACTICES AND METHODOLOGY

Foreign Material Exclusion Procedures

A good FME effort must include clearly written procedures addressing all aspects of training and implementation. Responsibilities must be clearly defined, as well as contingencies for loss of FME control. A typical FME procedure may include the following subjects: Responsibilities of: • Planner • Supervisor • FME monitor • FME qualified worker • Non-FME qualified worker • Escorts Issues addressed • Definitions • Controls of FME zones • FME zone designation • FME signs • FME covers • Appropriate clothing • Tool controls • Spare parts controls • Recovery of lost FME control • Final FME check Training • As required by the level of responsibility and the issues for each level described above

7.2 EXPERIENCE AND TRAINING Inspection of generators is not a trade that can be solely learned in the classroom. A combination of classroom learning, reading relevant literature, with years of hands on training, mainly while accompanying an experienced specialist, is what leads to the ability to decipher the root causes of conditions afflicting a generator. Misjudging root causes, or simply not understanding the condition of the machine, can lead to significant monetary losses. Therefore, it is always a good idea to have someone well experienced leading the inspection. When large units are inspected, repaired, or overhauled, most owners will arrange for a manufacturer’s specialist to be present, at least through part of the effort.

7.2 EXPERIENCE AND TRAINING

7.2.1

315

Safety Procedures: Electrical Clearances

When carrying out work in an industrial environment, nothing is more important than adhering to all required safety precautions. Machines opened for inspection often present obstacles in the form of big openings in the floor surface, crevices to crawl thorough, rods and protruding machine members, and so on. They all demand evaluation of required temporary additions to the site, such as beams over the open floor spaces, handrails, secure ladders, and so forth. The obstacles just mentioned are all visible to the people engaged in the inspection. However, an invisible and very powerful element to contend with is the voltage potential (or range of voltages) that may be present in a generator. Although rare, electrical accidents can occur when work is performed on these machines, therefore, ensuring the appropriate work protection is in place and becoming a member of the work group for that inspection and isolation is a must. “Walking the clearance or permit” is a jargon used by some to describe the process of inspecting all breakers, cables, switches, and connections that may be sources of electric power to any part of the machine, and making sure they are all de-energized and secured with locks and a lockout procedure. This means that none of these will be accidentally energized during the inspection. Further, becoming familiar with the “guarantee of nonrotation” part of the “clearance or permit” is critical in ensuring that the rotor cannot turn during the inspection. This should be discussed and verified before entering the safe work zone. A safe work zone can be considered as a zone where no kinetic, potential, or electrical energy can influence the zone thus protecting the worker within that zone. The safe work zone is defined by a physical barrier, such as tape or barricades that have signage clearly outlining the area as a safe work zone. The signage should also include the name of the person or persons to contact to get permission to enter the safe work zone so that a record of who is in the zone is established and can be managed. Each person that is inside this zone should be signed onto a log sheet to maintain real-time records of who is coming out or going in. This minimizes the risk of personnel already working inside the zone and those wishing to enter or exit. The preceding description of “safe work zone” does not replace existing procedures or definitions at the location of work that already exist at the generation facility. It is merely a process that can be considered to protect everyone involved. The following is a typical (but by no means all-inclusive) list of safety procedures: • Temporary grounding cables at both ends of the winding of each phase will minimize the possibility of receiving an unexpected electric discharge of dangerous magnitude as shown in Figure 7.2-1. These grounds are normally installed by the electrical crew before any work begins and is part of the overall electrical protection in the safe work zone. The portable grounds conductor’s cross section, ground clamps, and ground studs must be sized to withstand the available short circuit currents for the expected duration of the fault, given the protection settings on site, so

316

CHAPTER 7

INSPECTION PRACTICES AND METHODOLOGY

Figure 7.2-1 Ground leads applied to the generator bus at the switchgear.

