Compression Machinery for Oil and Gas 978-0-12-814683-5

Table of contents :
Cover......Page 1
Compression
Machinery for Oil
and Gas
......Page 3
Copyright......Page 4
Contributors......Page 5
The Editors......Page 7
Preface......Page 9
Acknowledgments......Page 12
Section I: Fundamentals of Compression
......Page 13
Dynamic Compressors......Page 14
Basic Thermodynamics......Page 15
First Law......Page 16
Intercooling......Page 18
Equations of State......Page 19
Basic Machinery Dynamics......Page 20
References......Page 22
Types of Compression Equipment......Page 23
Upstream......Page 25
Downstream......Page 26
Compression Ratio and Maximum Temperatures......Page 28
Gas Type (Sour Gas, Wet Gas, etc.)......Page 29
Emissions, Noise, and Safety......Page 30
Operating and Maintenance Costs......Page 32
Service Life......Page 34
Automated/Manual Controls......Page 35
Packaging Issues......Page 36
References......Page 37
Section II: Types of Equipment
......Page 38
General Requirements for Range, Pressure Ratio, and Performance for Inline Centrifugal Compressors in Various Applications?......Page 39
Elements of an Inline Centrifugal Compressor......Page 40
Horizontal Split......Page 41
Hermetically Sealed......Page 42
Multisection......Page 43
Impeller......Page 45
Diffuser and Return Channel......Page 46
Volute......Page 47
Balance Piston/Division Wall Seal......Page 48
Antiswirl Mechanisms......Page 49
Shaft End......Page 50
Static Seals......Page 53
Shaft......Page 54
Bearings......Page 55
Euler's Equations......Page 58
Operating Regimes of a Centrifugal Compressor......Page 63
Surge......Page 64
Stall......Page 65
Compressor Control......Page 66
Variable Speed......Page 67
Process Control With Centrifugal Compressors Driven by Two-Shaft Gas Turbines......Page 69
Multiple Units......Page 70
Pressure Vessel......Page 71
Lateral Rotordynamics......Page 72
Rotordynamic Modeling and Analysis......Page 73
Torsional Rotordynamics......Page 80
Blade and Disk Vibrations......Page 83
Deflection......Page 86
Hot/Cold Applications, Centerline Mounting......Page 87
Piping/Flange loads......Page 88
Thrust Loads......Page 89
Foundation......Page 96
Dry Gas......Page 98
Wet......Page 104
Liquid Film Seal Oil Supply System......Page 107
Bearing/Lube Oil......Page 108
Interaction With Reciprocating Compressors and Aerodynamic Stability......Page 111
Erosion/Fouling/Plugging of Components and Mitigation Strategies......Page 112
Parallel/Serial......Page 115
Typical Compressor Types and Their Market Space......Page 116
Gas Transmission Compressor......Page 121
Injection......Page 122
LNG......Page 125
Hydrogen Recycle......Page 127
Air/Nitrogen......Page 128
Design for Maintenance......Page 129
Predictive Maintenance......Page 131
Preventive Maintenance......Page 132
Sparing and Availability......Page 133
Principles of Gas Compressor Restage......Page 134
Restage Applications......Page 135
Running Slow......Page 136
Restage Criteria......Page 137
References......Page 138
Further Reading......Page 141
Introduction......Page 142
Advantages......Page 143
Design Speed......Page 144
Intercooling......Page 145
Variable Geometry......Page 146
Size......Page 147
Sealing......Page 148
Design Process......Page 149
Drivers......Page 151
Aerodynamics......Page 152
Flow Control......Page 154
Gears......Page 155
Thrust Management......Page 157
Bearings......Page 158
Lubrication......Page 160
Rotordynamics......Page 161
Casings......Page 163
Mechanical Design......Page 165
Operation......Page 167
Midstream......Page 168
Petrochemical......Page 169
Summary......Page 170
References......Page 171
Further Reading......Page 172
Compressors Without a Crosshead......Page 173
Other Types of Reciprocating Compressors......Page 174
Advantages and Disadvantages Compared to Other Compressors......Page 175
Average Linear Piston Speed......Page 176
Effect of Torque Pulsations......Page 177
Induction Motors......Page 178
Synchronous Motors......Page 179
Torque Pulsation......Page 181
Engines......Page 182
Turbines and Gearboxes......Page 183
Couplings......Page 184
Importance of Pulsation Studies......Page 185
Vibration Concerns......Page 186
Key Guidance From Standards......Page 188
Cylinder Swept Volume Process Overview......Page 189
Types......Page 191
Dynamics......Page 192
System Model......Page 195
Example Calculations......Page 196
Baseline Results......Page 197
Increased Gas Density......Page 199
Changing Piston Velocity......Page 201
Capacity......Page 204
Volumetric Efficiency......Page 206
Power......Page 207
Friction......Page 214
Pulsation Bottle Losses......Page 215
Cylinder Cooling Effects......Page 216
Recycle......Page 217
Head-End Fixed Volume Clearance Pocket......Page 218
Valve Clearance Spacer......Page 220
Finger-Type Suction Valve Unloader......Page 221
Head-End Bypass......Page 223
Pulsation Control......Page 225
Bottle Design......Page 226
Cylinder Orifice Plates......Page 228
Nozzle Pulsation......Page 229
Torsional Rotordynamics......Page 230
Torsional Mass Elastic Models......Page 232
Torsional Stiffening Effect of Induction Motor Webs......Page 235
Steady-State Analysis......Page 239
Allowable Stress Methodology for Torsional Systems......Page 241
Torsional Damping......Page 244
Transient Torsional Analysis......Page 245
Resolution of Typical Torsional Problems......Page 249
Testing Methods for Torsional Systems......Page 251
Reciprocating Compressor Lateral Rotordynamics......Page 255
References......Page 257
Working Principle of Screw Compressors......Page 259
Undercompression......Page 262
Overcompression......Page 263
Comparison of Positive Displacement Machines (Screw Compressor, Reciprocating Compressor) Versus Centrifugal Compressors......Page 264
Dry Screw Compressors......Page 266
Oil-Flooded Screw Compressors......Page 267
Design Features......Page 271
Rotor Design......Page 272
Shaft Ends and Coupling......Page 273
Two-Stage Casing Arrangement......Page 274
Shaft Seals for Dry Screw Compressors......Page 275
Driveshaft Seals for Oil-Flooded Screw Compressors......Page 282
Suction Flow and Power Consumption Versus Compressor Speed......Page 283
Actual Suction Volume Flow Versus Pressure Ratio......Page 285
Power Consumption Versus Discharge Pressure......Page 286
Actual Suction Volume Flow and Power Consumption Versus Molecular Weight for Dry Screw Compressors......Page 288
Actual Suction Volume Flow and Power Consumption Versus Molecular Weight for Oil-Flooded Screw Compressors......Page 289
Power Consumption Versus vi for Oil-Flooded Screw Compressors......Page 290
Starting of Dry Screw Compressors......Page 291
Stopping of Dry Screw Compressors and Settle-Out Pressure......Page 292
Stopping of Flooded Screw Compressors and SOP......Page 293
Injection Above Saturation Temperature......Page 294
Liquid Injection Flows for Dry and Oil-Injected Screw Compressors/Liquid Hammer......Page 295
Design Range for Dry Screws......Page 296
Design Range for Oil-Flooded Screws......Page 307
Lateral Rotordynamics......Page 308
Pulsation and Vibration......Page 310
References......Page 538
Overview of Gas Turbine Components......Page 314
Performance......Page 318
Reciprocating Engines......Page 319
Durability......Page 322
Basic Principles......Page 324
Rotor Circuit Resistance and Motor Torque......Page 329
Reciprocating Compressors......Page 331
Centrifugal Compressors......Page 336
Synchronous Motors......Page 337
Why Use Synchronous Motors?......Page 338
Natural Frequency......Page 339
Induction motor......Page 341
Lateral Pulsation......Page 342
Variable Frequency Drives......Page 343
Steam Turbines......Page 348
Types......Page 349
Application as a Compressor Driver......Page 353
Major components......Page 355
Selection......Page 361
Performance......Page 363
Operation......Page 364
Steam Turbine Maintenance/Reliability......Page 366
Performance Degradation......Page 369
Expanders in Cryogenic Applications......Page 370
Hot Gas Expanders......Page 374
Thermodynamics of Gas Expanders......Page 375
References......Page 377
Section III: Applications
......Page 378
Introduction......Page 379
Associated Gas (Flash Gas) Compression......Page 380
Nonassociated Gas Compression......Page 381
Gas Lift......Page 383
Reinjection......Page 384
Gas Processing Plant......Page 385
Corrosion......Page 387
Changing Process Conditions......Page 388
References......Page 389
Midstream Activities......Page 390
Pipeline Design......Page 392
Station Piping Design......Page 395
Piping Supports, Thermal Stress, and Vibration Control......Page 396
Foot Print Size/Weight/Unbalanced Forces/Foundations......Page 397
Metering, Pressure Regulation, and Custody Transfer......Page 400
Storage Reservoirs......Page 401
Local and National Pipeline Regulations......Page 402
Further Reading......Page 403
References......Page 427
Refinery Applications......Page 404
LNG Production......Page 405
Compressor Configurations and Compressor Technology Milestones......Page 406
Compressor Technology Milestones......Page 407
Horizontally Split and Barrel Designs......Page 409
Back-to-Back LNG Compressor Designs......Page 410
Centrifugal Compressor Design and Performance......Page 412
Two-dimensional (2D) and 3D Impellers......Page 413
Machine Mach Number and Inlet Relative Mach Number......Page 414
Impeller Head per Stage......Page 415
Compressor Design Trade-offs and Compromises......Page 416
Aerodynamic Mismatching of Stages......Page 417
Mechanical Run Tests......Page 418
Performance Tests......Page 419
String Tests......Page 420
Auxiliary Compressors......Page 421
Fuel Gas Compression (FGC)......Page 422
Reciprocating Compressors......Page 429
Centrifugal Compressors......Page 430
Operational Considerations......Page 431
Reciprocating Compressors......Page 432
Pulsation Dampeners......Page 433
Cylinder Orifice Plates......Page 435
Mixed Compression......Page 436
Flow-Induced Excitation Analyses: FIV, AIV, and FIT......Page 437
System Dynamic Analyses......Page 440
Transient Surge in Centrifugal Compressor Systems......Page 441
Station Dynamics......Page 443
Rotordynamic Analyses......Page 445
Mechanical: Skid and Piping Analyses......Page 447
References......Page 448
Further Reading......Page 449
Motivation, Justification, and Objectives......Page 450
Pressure......Page 451
Vibration......Page 452
Data Acquisition......Page 454
Uncertainty and Error Sources......Page 455
Axial and Centrifugal Compressors......Page 456
Reciprocating Compressors......Page 459
Screw Compressors......Page 461
References......Page 462
Introduction......Page 463
Centrifugal Compressors......Page 464
Reciprocating Compressors......Page 465
Electric Motor Drives......Page 466
Steam Turbines......Page 467
Couplings......Page 468
Torque Convertors......Page 469
Instrumentation and Data......Page 470
Piping and Pressure Vessels......Page 471
Performance Testing and Inspection......Page 472
Rotordynamics......Page 473
Machinery Installation......Page 474
Health, Safety, and Fire......Page 475
Further Reading......Page 477
Section IV: Technology Developments
......Page 482
Introduction......Page 483
Wet Gas Defined in the Upper Limit......Page 484
Experimental Testing......Page 487
General Arrangement......Page 488
Liquid Injection......Page 490
Sealing on the Compressor......Page 495
Pressure, Temperature, and Flow Rate Measurement and Control......Page 496
Screw Type......Page 498
Screw Pumps in Wet Gas Conditions......Page 499
Additional Observed and Tested Issues......Page 501
Reciprocating Compressors......Page 502
Erosion......Page 503
Corrosion......Page 505
Damage From Liquid Slugging......Page 506
Mechanical Vibrations and Lube/Seal Oil Interactions With Wet Gas......Page 508
Lube/Seal Oil Interactions in Reciprocating Compressor Applications......Page 511
Difficulties in Testing......Page 512
Important Test Variables......Page 513
Effects of Liquid Distribution......Page 514
Liquid Distribution......Page 516
Liquid Distribution Inside a Rotating Impeller......Page 519
Compressor Leading Edge......Page 520
Inside Compressor Impeller......Page 523
Compressor Trailing Edge and Diffuser......Page 527
Experimental Research......Page 530
Direct Integration Approach......Page 532
Evaporation Performance Models......Page 533
Compressor Aerodynamic Performance......Page 534
Compressor Rotordynamics......Page 536
Liquid Separation......Page 537
Subsea Compression......Page 540
Subsea Requirements......Page 543
Subsea Compression System Concepts......Page 547
Compressor Internals......Page 550
Subsea Compressor Concepts......Page 551
Bearings......Page 552
Motor......Page 553
OneSubsea: WGC (Wet Gas Compressor)......Page 554
Hermetically Sealed and Oil-Free Compression......Page 555
Stator Windings Exposed to the Process Gas Versus Canned Windings......Page 557
High-Pressure Compressors......Page 558
High-Speed Drive Trains......Page 561
Integrated Separator Centrifugal Compressor......Page 563
Cooled Diaphragms for Centrifugal Compressors......Page 565
Subsurface Compression......Page 567
Advanced Seals......Page 570
Gas Bearings......Page 571
Additive Manufacturing......Page 572
Gas Property Testing......Page 574
Linear Motor Compressor......Page 576
Project Goal......Page 577
Technical Overview......Page 578
Advanced Pulsation Control for Reciprocating Compressors......Page 579
Advanced Reciprocating Compressor Valve Technology......Page 581
Surge Force Predictions......Page 582
Liquid Packing Seals for Reciprocating Compressors......Page 584
References......Page 586
Further Reading......Page 588
C......Page 589
D......Page 590
G......Page 591
I......Page 592
N......Page 594
R......Page 595
S......Page 596
W......Page 598
Back Cover......Page 600