the inspector(s) should verify with the clearance or permit holder that these are the correct temporary grounds. Ensure the connections of the temporary grounds are tight and secure, and that the ground cable itself is tied off sufficiently to prevent whipping or severing against a sharp object should they be required to interrupt a fault. The amount of current flowing through the temporary ground will cause it to whip violently as it awaits the protection to operate and clear the fault. One of the authors has personally witnessed shortcircuit testing on temporary grounds at the 70 kA level, and it is an eyeopener to witness the amount of whipping the conductor is subject to. As the conductor is whipping violently, it works the clamp and stud trying to separate them. If the conductor is not tied off securely, separation can also occur between the clamp and conductor rendering the temporary ground useless. This is why tying off the conductor and restraining it is crucial to the temporary ground staying in one piece. • Phase leads must be open. • Neutral transformer (if present) must be disconnected or have its leads opened. • Voltage regulators and other excitation equipment must be disconnected. • Potential transformers are an additional source of voltage to the main windings and, therefore, they must be disconnected and secured. Space heaters are often overlooked. To keep moisture out, space heaters, when present, are

7.3 INSPECTION FREQUENCY

317

normally left “on” after disassembling the generator; thus, it is important to check that they are disconnected during the inspection. • All switches that may energize any part of the machine must be clearly tagged and if equipped for a lock, a lock(s) applied. Only the person who installed a tag can remove it, after verifying that the inspector(s) have left the machine. • Ensure that no electrical testing is going on while the inspection is taking place that will affect your inspection area. • Inquire if any other mechanical work is being done that involves the use of solvents, welding as the fumes from the solvents or welding may cause discomfort or nausea during inspection. • Ensure no overhead craning is ongoing while on top of the generator being mindful of overhead loads and the danger they present. • Any additional items as each specific case warrants. • Finally, it is required that the person performing the inspection should walk the clearance or permit. Do not take anyone else’s word for it; mistakes can happen anywhere at any time.

7.3 INSPECTION FREQUENCY Certain components in hydro machines require quick inspections (and sometimes maintenance) between scheduled major outages. Other more comprehensive inspections requiring various degrees of machine disassembly are performed during the more lengthy outages. However, experience shows that a major inspection after one year of operation is highly recommended for new machines. During the initial period, winding support hardware and other components experience a full spectrum of operating stresses. Verification of core compression is a good example of what is normally required during this first outage. Also, expiration of warranty may be of consideration; hence, a careful inspection is very important. Subsequent outages and inspections can be performed at longer planned intervals. How long an interval? Minor outages/inspections every 30 months, to major outages/inspections every 60 months, are typical periods recommended by machine manufacturers. The major outages include removal of the rotor, comprehensive electrical and mechanical (nondestructive) tests, and visual inspections. Of course, these intervals tend to be longer for machines spending long periods without operation. Most machine operators have electronic logs recording the actual number of hours the unit was running and the number of starts/stops. Together with the manufacturer’s recommendations, this information forms the basis for scheduling inspections and overhauls. Of particular interest are the number of start/stops, run time and average winding temperatures. Higher levels of this wear and tear should drive more frequent outages since the machine is may

318

CHAPTER 7

INSPECTION PRACTICES AND METHODOLOGY

be subjected to more mechanical and thermal stress. In today’s demanding electricity market and lucrative supply contracts, it is not uncommon for a machine to be cycled on and off anywhere from 1 to 6 or more times per day, with various cool down periods in between. If the machine were to run every day that would amount to 365–2190 start/stops per year. Large utilities that have many hydro generators in their systems and many years of experience running these machines have formed their own maintenance and inspection criteria and schedules. Although working closely with machine manufacturers, these utilities tend to extend the periods between outages for those machines that experience has shown to have good records of operation, and shorten the periods between outages/inspections for machines that have been characterized by more frequent failures.

7.4 GENERATOR ACCESSIBILITY The issue of accessibility is the same for inspection and for maintenance, as both activities often follow each other. The following list summarizes the various levels of access to the generator and what you can expect to see from each. The level of access required is commensurate with the purpose of a specific inspection activity, and it will vary somewhat for different machine constructions. Generator top covers/shrouds removed Stator endwinding access Stator core and core-clamping mechanism Slot wedges (few near top end of core) Airgap Inter-polar space Rotor poles and pole interconnections (depends on fan and/or shroud design for the rotor) Amortisseur interconnections on field poles if so equipped Generator bottom shrouds removed Similar visual access as the top covers/shrouds being removed Rotor removed Access to stator bore Full rotor inspection possible including bearings and brackets Stator surface air cooler removal if equipped Access to back of stator core in localized areas or ability if possible to crawl around back of machine Surface air cooler inspection and cleaning possible Stator frame access panels if equipped