Citation preview

Compression Machinery for Oil and Gas

Compression Machinery for Oil and Gas

Edited by

Klaus Brun Elliott Group, Jeannette, PA, United States

Rainer Kurz Solar Turbines, San Diego, CA, United States

Gulf Professional Publishing is an imprint of Elsevier 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States The Boulevard, Langford Lane, Kidlington, Oxford, OX5 1GB, United Kingdom © 2019 Elsevier Inc. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-814683-5 For information on all Gulf Professional publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Brian Romer Senior Acquisition Editor: Katie Hammon Editorial Project Manager: Gabriela D. Capille Production Project Manager: Kamesh Ramajogi Cover Designer: Victoria Pearson Typeset by SPi Global, India

Contributors Numbers in parentheses indicate the pages on which the authors’ contributions begin.

Timothy C. Allison (3, 375, 543), Southwest Research Institute, San Antonio, TX, United States Brian Bauer (309), Elliott Group, Jeannette, PA, United States Urs Baumann (543), MAN Diesel & Turbo Schweiz AG, Z€ urich, Switzerland Eugene “Buddy” Broerman (253, 569), Southwest Research Institute, San Antonio, TX, United States Dirk B€ uche (31, 485), MAN Diesel & Turbo Schweiz AG, Z€ urich, Switzerland Ryan Cater (485), Southwest Research Institute, San Antonio, TX, United States Hector Delgado-Garibay (449), Southwest Research Institute, San Antonio, TX, United States Kenneth Hall (309), Caterpiller Oil & Gas, Lafayette, IN, United States Martin Hinchliff (167, 309, 427), Dresser-Rand, Painted Post, NY, United States Justin Hollingsworth (31, 167, 253), Southwest Research Institute, San Antonio, TX, United States Kevin Hoopes (3), Southwest Research Institute, San Antonio, TX, United States Min Ji (427), Solar Turbines, Inc., San Diego, CA, United States Terry Kreuz (13, 387), National Fuel Gas, Williamsville, NY, United States Chris Kulhanek (31, 167, 463), Southwest Research Institute, San Antonio, TX, United States Rainer Kurz (3, 31, 309, 427, 449), Solar Turbines, Inc., San Diego, CA, United States Tim Manthey (253), Aerzen USA, Coatesville, PA, United States Franzisko Maywald (167), Burckhardt Compression AG, Winterthur, Switzerland Cyrus Meher-Homji (309, 401), Bechtel, San Francisco, CA, United States Kolja Metz (135), MAN Energy Solutions, Berlin, Germany Harry Miller (543, 569), Dresser-Rand, A Siemens Business, Olean, NY, United States J. Jeffrey Moore (463, 569), Southwest Research Institute, San Antonio, TX, United States Dave Moss (401), UE Compression, Henderson, CO, United States Grant Musgrove (309, 401, 485), Southwest Research Institute, San Antonio, TX, United States xv

xvi Contributors

Rob Pelton (135), Hanwha Power Systems Americas Inc., Houston, TX, United States Brian Pettinato (31, 309), Elliott Group, Jeannette, PA, United States Greg Phillippi (167, 449, 463), Ariel Corporation, Mount Vernon, OH, United States Nathan Poerner (449), Southwest Research Institute, San Antonio, TX, United States Aoron Rimpel (135, 167, 569), Southwest Research Institute, San Antonio, TX, United States Dragan Ristanovic (309), Bechtel, San Francisco, CA, United States Sarah Simons (387, 427, 569), Southwest Research Institute, San Antonio, TX, United States Avneet Singh (375), Solar Turbines, Inc., San Diego, CA, United States Natalie R. Smith (543), Southwest Research Institute, San Antonio, TX, United States Matt Taher (31, 309, 463), Bechtel, San Francisco, CA, United States George Talabisco (13, 31, 427), Dresser-Rand, Olean, NY, United States Leif Arne Tonnessen (543), TechnipFMC, Kongsberg, Norway Joseph Thorp (375), Saudi Aramco Energy Ventures - North America, Houston, TX, United States Christian Wacker (135), MAN Energy Solutions, Berlin, Germany J€ urgen Wennemar (253), MAN Energy Solutions SE, Oberhausen, Germany Ferdinand Werdecker (31), EagleBurgman, Houston, TX, United States Benjamin White (13, 387, 427, 463), Southwest Research Institute, San Antonio, TX, United States Jason Wilkes (31), Southwest Research Institute, San Antonio, TX, United States Karl Wygant (135, 569), Hanwha Power Systems Americas Inc., Houston, TX, United States Donghui Zhang (31, 449), Solar Turbines, Inc., San Diego, CA, United States

The Editors Dr. Klaus Brun, Elliott Group, Jeannette, PA, United States Dr. Brun is the director of Research & Development at Elliott Group where he leads a group of over 60 professionals in the development of turbomachinery and related systems for the energy industry. His past experience includes positions in product development, engineering, project management, and executive management at Southwest Research Institute, Solar Turbines, General Electric, and Alstom. He holds 9 patents, authored over 350 papers, and published 3 textbooks on energy systems and turbomachinery. Dr. Brun is a Fellow of the ASME and won an R&D 100 award in 2007 for his Semi-Active Valve invention. He also won the ASME Industrial Gas Turbine Award in 2016 and 11 individuals ASME Turbo Expo Best Paper awards. Dr. Brun is the chair of the 2020 Supercritical CO2 Power Cycles Symposium, past chair of the ASME-IGTI Board of Directors, the ASME Oil & Gas Applications Committee, and ASME sCO2 Power Cycle Committee. He is also a member of the API 616 Task Force, the ASME PTC-10 task force, the Asia Turbomachinery Symposiums Committee, and the Supercritical CO2 Symposium Advisory Committee. Dr. Brun is currently the executive correspondent of Turbomachinery International Magazine and Associate Editor of the ASME Journal of Gas Turbines for Power. Dr. Rainer Kurz, Solar Turbines, San Diego, CA, United States Dr. Rainer Kurz is the manager, Systems Analysis at Solar Turbines Incorporated, in San Diego, California. His organization is responsible for predicting compressor and gas turbine performance, for conducting application studies, and for field performance testing. Dr. Kurz attended the Universitaet der Bundeswehr in Hamburg, Germany, where he received the degree of a Dr.-Ing. in 1991. He joined Solar Turbines in 1993, and holds his current position since 1995. Dr. Kurz is the past chair of the ASME/IGTI Oil and Gas Applications Committee, a member of the Gas Machinery Research Council Project Supervisory Committee, the GMC Conference Organizing Committee, the Texas A&M Turbomachinery Symposium Advisory Committee, the Asian Turbomachinery Symposium Advisory Committee, and the SDSU Aerospace Engineering Advisory Committee. He was elected ASME Fellow in 2003. He has authored numerous publications on turbomachinery-related topics, with an emphasis on compressor applications, dynamic behavior, and gas turbine operation and degradation. Many of his publications were considered of

xvii

xviii The Editors

archive quality and were accepted for publications in Engineering Journals. He has received several “Best Paper” and “Best Tutorial” Awards at the ASME TurboExpo Conferences and is the recipient of the 2013 Industrial Gas Turbine Technology award.