7.5 INSPECTION TOOLS

319

Access to back of stator core in localized areas Terminal enclosures opened if so equipped Access to stator terminals (main and neutral)

7.5 INSPECTION TOOLS Probably the most important item for the inspector is not a gadget or an instrument, but it should be reports of previous visual inspections and electrical tests. Findings from past inspections act like a compass, helping to guide the inspector to areas already proved to be problematic. Most operators of large generators carry out a minimum set of electrical tests on the machines before disassembling them for a visual inspection. Results from these inspections have the advantage of, first, calling attention to problematic areas and, second, allowing comparison with the test results obtained after cleanup or refurbishment performed on the machine. It goes without saying that a comprehensive inspection report should always be written and archived. This report is a helpful reference for the next inspection, often several years later. An additional source of information that has served the authors of this book well is historical information obtained from identical or similar machines. Still other significant sources of information will be the manufacturers of the generator or many kinds hydro generator user groups. As for the “bag” of inspection tools carried by the inspector, it may include the following (partial bag of tools shown in Figure 7.5-1): • Writing pad or preprinted check-sheet attached to a clipboard, and a nonmetallic pen that can be clipped to the board or perhaps with a string. • Thick black permanent marker for labeling areas of interest. • Safety glasses.

Figure 7.5-1 Example set of inspection tools.

320

CHAPTER 7

INSPECTION PRACTICES AND METHODOLOGY

• Comfortable pair of gloves with good dexterity. • Disposable paper or cloth boot covers. • Disposable paper or cloth overall. • Safety shoes with soft rubber sole. • Magnetic-based flashlight so it can be set anywhere in the machine without losing it and a set of spare batteries • Clean paper towel roll or rags to be used as swabs; useful for taking samples of contamination. • Clear sandwich bags to store sample swabs for testing. • A mirror or set of mirrors with articulated joints and expandable handles. If the mirrors are at risk of falling to inaccessible places, they should be attached to the wrist with a string and tape. • An expandable handle magnet (with light if available) for lifting contamination that is magnetic and to test for fretting dust in hard to reach places. • A hammer with soft (rubber) and hard (plastic) heads, for probing wedges, insulation blocking, and so on. A small ball-peen hammer may also be used to survey wedges. • A set of magnifying glasses or handheld microscopes to probe for corrosion or electrically originated pitting on bearing surfaces and on other critical components. • Charts from manufacturers of commutator brushes depicting observable signs of bad commutation. • A Borescope with an articulating head especially suitable for inspecting air ducts, the airgap, and other inaccessible spots as shown in Figure 7.5-2.

Figure 7.5-2 Typical borescope with articulating head.

7.6 INSPECTION FORMS

321

Figure 7.5-3 Pocket knife used to spot check core tightness.

• Pocket knife (in lieu of a knife with a 1.397 mm (0.055 ) beveled blade if one is not available to assess core tightness as shown in Figure 7.5-3. • A set of feeler gauges to check clearances in various areas and between various components. • A good fully charged camera with a macro function for close-up shots. • A small or medium size pencil case to carry the smaller tools. • A compass to pinpoint magnetized items.

7.6 INSPECTION FORMS This section includes a total of 10 generator inspection and test report sheets. A similar set of forms good for any type of synchronous machine can be found in Ref. [2]. The inspection forms included in this chapter cover full inspections and routine inspections. They are examples of forms that can be used practically for any hydro generator, of any size. The forms are by nature generic. However, it should take very little effort to recast them in a way that can be quickly adapted to the needs of any machine operator and/or inspector. Forms 7.1 to 7.8 are designed for a full or partial inspection of a machine in various stages of disassembly. Form 7.9 is for testing when the machine is shut down for maintenance. Form 7.10 can be useful in tracking the wear of brushes in a generator. Given that each brush will rarely have to be changed more than once a month, using a form for an entire month should suffice. This way, 12 forms, one per month should be adequate to document brush changes for each brush location and date of replacement for the entire year. Completing this form entails a very small effort, but the rewards can be significant when searching for specific wear and commutation problems.