Preface Oil and natural gas provides the world with high-density energy and feedstock for transportation, power generation, chemical processes, and many industrial and consumer products. The oil and gas industry produces these hydrocarbons and then processes, transports, and distributes them to users. Almost all aspects of the oil and gas industry require some gas compression and there are many applications where compression is the most critical part of the production and transport chain of hydrocarbons. For example, to transport natural gas from the production well to a power plant, within a city gas distribution system, or to a chemical plant, several compression facilities with complex machinery trains are necessary. Other production-related oil and gas compression applications include gas gathering, flash gas compression near the well, the use of the gas for enhanced oil recovery, the recompression of gas after processing in a gas plant, chemical plants and refineries processes, and even fuel gas compression at power plants. Further downstream in the hydrocarbon product chain, refrigeration compression is used to liquefy natural gas for ease of transport by ship, rail, or truck. Fundamentally, all natural gas must be compressed so that it can be efficiently transported, stored, or processed. Besides conventional hydrocarbon gases, such as natural gas, there are many other oil and gas applications where other nonhydrocarbon gasses such as carbon dioxide, nitrogen, or hydrogen have to be compressed. Gas compression is thus vital for the transportation and processing of hydrocarbon and nonhydrocarbon gasses that are required by the oil and gas industry and their customers. For example, in North America alone there are over 9000 pipeline compression stations with nearly 40,000 individual compressors to transport natural gas from the producer to the consumer. Similarly, gas compressors are a critical part in the upstream production and downstream refining and distribution infrastructure. The design of compression systems for the oil and gas industry is challenging due to the environments in which these systems operate, the fact that the operating conditions may change significantly on all time scales, and the extremely cyclical operating demands. Additionally, the equipment is expected to operate for long-time intervals, often many years, without interruptions for maintenance, or unplanned shutdowns. There are many different types of compressors and drivers that are being used for these applications, each with different features, limits, and capabilities.

xix

xx Preface

This book provides a comprehensive overview of the compression machinery that is utilized in the oil and gas industry. In the first section of this book the thermodynamic foundation of gas compression is explained (Chapter 1) and an overview over the different types of typical used oil and gas compression equipment is provided (Chapter 2). The second section of this book provides a deeper look into this compression equipment including a discussion on the performance and aerodynamics of centrifugal compressors in Chapter 3, the performance of integrally geared compressors in Chapter 4, the performance of reciprocating compressors in Chapter 5, and screw compressors in Chapter 6. The section is concluded by a description of the different drivers for these compressors such as gas turbines, steam turbines, expanders, electric motors, and gas engines (Chapter 7). The book’s third section highlights relevant application topics. Here, we discuss in Chapter 8 how compressors are applied in the upstream section, where compressors are used for gas gathering, gas lift, gas reinjection, and for compression in gas plants. Then, in Chapter 9, the typical midstream applications, such as pipeline compression, and gas storage are explained. In the downstream business (Chapter 10), compression needs in LNG facilities, refineries, and for fuel gas compression are covered. Important system design and analysis issues are discussed in Chapter 11, and the section is completed with a discussion of testing methods and requirements in Chapter 12, and relevant codes and standards in Chapter 13. The final section (Section 4) gives an outlook to future challenges and technology developments in oil and gas compression. The important topic of wet gas compression is covered in detail in Chapter 14 and a description of new machinery concepts, such as subsea and downhole compression, or hermetically sealed compressors can be found in Chapter 15. Some leading edge concepts such as the use of linear motors, supersonic compression, and isothermal compression in Chapter 16 round out the topics in this book. The topics covered in this book were selected to provide engineers and practitioners interested in the oil and gas industry and the relevant compression machinery utilized with a comprehensive technology and applications overview. Each chapter provides sufficient background material to stand alone, and can be used on its own, although we attempted to avoid duplication throughout this book. We, the editors, are indebted to the chapter authors. They are all subject matter experts in their fields, who were selected from the engineering and scientific community based on their relevant contributions to the field. They represent a broad range of expertise, and come from a diverse range of backgrounds. Klaus Brun Rainer Kurz Contributing authors (*Indicates chapter lead): Timothy C. Allison (Chapters 1, 8*, 15)

Preface

xxi

Brian Bauer (Chapter 7) Eugene “Buddy” Broerman (Chapters 6*, 16) Dirk Buche (Chapters 3, 14) Jon Bygrave (Chapter 3) Ryan Cater (Chapter 14*) Hector Delgado (Chapter 12*) Kenneth Hall (Chapter 7) Martin Hinchliff (Chapters 5, 7, 11) Justin Hollingsworth (Chapters 3, 5*, 6) Kevin Hoopes (Chapter 1*) Min Ji (Chapter 11) Terry Kreuz (Chapters 2, 9) Chris Kulhanek (Chapters 3, 5, 13*) Rainer Kurz (Chapters 1, 3, 7, 11, 12) Tim Manthey (Chapter 6) Franzisko Maywald (Chapter 5) Cyrus Meher-Homji (Chapters 7, 10) Kolja Metz (Chapter 4) Harry Miller (Chapters 15, 16) Jeff Moore (Chapters 13, 16) Dave Moss (Chapter 10) Grant Musgrove (Chapters 7*, 10*, 14) Rob Pelton (Chapter 4) Brian Petinato (Chapter 3*, 7) Greg Phillippi (Chapters 5, 12, 13) Nathan Poerner (Chapter 12) Aaron Rimpel (Chapters 4*, 5, 16) Dragan Ristanovic (Chapter 7) Sarah Simons (Chapters 9, 11*, 16*) Avneet Singh (Chapter 8) Natalie Smith (Chapter 15*) Matt Taher (Chapters 3, 7, 13) George Talabisco (Chapters 2, 3, 11) Leif Tonnessen (Chapter 15) Christian Wacker (Chapter 4) J€urgen Wennemar (Chapter 6) Ferdinand Werdecker (Chapter 3) Benjamin White (Chapters 2*, 9*, 11, 13) Jason Wilkes (Chapter 3*) Karl Wygant (Chapters 4, 16) Donghui Zhang (Chapters 3, 12)

Acknowledgments We would like to thank Dorothea Martinez for her tireless efforts and assistance while putting this book together.

xxiii

Chapter 1

Oil and Gas Compressor Basics Kevin Hoopes*, Timothy C. Allison* and Rainer Kurz† *

Southwest Research Institute, San Antonio, TX, United States, †Solar Turbines, Inc., San Diego, CA, United States

Overview of Compressor Types Gas compressors operate by adding work to a gas to increase the pressure of that gas as it flows through them. They are used in many different applications from everyday items such as vacuum cleaners, automobiles, and air conditioners to large industrial scale compressors for chemical processing, jet engine propulsion, and natural gas processing and transmission. They are separated into two distinct groups: positive displacement compressors and dynamic compressors.

Positive Displacement Compressors Positive displacement compressors operate by decreasing the volume of a gas in a trapped volume. Because they operate on a trapped volume of fluid, positive displacement machines operate on distinct portions of the fluid at a time; as such their mechanical behavior, operating speed, etc., is very different than dynamic machines. Examples of compressors of this type include reciprocating compressors, screw compressors, and scroll compressors.

Dynamic Compressors Dynamic compressors operate by continuously increasing the momentum of a gas as it flows through them and do not rely on a trapped volume. Examples of compressors of this type include centrifugal (also called radial) compressors, axial compressors, and mixed flow compressors. The major distinctions between these categories come from how the fluid enters and exits the machine. In a centrifugal machine, the fluid flows into the machine parallel to the axis of rotation and out of the machine radially or perpendicular to the axis of rotation. In axial machines, the gas enters and exits the machine parallel to the axis of rotation. As their name suggests, mixed flow machines are a mixture between purely centrifugal and purely axial machines. Compression Machinery for Oil and Gas. https://doi.org/10.1016/B978-0-12-814683-5.00001-8 © 2019 Elsevier Inc. All rights reserved.

3

4 SECTION

I Fundamentals of Compression

The appropriate type of compressor for a particular application is a function of the required flow rate and pressure ratio. A chart describing the approximate operating envelopes of different compressor types has been provided by the Natural Gas Processor Suppliers Association and is shown in Fig. 1.1. Although the exact capabilities of a particular compressor type may deviate from these conditions based on a specific design, the general trends are valid. In general, there is significant overlap between the three compressor types, although reciprocating compressors uniquely cover low-flow applications with high pressures and centrifugal compressors uniquely cover high-flow applications.

Basic Thermodynamics The working principles of gas compressors can be understood by applying the basic laws of physics. Using the first and second law of thermodynamics together with basic laws of fluid dynamics, such as Bernoulli’s law and Euler’s law allows us to explain the fundamental working principles, and by extension, can increase the understanding of the operational behavior of gas compressors. Most descriptions of compressors presented here are specifically geared toward pipeline applications. They are usually also applicable to many other gas compression applications. The general description of the thermodynamics of gas compression applies to any type of compressor, independent of its detailed working principles.

Discharge pressure vs. inlet flow 100,000

Discharge pressure (psig)

10,000

1000

100 Rotary-screw Recip-single stage Recip-multistage

10

Centrifugal-single stage Centrifugal-multistage

1 1

10

100

1000

10,000

100,000

1,000,000

Inlet flow (acfm)

FIG. 1.1 Compressor types and application conditions. (Modified from NGPSA Engineering Data Book, vol. 1, Revised tenth ed., 1994. Compiled and edited in cooperation with the Gas Processors Association. Copyright 1987 Gas Processors Association.)

Oil and Gas Compressor Basics Chapter

1

5

First Law For a compressor receiving gas at a certain suction pressure and temperature, and delivering it at a certain output pressure, the isentropic head represents the energy input required by a reversible, adiabatic (thus isentropic) compression. The actual compressor will require a higher amount of energy input than needed for the ideal (isentropic) compression. It is important to clarify certain properties at this time, and in particular find their connection to the first and second law of thermodynamics written for steady-state fluid flows. The first law (defining the conservation of energy) becomes:

 w22 w21 + gz2
h1 + + gz1 ¼ q12 + Wt12 h2 + 2 2 with q ¼ 0 for adiabatic processes and gz ¼ 0 because changes in elevation are not significant for gas compressors. We can combine enthalpy and velocity into a total enthalpy by ht ¼ h +

w2 2

where Wt12 is the amount of work we have to apply to affect the change in enthalpy in the gas. The work Wt12 is related to the required power, P, by multiplying it with the mass flow. _ t12 P ¼ mW Power and enthalpy difference are thus related by P ¼ m_ ðht,2
ht,1 Þ If we can find a relationship that combines enthalpy with the pressure and temperature of a gas, we have found the necessary tools to describe the gas compression process. For a perfect gas, with constant heat capacity, the relationship between enthalpy, pressures, and temperatures is Δh ¼ cp ðT2
T1 Þ Because, for an isentropic compression, the discharge temperature is determined by the pressure ratio (with k ¼ cp/cv):
 p2 k
1 k + T1 T2 ¼ T1 p1 We can, for an isentropic compression of a perfect gas, relate the isentropic head, temperature, and pressures by 
  p2 k
1 k
1 Δhs ¼ cp T1 p1

6 SECTION

I Fundamentals of Compression

For real gases (for which k and cp in the above equations become functions of temperature and pressure), the enthalpy of a gas h(p, T) is calculated in a more complicated way using equations of state [1]. They represent relationships that allow the calculation of the enthalpy of gas of known composition, if any two of its pressure, its temperature, or its entropy are known. We therefore can calculate the actual head for the compression by Δh ¼ hðp2 , T2 Þ
hðp1 , T1 Þ and the isentropic head by Δh ¼ hðp2 , s1 Þ
hðp1 , T1 Þ s1 ¼ sðp1 , T1 Þ The performance quality of a compressor can be assessed by comparing the actual head (which directly relates to the amount of power we need to spend for the compression) with the head that the ideal, isentropic compression would require. This defines the isentropic efficiency: ηs ¼

Δhs Δh

The second law tells us: Z m_ ðs2
s1 Þ ¼ 1

2

dq + Sirr T

For adiabatic flows, where no heat q enters or leaves, the change in entropy simply describes the losses generated in the compression process. These losses come from the friction of gas with solid surfaces and the mixing of gas of different energy levels. An adiabatic, reversible compression process therefore does not change the entropy of the system, it is isentropic. Our equation for the actual head implicitly includes the entropy rise Δs, because Δh ¼ hðp2 , T2 Þ
hðp1 , T1 Þ ¼ hðp2 , s1 + ΔsÞ
hðp1 , T1 Þ If cooling is applied during the compression process (e.g., with intercoolers between two compressors in series), then the increase in entropy is smaller than that for an uncooled process. Therefore, the power requirement will be reduced. Using the polytropic process [2] for comparison reasons works fundamentally the same way as using the isentropic process for comparison reasons. The difference lies in the fact that the polytropic process uses the same discharge temperature as the actual process, while the isentropic process has a different (lower) discharge temperature than the actual process for the same compression task. In particular, both the isentropic and the polytropic process are reversible processes. In order to fully define the isentropic compression process for a given gas, suction pressure, suction temperature, and discharge pressures have to be known. To define the polytropic process, in addition either the polytropic compression efficiency, or the discharge temperature has to be known.

Oil and Gas Compressor Basics Chapter

1

7

The polytropic efficiency ηp is defined such that it is constant for any infinitesimally small compression step, which then allows to write Z 1 p2 Δhp vdp ¼ Δh ¼ ηp ηp p1 and

Z Δhp ¼

p2

vdp p1

or, to define the polytropic efficiency: ηp ¼

Δhp Δh

For designers of compressors, the polytropic efficiency has an important advantage: If a compressor has five stages, and each stage has the same isentropic efficiency ηs, then the overall compressor efficiency will be lower than ηs. If, for the same example, we assume that each stage has the same polytropic efficiency ηp, then the polytropic efficency of the entire machine is also ηp. Because the enthalpy definition above is on a per mass flow basis, the absorbed gas power Pg (i.e., the power that the compressor transferred into the gas) can be calculated as _ Pg ¼ mΔh The mechanical power P necessary to drive the compressor is the gas absorbed power increased by all mechanical losses (friction in the seals and bearings), expressed by a mechanical efficiency ηm (typically in the order of 1% or 2% of the total absorbed power): P¼

_ s 1 mΔh _ mΔh ¼ ηm ηs ηm

We also encounter energy conservation on a different level in turbomachines: The aerodynamic function of a turbomachine relies on the capability to trade two forms of energy—kinetic energy (velocity energy) and potential energy (pressure energy). This will be discussed in a subsequent section.

Intercooling A compression process where the gas is cooled as part of the compression is no longer adiabatic. It is thus not appropriate to state isentropic or polytropic processes for comparison. In some instances, an isothermal efficiency might be suitable to compare different configurations. Since the cooling process moves entropy from the compressed gas to the environment, the overall compression will consume less power than the same process without intercooling.

8 SECTION

I Fundamentals of Compression

Equations of State Understanding gas compression requires an understanding of the relationship between pressure, temperature, and density of a gas. An ideal gas exhibits the following behavior: P ¼ RT ρ where R is the gas constant, and as such is constant as long as the gas composition is not changed. Any gas at very low pressures can be described by this equation. For the elevated pressures we see in natural gas compression, this equation becomes inaccurate, and an additional variable, the compressibility factor Z, has to be added: P ¼ ZRT ρ Unfortunately, the compressibility factor itself is a function of pressure, temperature, and gas composition. A similar situation arises when the enthalpy has to be calculated: For an ideal gas, we find Z T2 Cp dT Δh ¼ Cp ΔT ¼ T1

where Cp is only a function of temperature. In a real gas, we get additional terms for the deviation between real gas behavior and ideal gas behavior (Poling et al., 2001): Z T2     Cp dT
h0
hðp2 Þ T2 Δh ¼ h0
hðp1 Þ T1 + T1

The terms (h
h(p1))T1 and (h
h(p2))T2 are called departure functions, because they describe the deviation of the real gas behavior from the ideal gas behavior. They relate the enthalpy at some pressures and temperatures to a reference state at low pressure, but at the same temperature. The departure functions can be calculated solely from an equation of state, while the term R T2 C dT is evaluated in the ideal gas state. Fig. 1.2 shows the path of a calcuT1 p lation using an equation of state. Equations of state are semiempirical relationships that allow to calculate the compressibility factor, as well as, the departure functions. For gas compression applications, the most frequently used equations of state are Redlich-Kwong, Soave-Redlich-Kwong, Benedict-Webb-Rubin, Benedict-Webb-RubinStarling, and Lee-Kessler-Ploecker (Poling et al., 2001). In general, all of these equations provide accurate results for typical applications in pipelines, that is, for gases with a high methane content, and at 0

0

Oil and Gas Compressor Basics Chapter

1

9

FIG. 1.2 Temperature-entropy diagram for a Brayton cycle.

pressures below about 42 MPa. Kumar et al. [3] and Beinecke and Luedtke [2] have compared these equations of state regarding their accuracy for compression applications. It should be noted that the Redlich-Kwong equation of state is the most effective equation from a computational point of view (because the solution is found directly rather than through iteration).

p-h and T-s Diagrams The state of any gas of known composition is fully defined if exactly two parameters are known. These parameters could be pressure and temperature, pressure and entropy, enthalpy and entropy, or specific volume and temperature. This fact allows the use of p-h (pressure-enthalpy) or T-s (temperature-entropy) diagrams to graphically describe thermodynamic processes such as the gas compression process, or thermodynamic cycles like the gas turbine Brayton cycle. Any gas or gas mixture can be displayed as a p-h or T-s diagram. A p-h diagram displays the same information that can be calculated by an equation of state. Typically, p-h diagrams show lines of constant pressure, constant volume, constant entropy, constant temperature, as well as, the two-phase areas. T-s diagrams, often show constant pressure or constant volume lines. For practical purposes, p-h and T-s diagrams are available for pure gases and air in many textbooks examples of which are shown in (Figs. 1.3 and 1.4).

Basic Machinery Dynamics Rotordynamics includes the lateral and torsional vibrations of machinery trains. Torsional analysis is typically required for both reciprocating and centrifugal compressors, but lateral analysis is often performed only for centrifugal

10 SECTION

I Fundamentals of Compression

Pressure

= Efficiency

P2

1 = Suction 2 = Discharge

2isen

2

S1 T1 P1 1 Enthalpy

h1

h2isen h2

Enthalpy(2isen) – Enthalpy(1) Efficiency =

Enthalpy(2) – Enthalpy(1)

FIG. 1.3 Pressure-enthalpy diagram for gas compression.

FIG. 1.4 Pressure enthalpy diagram for methane, with the path for actual and isentropic compression. The range where methane can be treated as an ideal gas (i.e., at low pressures, where enthalpy is only dependent on temperature, but not on pressure) is highlighted. (From GPSA Handbook.)

Oil and Gas Compressor Basics Chapter

1

11

compressors since they are not often a concern for reciprocating units due to their low operating speed and resulting frequency separation between lateral natural frequencies and excitation frequencies. Regarding lateral rotordynamics, there are two types of vibration: forced response (typically from unbalance) and self-excited (usually referred to as stability). Rotordynamic stability in centrifugal compressors depends on forces in the bearings, seals, and secondary flow passages that can be stabilizing or destabilizing. The stabilizing forces include the bearing and seal damping which works to dissipate energy. The destabilizing forces arise from tangential forces that increase rotor vibration amplitudes. Assessing rotordynamic stability is an accounting exercise, quantifying the rotordynamic forces of each component on the stability of the overall rotor system. Destabilizing forces can be represented by cross-coupled stiffnesses, which will be described in more detail in a later chapter. Destabilizing forces in compressors extract energy from the process gas and rotation to excite a rotor vibration mode (usually the first forward whirling mode). If these forces grow large enough and overcome the damping forces in the rotor system, the vibration amplitude of that mode will grow unbounded until a rotor-stator rub occurs. Vibration monitoring systems will shutdown the unit after an instability, but usually the vibration amplitude grows so quickly that seal rubbing will typically result. Further aggravating the situation is that many operators will restart the compressor and repeat the exercise many times before they realize they have a serious problem. The damage can go beyond just replacing the internal labyrinth seals and more serious damage to shafts, impellers, bearings, and diaphragms can result.

References [1] B.E. Poling, J.M. Prausnitz, J.P. O’Connell, The Properties of Gases and Liquids, McGraw-Hill, 2001. [2] D. Beinecke, K. Luedtke, Die Auslegung von Turboverdichtern unter Beruecksichtigung des realen Gasverhaltens, VDI Berichte (1983) 487. [3] S. Kumar, R. Kurz, J.P. O’Connell, Equations of State for Compressor Design and Testing, (1999) ASME Paper 99-GT-12.

Chapter 2

Equipment Overview Benjamin White*, Terry Kreuz† and George Talabisco‡ *

Southwest Research Institute, San Antonio, TX, United States, †National Fuel Gas, Williamsville, NY, United States, ‡Dresser-Rand, Olean, NY, United States

Types of Compression Equipment Compression equipment is available in a wide range of types and sizes for use in many different industry applications. Compressors are generally divided into two main categories: positive displacement compressors and dynamic compressors. Positive displacement compressors typically include reciprocating compressors (single acting, double acting, hyper), rotary compressors (screw, scroll, vane), and diaphragm compressors. Dynamic compressors include centrifugal compressors and axial compressors. A summary of compressor types is shown in Fig. 2.1. Each compressor type is usually best suited for a particular range of flow rates, pressures, and fluid types. However, there is also a lot of overlap in some applications. For industrial oil and gas applications, the two most common compressor types are reciprocating and centrifugal compressors. In very broad terms, reciprocating compressors are usually best for applications with lowto-medium flow and low-to-high pressures. Reciprocating compressors also tend to be very flexible in terms of varying flow rates, gas compositions, and fluid densities. Multistage reciprocating compressors can generate very highpressure ratios across a single machine. Centrifugal compressors are usually best for applications with medium-to-high flow and low-to-medium pressures. Centrifugal compressors typically have smaller physical footprints for a given power rating, lower maintenance costs, and longer run times between maintenance intervals. Centrifugal compressors also have lower unbalanced forces and are typically much less susceptible to pulsation and vibration-related issues. However, as discussed later in this chapter, there are also many other factors to consider when selecting the ideal compressor for a particular application. Fig. 2.2 shows typical coverage ranges for various types of compressors. These range from axial compressors which can produce very high flow rates with

Compression Machinery for Oil and Gas. https://doi.org/10.1016/B978-0-12-814683-5.00002-X © 2019 Elsevier Inc. All rights reserved.

13

14 SECTION

I Fundamentals of Compression

Compressor types

Dynamic

Positive displacement

Reciprocating

Single-acting

Rotary

Axial

Centrifugal

Double-acting Lobe

Screw

Vane

Diaphragm Liquid ring

Scroll

FIG. 2.1 Family tree of compressor types [1]. https://commons.wikimedia.org/wiki/File:Compressor_ Types.png.