322

CHAPTER 7

INSPECTION PRACTICES AND METHODOLOGY

HYDRO GENERATOR INSPECTION AND TEST REPORT FORM 7.1: BASIC INFORMATION Company: Station: Unit no.: Manufacturer Frame type or size: Date of manufacture: Year installed: Date of last rewind: Stator

Rotor

Date of last major inspection: Operating years since last overhaul: Total number of starts/stops since last overhaul: Inspection performed by: Date of inspection:

HYDRO GENERATOR INSPECTION AND TEST REPORT FORM 7.2: MACHINE INFORMATION Rated MVA:

Power factor:

Short circuit ratio:

Runaway speed:

Rated voltage:

Rated current:

Rated field voltage:

Rated field current:

Rated speed:

No. of poles:

Number of coil turns per field pole:

Number of coil turns per stator coil/bar:

Number of slots:

Frequency:

Stator cooling: Open air or Air/water or Direct water

Stator cooling water operating pressure (direct cooled units):

Stator insulation: Asphalt, Polyester, Epoxy

Serial no.:

7.6 INSPECTION FORMS

323

HYDRO GENERATOR INSPECTION AND TEST REPORT FORM 7.3: MACHINE ACCESSIBILITY Item

Yes

No

Upper bracket covers removed Surface air coolers removed Rotor removed Stator scaffold in place Stator frame cover plates removed Stator winding shrouds removed top and bottom Rotor air seal removed top and bottom

HYDRO GENERATOR INSPECTION AND TEST REPORT FORM 7.4: STATOR INSPECTION ITEMS Type of endwinding blocking: Maple _____ Epoxy/Polyester glass Felt _____ Ties: Glass roving _____ Rope/String/Cord _____ Other _____ Wedges: Ripple spring _____ Piggyback _____ Retainer _____ Other _____ Side fillers: Flat Ripple spring _____CRTV Ripple _____CRTV/RTV flat _____ Wrapper _____ Top bar/coil ripple filler under wedge: Yes _____ No _____ Borescopic inspection of the air duct area performed: Yes ____ No _____ If wedge survey performed: % of loose wedges _____ Description

Stator frame – Soleplates 1

Frame able to move freely on soleplates or locked down so frame cannot move

2

Dowel or key position (not twisted or backed out)

3

Grouting and foundation condition – cracking, spawling, etc. Stator frame – General

4

Broken or cracked welds on frame members

5

Stator frame splits, alignment, fretting, movement

6

Stator frame shelves, fretting, broken, or cracked welds

7

All stator hold down bolts set to the correct torque

Comments

8

Cleanliness

9

Stator frame – overall condition Stator core – Air ducts

10

I-beam/vent/duct spacers ability to support laminations, is there migration

11

Obstructions (dirt, grease, grime) preventing air flow leading to high temperatures on winding Stator core – Laminations

12

Broken teeth or breaks at keybars

13

Buckling or wave

14

Movement of laminations (into airgap or winding)

15

Looseness using pocket knife

16

Core to keybar fretting

17

Alignment – vertical at core split

18

Smearing

19

Signs of overheating on core ends or main core

20

Oil and dust from brakes or other dust like debris

21

Fretting at core splits

22

Chevroning at core splits

23

Evidence of arcing, fusing, or fretting on core packets

24

Flux test required? Stator core – Clamping system

25

Finger condition

26

Core clamping/stud bolt tension

27

Core stud or key bar welds on frame

28

Free hanging core stud condition (not welded to frame)

29

Circularity and concentricity as per CEATI standards Stator coils/Bars

30

Carbonized tracking paths

31

Cleanliness

32

Evidence of abrasion or impact damage due to foreign material

33

Evidence of girth cracking (bitumen windings) at location where coil/bar leaves the slot

34

Semi conducting/grading tape or paint interface (white powder)

35

Coil/bar puffiness (bitumen winding)

36

Dripping bitumen into generator pit

37

Evidence of corona activity in endwinding components

38

Temperature distribution throughout the winding

39

Water boxes – connections leaking on water or winding side

40

Flow restriction in water cooled stator windings

41

Hoses, fittings, and gaskets in water cooled stator windings

42

Insulation Resistance (IR) testing

43

Polarization Index (PI) testing

44

Thermocouple/RTD operation check

45

Partial Discharge (PD) testing

46

High potential test per IEEE Std 95

47

Bypassed coils Stator wedging system

48

Looseness – touch with one finger while tapping with the other

49

Unacceptable conditions – more than 25% of wedges are loose in a slot, or too or bottom wedge is loose