FIG. 2.2 Compressor coverage chart [2]. (Courtesy of Southwest Research Institute).

limited pressure ratio and discharge pressure capabilities, to diaphragm compressors that produce very limited flow rates with usually very high discharge pressures. The others compressors listed cover parts of the displayed pressures and flow ranges. As a type, reciprocating compressors generally cover the

Equipment Overview Chapter

2

15

broadest area of the range. Detailed information on each compressor type and driver options can also be found in ***Chapters 3–7.

Types of Applications There are a wide variety of different compression applications, typically divided into three main categories: Upstream, Midstream, and Downstream. Note that this section is just a brief summary of what types of equipment would commonly be used in each application. Refer to Chapters 8–10 for more detail.

Upstream 1. Gas gathering application can typically utilize screw, rotary vane, reciprocating, and centrifugal compressors to collect gas at producing wellheads and move the gas to a central location for processing and/or transport through a series of various pipelines. 2. Export compression increases the pressure of processed gas to be suitable for exporting from a processing facility to a pipeline for use by downstream consumers. Typical compressors used are centrifugal and reciprocating. 3. Gas lift is a process where high-pressure gas is reinjected into the well riser to mix with the fluid, thus helping with lifting oil from a well by making the weight of the fluid column lighter in weight. This helps to maintain or enhance production from a well. Both centrifugal and reciprocating can be utilized in this process. 4. Gas reinjection can utilize centrifugal and reciprocating compressors to maintain or enhance production in an oil reservoir that contains both oil and gas. In this process the gas is injected into the reservoir to maintain pressure and thus oil production rate. 5. Vapor recovery units—in vapor recovery, gas that separates from oil in a storage system (tank or vessel) is recovered by removing the gas that collects on the top of a vessel thus reducing the pressure within a storage tank and capturing the gas for use elsewhere. Reciprocating, rotary screw, rotary vane, or centrifugal compressors are typically employed as the compression equipment for this application process. 6. Subsea compression equipment located in a subsea environment is used to maintain reservoir production by reinjection into a well formation. Centrifugal compressors with direct coupled high-speed motors are typically used for this application. 7. Enhanced oil recovery (EOR) can include gas injection, water flood, and chemical or steam injection. EOR is used to maintain reservoir pressure by injection of gas either recycled from the reservoir itself or from an alternate source using a different fluid like carbon dioxide (CO2), nitrogen, or other gas streams can be injected. Typical compression equipment used can be reciprocating and centrifugal compression.

16 SECTION

I Fundamentals of Compression

Midstream 1. Pipeline transmission compression is the transportation of natural gas from the source (wells) to the distribution system. This is a typical application for centrifugal and reciprocating compression systems. 2. Gas storage injection/withdrawal: old well formations are sometimes used as peaking storage locations for natural gas transmission and distribution systems. They are used to provide additional capacity at various locations where peak demand can exceed capability of existing pipelines to provide the necessary product at the required pressure levels. Typical compression equipment consists of both reciprocating and centrifugal compressors that are used to inject the gas into the storage formation and to boost the pressure during removal to meet pipeline pressure requirements. 3. Natural gas processing: this application removes unwanted components from raw and untreated natural gas to make it suitable to be classified as pipeline quality gas. Centrifugal and reciprocating compressors can be used in this application.

Downstream 1. Refinery/process applications: compressors constitute an important part of the mechanical equipment in oil and gas refineries and petrochemical plants. Compressors are used for different applications in the main and auxiliary process cycles: a. Recycling compressors designed to provide a steady flow of process gas through a closed circuit in order to maintain the required process parameters in the plant units (e.g., hydrogen-rich gas recycling in a hydrotreater) b. Feed compressors supplying process gas to reactor c. Booster compressors d. Sales gas compressors (e.g., methane) e. Electrically driven reciprocating and centrifugal compressors are most commonly used in oil and gas refining facilities 2. Ethylene/Low-density polyethylene (LDPE): due to the extremely high discharge pressures required in these applications, large horizontal opposed reciprocating compressors for the highest discharge pressures are required (aka hyper compressors). This type of compressor is used in the LDPE production process (low-density polyethylene) as a secondary compressor in combination with a process gas compressor used as a booster/primary compressor. 3. CNG (compressed natural gas) is a fuel for vehicles or other commercial applications made by compressing natural gas or biogas to less than 1% of its uncompressed volume. Compressors are used to boost the pressure of the gas and they are the primary equipment of a CNG refueling station

Equipment Overview Chapter

4.

5.

6.

7.

2

17

for vehicles. This application involves suction pressures ranging from a slight vacuum to more than 5.2 MPa, with discharge pressures exceeding 31 MPa, thus requiring multistaged compressor technologies. Typical compression types utilized include two-to-five stage reciprocating piston compressors. LNG (liquefied natural gas) is a growing application with increasing demands for use in industry and domestic heating making it necessary to use huge quantities of gas from far away production sources which create the need to transport natural gas in its liquid form via ships and tank trucks. Among the equipment required for this technology, centrifugal compressors and rotary machines prove to be very versatile for use in gathering produced gas, as well as, in later stages of pipeline transportation, liquefaction, and regasification/expansion at the point of use by residential, commercial, and industrial consumers. Industrial gases (hydrogen/CO2/ammonia) hydrogen compression is applied to reduce the volume resulting in compressed hydrogen or liquid hydrogen. The compressor reduces the volume of hydrogen gas, to allow the liquid hydrogen to be transported elsewhere. A proven method to compress hydrogen is to apply reciprocating piston compressors. Nonlubricated compressors are preferred to avoid oil contamination of the hydrogen. Ammonia compression to manufacture ammonia is a part of a complex chemical refinery process wherein the synthesis gas needs to be compressed to extremely high pressures, ranging from 100 to 250 bars (1500–4000 psi) for ammonia synthesis. Modern plants employ centrifugal compressors which are usually driven by steam turbines that use the steam produced from excess process heat. Ammonia is also used as a process fluid for refrigeration systems, the most common type of refrigerating system is a vaporcompression refrigerator. This approach uses anhydrous ammonia as a refrigerant as the means of moving heat around. CO2 compression: CO2 is often used as an inexpensive, nonflammable pressurized gas and it is one of the most commonly used compressed gases for pneumatic (pressurized gas) systems in portable pressure tools. CO2 gas producing process plants produce CO2 in mainly two forms—liquid and solid. Solid CO2 is also known as “dry ice” and is used as a refrigerant in food industry and for small shipments. CO2 is widely utilized during the storage and shipping of ice cream and other frozen foods. CO2 is also used as an atmosphere for welding. CO2 is compressed to the desired usage pressure using a gas compressor or is liquefied at lower pressures by using compressor driven refrigeration systems and then pumped to the desired pressure for bulk storage for CO2 capture or large-scale industrial uses. Large-scale CO2 compressors are responsible for a large portion of the enormous capital and operating cost penalties expected with any carbon capture and sequestration (CCS) systems. The compressor power requirements for

18 SECTION

I Fundamentals of Compression

gas-phase compression of CO2 range from 1.7 to 15.27 MPa versus CO2 liquefaction at 1.7 MPa and pumping that liquid CO2 to 15.27 MPa. 8. Distribution: After natural gas has been produced and cleaned, it must be pumped/compressed and stored for consumption by a multitude of residential, commercial and industrial users primarily for residential heat and industrial processes. To maximize capacity and deliverability to midstream and transmission pipelines feeding lower pressure gas distribution systems, gas is typically compressed to higher pressures ranging 3.4–10 MPa. The required boosting compression function power is delivered using a mix of reciprocating engines, turbines, and electric motors. The prime power devices are used to drive a mix of reciprocating piston, centrifugal, and screw compressors.

Factors to Consider When Selecting Compression Equipment Selecting the ideal compressor for a given application should include consideration for a wide variety of complex factors. Many of these key considerations are presented in the following.

Flow Rates and Absolute Pressures The maximum flow of a reciprocating compressor is limited by the cylinder size (bore and stroke), the number of compressor cylinders, and the operating speed. The maximum flow of a centrifugal compressor is typically limited by driver power or some choke point internal to the compressor where the flow velocity nears the speed of sound of the fluid. Both reciprocating and centrifugal compressors can reach maximum discharge pressures in the range of 70–100 MPa. However, discharge pressures of approximately 10 MPa or less are most common. Special “hyper” reciprocating compressors, often used in low-density polyethylene production, can reach discharge pressures up to 340 MPa. Fig. 2.3 provides an overview comparison of typical ranges of flow rates and discharge pressures that can be attained with different types of compressors. These criteria are a good starting point for selecting a compressor type. However, when multiple compressor types overlap for a given application, the following factors should be considered.

Compression Ratio and Maximum Temperatures For a single-stage reciprocating compressor, the compression ratio typically varies from 1.1 up to 4.0. The maximum compression ratio is usually limited by a goal of preventing discharge temperatures from reaching no more than approximately 296.33 °F. However, multistage reciprocating compressors with

Equipment Overview Chapter

2

19

Table 6-1 Comparison of prime movers* High-speed diesel (900–1800 rpm)

Low-speed diesel (400–600 rpm)

100–150

175–250

0.75 95 1200/1800/3600 99.9 100,000

0.48 0–10,000 25–45 35 900–1800 90 30,000

0.59 10–12,000 5–10 30 400–600 99 75000

25,000

4000

20,000

Electric motor Initial cost, $/bhp Fuel rate A. No. 2 diesel Ib/bhp/h B. Gas, BTU/bhp/h Maintenance,† $/bhp/yr Efficiency,‡ % Speed, rpm Availability on load base, % Time between overhauls, h Time between major inspections, h

20–30 (energy rate + demand)

*1000 to 5000 bhp in base-load operation. †Maintenance includes all parts and contact labor (excludes fuel) for U.S.A. onshore. ‡Turbine overall efficiency can be doubled or tripled by applying heat recovery. Source: Oil & Gas Journal 16 February 1981, p. 87

FIG. 2.3 Comparison of prime movers [3].

interstage cooling of the process gas can achieve much higher overall compression ratios. Single-stage centrifugal compressors are typically designed to operate with compression ratios in the range of 1.1–1.4.

Gas Type (Sour Gas, Wet Gas, etc.) Many natural gas streams carry contaminants in varying amounts warranting careful consideration to the design, installation, and maintenance of adequate gas cleaning facilities upstream of compression equipment to adequately protect them from serious problems resulting from operating with dirt, wet, or sour gas. Wet or dirty gas entrained with contaminates in reciprocating compressors can result in compressor valve damage and accelerated wear of compressor pistons by wiping lubrication from critical areas of the compressor. Worst case, entrained liquids could fill the compressor cylinders resulting in serious damage of compressor rods, bolting mechanisms, or catastrophic failure of the unit. Contaminants in centrifugal compressors can quickly cause erosion of impellers, volutes, diaphragms and shaft, and pressure seals. For all of these reasons gas contaminants must be removed to avoid efficiency losses, breakdowns, and high maintenance costs. For wet gas (primarily water), a well-designed system consisting of a combination of separators, scrubbers, filters, and gas dehydrators are the industry standard to provide a palatable gas quality upstream of the compression process so as to not risk wet gas detriments. “Wet gas” can also refer to unprocessed gas streams containing higher quantities of hydrocarbon liquids that require more advanced processing and treatment equipment to process out or “strip” these liquids prior to the gas compression process to reduce fallout of liquids in the compression cycle.