50

End wedges – moving out of slot/Blocking of ventilation ducts

51

Slot packing filler migration

52

Greasing or powder along the wedges groove

53

Mechanical damage Stator endwinding

54

Evidence of partial discharge activity between coil/bar endwinding (white powder)

55

Vibration damage, loose bracing

56

Vibration probe condition

57

Bull ring (surge ring) top and bottom and supports – good condition

58

Circuit rings – bracing, blocking, lashing tight

59

Connections – jumpers, series, epoxy filled cap condition – overheating Main and neutral end leads, cables, VTs, CTs, and insulators

60

Discoloration due to overheating

61

Looseness due to vibration

62

Tracking

63

Connections tight

64

Continuity

65

Signs of partial discharge

326

CHAPTER 7

INSPECTION PRACTICES AND METHODOLOGY

HYDRO GENERATOR INSPECTION AND TEST REPORT FORM 7.5: ROTOR INSPECTION ITEMS

Description

Rotor–Spider with shrunk laminated rims 1

Spider arm rim support shelf fretting or cracking – NDE required

2

Spider (Drum style) support shelf fretting – NDC required

3

Welds – Visual and NDE, cracks at welds at start point on material or along length

4

Fretting at shaft coupling Rotor–Rim

5

Circularity, concentricity and verticality as per OEM specifications for in service (poles on or off rotor)

6

Rim-keys – fretting, cracked welds

7

Ventilation duct obstructions

8

Balance weights mounting – weld or mounting condition

9

Fan blades Rotor–Poles

10

Is airgap uniformity safe for operation?

11

Axial elevation with respect to the stator – within OEM specifications

12

Physical damage to pole bodies

13

Pole face burning

14

Amortisseur condition

15

Pole collar condition

16

Turn insulation migration

17

Interpole wedging (blocking) condition

18

Interpole connections – evidence of looseness, cracking, or heating

19

Pole-keys – fretting, movement, looseness

20

Pole fixation/attachment cracking

21

DC leads condition from sliprings to field connections, bolted joints, support blocks, heating

Comments

22

Field winding copper resistance test

23

Micro-ohm check for connections

24

Insulation Resistance (IR) test

25

Polarization Index (PI) test (for units that have encapsulated field coils only)

26

Impedance test Rotor–Brakes

27

Non-uniformly worn shoes

28

Adjacent brake plates are within 0.010 (0.25 mm) of each other

29

Flatness over entire track is within 0.250 (6.4 mm)

30

NDE for cracks on brake track segments

31

Blue spots with hard oxide film

32

Warping of track, broken, or loose bolts

33

Broken brake cylinder return springs

34

Stuck brakes alarms – brake micro switches

35

Operate brakes with unit shut down and check for entire system air leakage

HYDRO GENERATOR INSPECTION AND TEST REPORT FORM 7.6: EXCITATION SYSTEM INSPECTION

Description

Excitation – Field breaker 1

Contact condition

2

Cable or bus condition

3

Shunt discoloration

4

Discharge resistor

5

Insulation resistance of cable and field breaker

6

Timing Static exciter

7

Thyristor bridge condition

8

Cooling fans operational

Comments

328 9

CHAPTER 7

INSPECTION PRACTICES AND METHODOLOGY

Brushless exciter

10

Cleanliness

11

Diode or thyristor condition

12

Connections tight

13

Cable/bus condition

14

Insulation condition Static exciter transformer

15

Cleanliness

16

Insulation resistance and ratio tests Excitation – Rotating exciters

17

Cleanliness armature and stator

18

Insulation resistance readings – armature, stator, and interpole

19

Winding condition – armature, stator, and interpole

20

Wedging, mounting, banding

21

Air gap clearance

22

Housing condition, cleanliness

23

Brush neutral (inductive kick test) if exciter has been removed and replaced

24

Excitation – sliprings, commutator, and brushes

25

Cleanliness, brush replacement frequency

26

Commutator shows evidence of streaking, threading, grooving, bar edge burning bar face burning, high mica

27

Sliprings show evidence of flat spots, burning, general wear

28

Sliprings insulation condition

29

Sliprings polarity reversed occasionally

30

Sliprings runout