20 SECTION

I Fundamentals of Compression

Corrosive gases (aka “sour” gas) such as sulfur rich or hydrogen sulfide compounds may be present in some raw production gas streams. This always requires special treatment and added design and protection to the compressor’s internal steel and sealing materials to withstand the corrosive gas streams as much as technically possible. Industry standard treatment systems for corrosive gases include wetting the gas stream with specialized chemicals (amine or similar) in contactor towers then separating the liquid from the gas stream with separators. Specialized dry chemical absorbents are also commonly used to remove corrosive gases in lower flow designs.

Driver Type and Fuel Source There are generally three main types of “prime movers,” which include reciprocating engines, gas turbines, and electric motors. Steam turbines are also used in process applications where steam is generally available for other process reasons. Reciprocating engines are typically used to drive reciprocating compressors and they can often be subdivided into “slow speed” (
600 rpm) and “medium-to-high speed” (600–1800 rpm) classifications. These engines can have two-stroke or four-stroke cycles and may or may not be turbo-charged. These “slow speed” engines are often used in an “integral” configuration where the compressor and engine share a common crankshaft. Separable systems have physically separate drivers and compressors connected by a gearbox or coupling. Gas turbines can be industrial type or aeroderivative type (based on designs originally intended for aviation applications). Gas turbines operate at much higher operating speeds and are well suited for operation with centrifugal compressors or where minimizing weight is a priority (such as offshore). Electric motors can be induction, synchronous, or DC, with induction being the most common. Electric motor drives pair well with reciprocating compressors based on their similar operating speeds, but electric motors can also be paired with centrifugal compressors. One key factor in determining the preferred driver type will be the available fuel sources. Both gas engines and gas turbines can often run off of fuel gas pulled from the main process gas. However, electric motors obviously require an uninterrupted source of electrical power. Other factors such as installation cost, operating and maintenance costs, efficiency, noise, emissions, reliability, etc. should also be considered. See additional discussion on drivers in Chapter 7.

Emissions, Noise, and Safety Factors relating to emissions, noise, and safety must be considered in any new compression facility. There are numerous regulatory and permitting requirements that must be met in regard to these factors to protect the public health.

Equipment Overview Chapter

2

21

The operation of compression equipment will typically generate emissions to the air during the combustion process (depending on the type of driver) and will typically be subject to various emission regulations, such as the Clean Air Act. Planned and unplanned emissions tests are often conducted on a regular basis by local regulatory representatives. Intentional or unintentional discharges of the process gas from the system should also be considered (leaks, blowdowns, etc.). Methane, the primary component in natural gas, is a known greenhouse gas. In addition, CO2 and nitrous oxides are natural by-products of the combustion process. The common sources of air emissions include: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

combustion gases, boilers, release of gas, fugitive omissions, gas vents, gas leaks, storage tanks, reciprocating engines, heaters, dehydration, blowdown, crankcase vents of reciprocating compressors, packing case vents for reciprocating compressors, cooling system vents, and other miscellaneous factors.

Given the proximity of many compressor stations to residential neighbors, noise generated by the compression equipment can be a factor, particularly when ambient noise conditions are low, such as at night in rural settings. Sources of noise could include the turbine or engine air intakes and exhaust systems, blowdown systems, gas coolers, auxiliary air compressors or generators, etc. Pulsation and vibration in the piping can also be similar irritant to nearby residential neighbors. Valves especially antisurge-recycle valves can be a source of noise and require noise attenuation trim along with acoustic insulation to minimize noise. A maximum noise limit of 55 dB(A) is often required by FERC (Federal Energy Regulatory Commission), which can difficult to meet in some installations. Noise control is usually achieved by employing a combination of techniques including the installation of engine exhaust and inlet silencers, noise insulation on the piping, acoustically insulated building and enclosures, etc. A comprehensive noise assessment should be conducted at the design phase. Compressor stations are required to be designed, constructed and operated in accordance with Pipeline and Hazardous Materials Safety Administration (PHMSA) safety standards. These standards are intended to minimize the risk

22 SECTION

I Fundamentals of Compression

of pipeline accidents and to protect the public safety. PHMSA conducts inspections during various phases of design, construction, and operation. State and local laws also apply, depending on the location of the facility. Surface and groundwater, as well as, local wildlife and vegetation, should be also protected as required by the EPA, US Bureau of Land Management, and other local agencies. The safety and security at a compression facility can be enhanced by installed safety features such as fire and combustible gas detection, security systems, environmental monitoring systems, emergency shutdown systems, and overpressure protection systems. Initiating site-specific safety programs for on-site personnel are also critical.

Physical Environment Compressor stations can vary in size from a few acres up to 20 or more. The most common and important location attributes that should be considered in the selection of a compressor system site include: l l l l

l

available utilities, accessibility to services and support, environmental conditions, depending on the application, other location variables may have to be considered as well, and the relative importance of the location variables should also be established to ensure that the best type and size of compressor is selected.

Operating and Maintenance Costs Maintenance costs for compressor stations and systems are usually tracked and quantified as $/BHP-hr. Fuel and/or power costs to power the prime mover obviously represent a large part of the operating costs for engines and compressors. Although fuel represents a portion of the operating cost for engines and compressors, maintenance and other consumables can be quite expensive. Comparing the cost of parts and consumables on dollars only can be misleading for different size engines and operating hours. Dollars per brake horsepower per hour is calculated by simply dividing maintenance costs by horsepower and by run hours. Maintenance costs are generally captured on a level designated by the operating company’s accounting system at the most detailed level maintenance costs are tracked by. Horsepower can be tracked based on an actual or rated basis. Either is okay as long as the same standard is used for all engines. Maintenance costs should be tracked over time (generally over several years). These costs are very useful when compared against other units and prime mover types (internal combustion

Equipment Overview Chapter

2

23

engine, gas turbine, or electric motor). It can be used to identify best practices and opportunities for improvement. Reliability is a relative percentage-based measure of planned versus unplanned downtime for a given compressor it measures how often the compressor is down unplanned for repairs, or planned to be preventatively maintained. Although each company defines it differently, annual reliability is most simply calculated as: Annual reliability ¼ ð8760h  unplanned down hours + planned down hoursÞ= ð8760  planned down hoursÞ: Availability is a slightly different measure of unscheduled downtime for a given compressor. It measures whether the compressor is ready for use when needed. Although each company defines it differently, annual availability is simply calculated as: Annual availability ¼ ð8760h  unplanned down hoursÞ=8760: Generally, and as supported by industry studies, gas turbine or electrical driven centrifugal compressors have proven to be lower cost in overall maintenance and have higher reliability/availability factors than internal combustion engine driven reciprocating compressors. The following are typical observational comparisons often used in operating gas turbines and centrifugal compressors: 1. output power versus control temperatures, 2. output power versus fuel flow rate or fuel pressure (increases in fuel flow rate or fuel pressure can be an indication of compressor fuel nozzle or meter calibration issues), 3. compressor discharge pressure versus percent gas producer speed (decreases in compressor discharge pressure at a given gas producer speed can indicate fouling of the compressor or an increase in turbine nozzle area), 4. vibration level (increases in vibration at constant operating conditions can indicate a number of problems such as compressor internals fouling, bearing damage, mechanical damage, and loose connections), and 5. exhaust temperature variation (variation across the exhaust collector can indicate: fuel distribution problems, combustor problems, or air/fuel mixing problems). Factors that may adversely affect maintenance intervals on gas turbines include these factors: 1. starting/Stopping frequency (rule of thumb says one start/stop ¼ 30 run hours equivalence), 2. frequent rapid load changes, and 3. condition of inlet air to the turbine

24 SECTION

I Fundamentals of Compression

The highest impact operator and maintenance activities for gas turbine and centrifugal compressors include: 1. active fluids and mechanical analysis program, 2. ensuring the fuel system is free from contamination that will lead to nozzle or torch plugging, 3. the most common sources of contamination are: liquids, iron sulfide, and glycol, 4. clean inlet combustion air supply, 5. regular detergent washing of the combustion air compressor/air inlet turbine section of the package, and 6. eliminate frequent starts and stops.

Service Life The key to extended service life of gas compressors and prime power movers is developing and maintaining a sound compressor maintenance program which includes: 1. regular original equipment manufacturer (OEM) recommended preventive maintenance (varies with the type and age of equipment), 2. load factor, use factor, and operating environment, 3. spare parts availability and maintenance, 4. personnel and training requirements, and 5. monitoring and inspection techniques. Maintenance scheduling is determined by company policy or OEM recommendations and can be formed on the basis of fixed intervals or predictive failures. Emergency maintenance is performed after a failure or major breakdown has occurred and usually only the component which has failed will be repaired or replaced. Maintenance and repair of compressor station equipment should be performed by fully trained and qualified company and contracted personnel. Prime power movers of the connected gas compressor section of a compressor equipment package can be divided into three major divisions: internal combustion gas fueled engines, gas fueled power turbines, and electric motors. All three prime mover types can be protected from catastrophic failures and maintained at top operating and fuel efficiency by installing extensive monitoring/ shutdown/alarm devices and adopting a robust preventative and predictive maintenance program. A condition monitoring plan that utilizes machine-type specific data analytics to move toward a predictive maintenance program is a worthwhile undertaking for all prime movers to preserve and extend service life of the equipment. The application of a good condition monitoring plan can yield such benefits as preemptive repairs, reduced maintenance, increased unit availability, reduced

Equipment Overview Chapter

2

25

risk of catastrophic failures, and potential for process improvements leading to greater efficiency and profitability. Relative historical service lives of the prime mover categories before extensive and expensive unit overhauls are: l l l l

slow-speed gas engine (300–600 rpm): 75,000 operating hours, high-speed gas engine (900–1800 rpm): 25,000–50,000 operating hours, gas turbine: 30,000 operating hours, and electric motor: 100,000 h.

Pulsation and Vibration Considerations Most positive displacement compressors (primarily reciprocating and screwtype compressors) can be subject to significant pulsation and vibration-related concerns and consideration of these issues should be included in the overall design effort. Centrifugal compressors inherently have lower unbalanced forces and are typically less susceptible (but not immune) to pulsation and vibration related issues. See Chapter 11 for detailed discussion on this topic.

Automated/Manual Controls The wide variation of compressor station size, complexity, type, and yearly utilization affects the sophistication and configuration of manual versus automated control systems selected. Generally compressor stations can be divided into three types or modes of operation. The classification of each mode depends on the degree of automation required by the company. Usually as the degree of automation increases the sophistication of the controls and instrumentation also increases. Typical terms used in three “modes of operation” at varying levels for compressor stations are: “local—manual,” “local—automatic,” and “remote—automatic.” The complexity and selection of the control mode varies with the amount of remote controlled automatic operation required. They may range from simple manually operated actuators to remotely controlled computerized systems with data logging/reporting and remote diagnostic capabilities. Regardless of complexity, control systems should be modular and constructed to facilitate expansion, troubleshooting, and ease of maintenance. Controls are becoming increasingly electronically operated or activated. There are also numerous options in the industry for preengineered “off the shelf” mechanical and electropneumatic individual unit control panels for a shorter list of critical monitoring and safety shut-down functions specific to standard industry compression packages allowing on-off control tasks and monitoring/alarming of key machine package operations. Controls of this nature are relatively inexpensive and can be interfaced easily with engine and compressor equipment. They are also advantageous in that they can be used safely in locations classified as having hazardous atmospheres.

26 SECTION

I Fundamentals of Compression

Industry standard sophisticated programmable logic controller (PLC) equipment is utilized most often in complex applications or regulation of equipment that requires more than simple on/off control, or where it is not feasible to transmit data and signals via communications lines to control equipment. Modern electronic controls are advantageous because they can be interfaced directly to commercially standard and available PLCs and computers, providing flexibility in operation and ease of expansion. Because of the variety of technical operating equipment found within a compressor station, the most efficient and cost-effective control systems are combined systems which include components of all types of control equipment. Electronic and electric controls are interfaced with mechanically operated controls on devices to provide safe and extremely accurate control systems with standard designs and reasonable costs.

Air Filtration Requirements Ambient air to be used for the combustion of fuel in an internal combustion engine or gas turbine must be provided in the right quantity and quality. The air inlet system primarily consists of an inlet air filter in the piping or ducting to direct the air to the engine or turbine. Silencers may also be required to control with the noise created by the flow of air and by the engine or turbine itself. Ambient-rated turbocharged engines as well as gas turbines also require some means of controlling the temperature of the air entering the engine if the full power is to be developed.

Operating Company Preferences Certain operating companies may have preferences for new compression needs based on their historical experience and existing installed compression machinery. There may be logistical benefits to selecting certain compressor types based on staff familiarity, training, spare parts, local suppliers, etc.

Packaging Issues Most high-horsepower, slow-speed reciprocating compressors (typically less than 500 rpm, either separable or integral units) are installed with a direct mount to a concrete foundation. This allows for the most rigid foundation to react to any substantial unbalanced forced. This direct to foundation mounting approach involves significant effort in the field during the installation process. Most modern high-speed compressors (over 500 rpm) are typically packaged on a steel skid that can be assembled in a shop and then transported to site. The compressor skid package allows for the compressor, driver, pulsation vessels, piping, auxiliary system (lubrication, cooling, etc.) to all be mounted on a single, portable skid. This packaging approach provides advantages in assembly time,

Equipment Overview Chapter

2

27

assembly cost, quality control, etc. Most reciprocating compressors are packaged by third party companies (not the OEMs). Therefore, each package can be a unique configuration and combination of components. Centrifugal compressors and their drivers are also typically skid mounted. However, most of the centrifugal compressor manufacturers will package the compressors themselves to ensure maximum compatibility of all components.

References [1] https://commons.wikimedia.org/wiki/File:Compressor_Types.png. [2] GPSA Engineering Data Book, 14th ed., Gas Processors Association, 2017. [3] Oil Gas J. (1981) 87.

Chapter 3

Centrifugal Compressors Jason Wilkes*, Brian Pettinato†, Rainer Kurz‡, Justin Hollingsworth*, Donghui Zhang‡, Matt Taher§, Chris Kulhanek*, Ferdinand Werdecker¶, € chek and George Talabisco# Dirk Bu *

Southwest Research Institute, San Antonio, TX, United States, †Elliott Group, Jeannette, PA, United States, ‡Solar Turbines, Inc., San Diego, CA, United States, §Bechtel, San Francisco, CA, United States, ¶EagleBurgman, Houston, TX, United States, kMAN Diesel & Turbo Schweiz AG, € Zurich, Switzerland, #Dresser-Rand, Olean, NY, United States

Basics of Inline Centrifugal Compressors Inline centrifugal compressors can be described as a combination of one or more centrifugal compressor stages operating about a common axis. Each stage is comprised of rotating impellers that impart kinetic energy into the gas that is then converted to a rise in static pressure as the gas decelerates. For the remainder of this chapter, it can be assumed that all of the impellers are operating on the same shaft. The different working principles cause differences in the operating characteristics of the centrifugal compressors compared to those of the reciprocating unit. Centrifugal compressors are used in a wide variety of applications in chemical plants, refineries, onshore and offshore gas lift and gas injection applications, gas gathering, and in the transmission and storage of natural gas. Centrifugal compressors have been used for outlet pressures as high as 70 MPa, thus overlapping with reciprocating compressors over a portion of the flow-rate/ pressure domain. Centrifugal compressors are usually either gas turbine or electric motor driven, although in downstream applications, steam turbines or turboexpanders may be used. Typical operating speeds for centrifugal compressors in gas transmission applications are about 14,000 rpm for 3700 kW units and 8000 rpm for 15,000 kW units.

General Requirements for Range, Pressure Ratio, and Performance for Inline Centrifugal Compressors in Various Applications? The requirements for Range, Pressure ratio, and Performance vary greatly between applications. While pipeline compressors typically operate at low pressure ratios (say, from 1.1 to 1.7), other applications may require significantly Compression Machinery for Oil and Gas. https://doi.org/10.1016/B978-0-12-814683-5.00003-1 © 2019 Elsevier Inc. All rights reserved.

31

32 SECTION

II Types of Equipment

higher ratios. In gas gathering, we may find pressure ratios of 25. The limits for centrifugal compressors are typically either due to rotordynamic concerns, which limit the speed and length of the rotor, or by temperature limits. Higher pressure ratios may require multiple compressor bodies per train, and the capability for intercooling. The operating range required also depends on the application. Since for many applications, pressure ratio and flow are correlated by the compression system, range has to be seen in that context. On the other hand, many applications only see one distinct operating condition. Depending on the range of operating conditions, different control methods (variable speed, variable geometry, recycle) may be used. For applications that require the compressor to cover a wide range of operating conditions, good compressor efficiency over a wide range is usually more important than high peak efficiency.

Elements of an Inline Centrifugal Compressor Fig. 3.1 shows a cross section of a barrel-style centrifugal compressor. While the elements of a compressor will be discussed in detail later, they will be introduced briefly here. Starting at the suction flange of the compressor, gas enters the compressor through the inlet, whose purpose is to transition the flow from radial to axial. In most barrel machines, fixed guide vanes help to reduce swirl and provide an axisymmetric flow field coming into the compressor; however, Inlet guide vane

Inlet housing

Casing

Stator assembly

Discharge volute Impeller

Suction flange (Port)

Discharge bearing and seal assembly Balance piston

Coupling hub or balance sleeve

Discharge flange (Port) Suction bearing and seal assembly

Stator Diffuser passage

FIG. 3.1 Barrel-style centrifugal compressor. (Courtesy of Solar Turbines, Inc.)

Centrifugal Compressors Chapter

3

33

some machines employ variable guide vanes to induce swirl and alter the relative flow angle seen by the compressor. After passing through the inlet, the flow is directed into the impeller. It is the impellers responsibility to increase the kinetic and potential energy of the gas by pushing it tangentially. This requires torque that is transferred to the impeller by the shaft, which is ultimately turned by a driver. After the flow leaves the impeller at high velocity, it slows down in the diffuser building pressure. Due to losses and the heat of compression, the gas builds up temperature as pressure increases. If there are more compression stages, the gas leaves the diffuser and enters a return channel, which directs the flow back to the inlet of the next stage. Once the gas leaves the final stage of compression in a given section, it passes from the diffuser into a collector or volute that directs the flow radially outward through the discharge flange. Other components consist of the casing that acts as an overall structure and pressure vessel. Seals and walls separate the stages and ensure that gas follows the primary path, minimizing leakage from the flow path. The shaft acts as the inner boundary of the flow and also supports and drives the impellers, which in turn performs the work and drives the gas. Bearings act as shaft support both radially and axially and as part of an overall rotor-bearing dynamic system that is tuned to allow robust operation of the machine.

Types/Configuration Inline compressors typically have centrifugal stages located in between the bearings; however, the casing can be packaged in a few different configurations. They are typically a barrel style, horizontally split, or hermetically sealed.

Barrel Fig. 3.2 shows a barrel-style compressor. In this photo, the rotor bundle, composed of rotor and stator blade passages that composes the compressor, is being removed axially from the casing. The casing has the responsibility of locating the centerline of the compressor relative to the ground and containing the pressure, while the bundle is responsible for directing the flow during the compression process. Having no horizontal split, a barrel-style casing can operate at very high pressures. Horizontal Split Compressors with horizontal split casings as shown in Fig. 3.3 are used in relatively low-pressure applications having a pressure limitation of roughly 7 MPa that is governed mostly by the ability to seal the split line against leakage. These horizontal split casing designs may also be described as “axially split.” The horizontal split of the casing is flat and bolted around the periphery with numerous split-line bolts to prevent leakage.

34 SECTION

II Types of Equipment

FIG. 3.2 Photo of a barrel-style casing with bundle removed. (Courtesy of Solar Turbines, Inc.)

FIG. 3.3 Horizontally split casing.

One advantage of the horizontal split casing design is the easy access to the compressor internals from overhead after the lid is removed. This feature is especially useful when the compressor is used as a drive-thru unit having one or more compressors attached at the nondrive end.

Hermetically Sealed A hermetically sealed inline compressor as shown in Fig. 3.4 is unique in that the driver is located inside of the pressure containing casing. The driver in a hermetically sealed compressor is a high-speed motor since there are no gears. Although the motor cavity may be at a reduced pressure or purged with helium

Centrifugal Compressors Chapter

3

35

FIG. 3.4 Blue-C hermetically sealed compressor.

to reduce windage, there is no possibility that gas from within a hermetically sealed casing may escape to the environment since there are no shaft end seals. For this reason, there are two primary uses for hermetically sealed units. Common applications are subsea compression applications for gas gathering or artificial lift. In this application, the hermetically sealed compressor is located on the surface of the ocean floor. The blue-C compressor is one example of the technology as pictured below. Hermetically sealed compressors also have the potential to be used to compress gases that are toxic; however, this is less common.

Overhung/Inline Overhung inline compressors are the least common configuration seen in oil and gas applications. These units are similar to an integrally geared compressor, however, the gearbox is not located on the structure of the compressor, but is instead driven by an external gearbox. Multisection So far, the flow through the casing has been described as passing from stage-tostage as the gas is compressed, ultimately leaving the compressor from the discharge flange. This is true for a straight-through single-section compressor as shown in Fig. 3.5 (top left). A two-section compressor, as shown in the other windows in Fig. 3.5 have two distinct compressor sections, where gas passes from one section to the next by leaving the compressor casing and entering through a second inlet. This process allows for an intercooler to be used to

Rotation

Intake

Discharge

Discharge

36 SECTION

Intake

II Types of Equipment

Rotation First intake

First discharge

Second discharge

Second intake

Intake

Discharge

Intake

FIG. 3.5 (top left) Single section straight through compressor, (top right) two-section straight through compressor, (bottom left) two-section back-to-back, and (bottom right) two-section dual flow. (Image Courtesy of Dresser Rand.)

Centrifugal Compressors Chapter

3

37

improve compressor efficiency. In the case of a dual flow arrangement, the sections run in parallel allowing for twice the flow through the compressor.

Compressor Components Impeller Impellers are at the heart of the compressor. Each impeller is part of a stage that is selected to meet performance requirements. Impellers are selected based on aerodynamic performance and mechanical integrity. Mechanical integrity is typically evaluated by finite element analysis (FEA), while aerodynamic performance is evaluated by a combination of one-dimensional (1D) and computational fluid dynamic (CFD) tools. Impellers are either a semiopen design or closed as depicted in Fig. 3.6 (left) and (right), respectively. Semiopen impellers consist of a hub and blades manufactured as a single piece. Closed impellers also consist of a hub and blades, but further adding a cover or shroud. The manufacture of modern closed impellers may be as a single piece or as two pieces joined together by welding or brazing. In the two-piece design, the hub and blades are typically milled together and the cover is joined to the blades by welding or brazing. An alternative two-piece design consists of the blades and cover milled together then joined to the hub by welding or brazing. Single-piece impellers are either milled or cast. Closed impellers are more common for inline centrifugal compressors. The reason for this is that assembly tolerances and long spans between impellers and thrust surfaces results in a high degree of axial motion of the impellers relative to the stationary flowpath elements. Since a typical eye seal on a compressor impeller is a radial seal, the impeller can shift axially without degrading performance. In an open wheel, shroud clearance is directly linked to performance, and large clearances required in multistage centrifugal compressors would result in poorer performance than a closed wheel counterpart.

Rib Puller hole for impeller removal Hollow inner hub

Hollow between ribs

FIG. 3.6 Cross section of a closed (left, [1]) and open (right, [2]) centrifugal impeller.

38 SECTION

II Types of Equipment

FIG. 3.7 Flow coefficients of various impellers. (Courtesy of Solar Turbines, Inc.)

Impeller welds are subjected to stringent nondestructive examination to ensure their quality. They are then overspeed tested to confirm their integrity. Although many impeller wheels may look similar, there are a few terms that can be used to help describe them more fully. Flow coefficient is one such term. The flow coefficient of an impeller describes the volumetric flow passing through the compressor as a function of tip diameter and speed. It is given by Φ1 ¼ ðDQÞ13 N, where Q1 is the inlet flow, D2 is the tip diameter, and N is the rota2 tional speed. Fig. 3.7 shows a representative range of flow coefficients for a theoretical impeller.

Diffuser and Return Channel Fig. 3.8 shows a photo of a split casing diaphragm. The diaphragm is the stationary portion of a stage that creates the bundle when stacked to together. In this photo, the impeller sits with inlet facing downward spinning about a vertical axis. When the gas leaves the exit of the impeller, it enters the entrance region of the diffuser. This region typically has a converging feature to smoothly take the Diaphragm

Return vanes Return Impeller exit Diffuser

FIG. 3.8 Diffuser and return channel.

Centrifugal Compressors Chapter

3

39

exit flow from the compressor and direct it into the diffuser which is often 10%–20% narrower than the impeller flowpath at the exit. This reduction in diffuser height is called pinch. Pinch is used to further accelerate the flow as it enters the diffuser, which improves the stability of the flow in the diffuser at low flows, and increases compressor range. Once the flow is inside of the diffuser, it may be diffused in a vaneless space, or there could be vanes. In most oil and gas applications, a vaneless diffuser is employed since they tend to have wider range than vaned diffusers, however, vaned diffusers tend to offer better performance at an optimized flow point. After the flow is sufficiently diffused, it returns through the u bend and starts traveling back along the return channel where the flow enters the next compressor stage.

Inlet The inlet of a compressor takes the flow from a pipe flange, and attempts to distribute it circumferentially equally in the axial direction. This can be accomplished with inlet vanes. Collectors Collectors take the flow exiting the last stage of a section and route the flow out through a flange on the casing to the discharge pipe. These generally take two forms, constant area collectors and volutes. Volute Volutes are more common than collectors, as they generally have less losses, improving the efficiency of the compressor. A representative volute is shown in Fig. 3.9. The tongue of the volute is generally the “start” of the volute, which

FIG. 3.9 Volute.

40 SECTION

II Types of Equipment

increases in area as more flow is added in the direction of rotation. As you approach the tongue from the other side, the goal is to have the entire flow exit the volute, preventing recirculation of flow. Similar to an impeller, volutes have a design point flow where they have optimized efficiency, and tend to perform worse as flow increases or decreases away from the design point. Collector A collector is employed in applications where performance is not as critical. Often, these elements are cheaper to manufacture, hence their benefit. A collector is a constant area annular cavity that the diffuser abruptly discharges into. This abrupt transition into a slow velocity field results in irreversible losses.

Seals Seals in a compressor function to maintain a pressure difference between two cavities. These seals can exist between stationary bodies (static seals), or they seal between a rotating and stationary component. Fig. 3.10 illustrates a compressor along with common seal locations. Balance Piston/Division Wall Seal Balance piston or division-wall seals are often employed in centrifugal compressors to counteract (balance) the net axial force produced by the impellers during operation. In a straight-through compressor, this seal is typically referred to as the balance-piston seal, while in back-to-back machines, this seal is the division wall seal, as it divides the sections of the machine.

FIG. 3.10 Seal locations in a blade-to-blade compressor.

Centrifugal Compressors Chapter

3

41

These seals are normally of the labyrinth or hole-pattern variety. Labyrinth types generally allow low leakage rates (which is beneficial from a compressor performance standpoint) since tighter clearances may be used because the seals are usually made of soft materials (e.g., aluminum). These materials are unlikely to damage the shafting during intermittent contact. However, these seals do little to counteract the destabilizing cross-coupling forces that are generated in the impellers and seals. This can increase the potential for self-excited rotordynamic instabilities. Hole-pattern seals (which include the popular honeycomb variety) require additional clearance, as they are generally made of harder materials (e.g., Hastelloy) which can easily damage shafting. As a result, these seals impose a slight penalty to compressor performance due to higher leakage levels when compared to labyrinth types, but have the potential to add significant stiffness and damping to the system, which can be very beneficial from a rotordynamic stability standpoint. Eye/Interstage Seals Seals are commonly included at the impeller eyes and between stages of centrifugal compressors. These seals are almost always of the labyrinth variety. As such, they are beneficial from a performance standpoint, due to the tight clearances made possible by using abradable materials. However, they can generate destabilizing cross-coupling coefficients that should be included in an American Petroleum Institute (API) level II rotordynamic analysis (Fig. 3.11). Antiswirl Mechanisms The flows entering balance piston and eye/interstage seals are commonly subjected to significant swirling effects, primarily due to leakage between the impellers and shrouds. This swirl tends to increase the cross-coupled stiffness generated by the seals, which can result in diminished rotor stability. To counter this effect, two mechanisms are commonly utilized: swirl brakes and shunt holes. Swirl brakes redirect the flow with physical barriers (similar to flow straightening vanes) oriented at an angle which reduces (or in some cases Eye seal

Interstage seal FIG. 3.11 Eye seal and interstage seal location.

42 SECTION

II Types of Equipment

completely negates) the swirl. These devices are particularly effective for labyrinth types, and may be used in either eye/interstage or balance piston seals. Shunt holes, which are commonly utilized in either labyrinth or hole-pattern balance piston seals, involve the use of redirected flow which is introduced radially, thus opposing and reducing the swirl. Shaft End Shaft end seals are required to seal the gas inside the compressor at the point where the compressor or piston shaft penetrates the casing. This vital sealing function is necessary to prevent escape of process gas to the environment surrounding the compressor. With the exception of reciprocating compressors, all other types of compressors have rotating shafts. Accordingly, the applied type of seals differs. Reciprocating compressors mainly use serial-arranged packing rings for sealing the piston rod. For compressors with rotating shaft, however, dry-gas-lubricated mechanical seals (dry gas seals (DGSs)) are commonly used. The importance of oil-lubricated mechanical seals is reduced to applications where requirements in pressure, speed, and power consumption are low. Bushing-type seals as the shaft end seal are reduced to applications where leakage requirements play a minor role. Labyrinths can be considered technically obsolete. However, bushing-type carbon ring seals or labyrinth seals are commonly used as separation seal mounted between the DGS and the shaft bearing. Their function is both to protect the DGS from spray oil and to prevent uncontrolled process gas entry into the bearing cavity in the event of a total failure of the DGS. The today’s most common type of shaft end seal in the oil and gas industry is based on DGS technology. This triumphal procession goes back to the end of the 1980s of the last century. The use of DGSs in the place of oil lubricated mechanical seals improves mechanical efficiency as the shear and friction power of the seal is significant reduced. The application limit of a single oil-lubricated seal stage is about 100 m/s peripheral speed at the outer diameter of the seal faces and 50 bar differential pressure. By comparison, gas seals achieve more than 200 m/s and a pressure difference of more than 450 bar per single seal stage. The resulting degrees of freedom in the compressor design justify the success of gas sealing technology. Fig. 3.12 shows the design of a DGS is basically similar to that of an oillubricated mechanical seal and consists of a stationary face (2) sealed by an axially displaceable seal element (4) and a rotational face (1). At standstill and depressurized condition, the stationary face is pressed by a set of circumferentially arranged springs (3) against the rotational face. A thrust ring is used to transmit the singular spring forces. The rotational face is centered by a shaft sleeve mounted on the compressor shaft. On its back side, another secondary seal element (4) is placed. The spring-loaded stationary seal face is centered by the seal housing mounted in compressor casing parts. In general, DGSs are pressurized from the outer side.

Centrifugal Compressors Chapter

3

43

FIG. 3.12 Block representation of a dry-gas seal.

FIG. 3.13 Force balance on a dry-gas seal.

The diameters of the faces are designed to achieve hydrostatic liftoff at standstill with a certain pressure difference across the seal as shown in Fig. 3.13. The axially displaceable seal element on the back of the stationary face seals is the so-called hydraulic diameter, which is always larger than the inner diameter of the sealing surface. This means that the effective pressure area on the back of the stationary face is smaller than the sealing surface and the prerequisite for a stationary liftoff is given.

44 SECTION

II Types of Equipment

With appropriate evenness and parallelism of the facing surfaces, only the pressure pad in the sealing gap is able to keep the closing force on the back of balance. This gives the seal the ability to regulate the height of the sealing gap itself. Once any additional force tries to disturb the balance the gap height will change, and the pressure pad in the seal gap will create immediately an appropriate counter force. The height of the sealing gap is only a few micrometers. The spring force is only dimensioned so that it can safely overcome the frictional force of the dynamic sealing element in the unpressurized state, and therefore, plays a minor role. The closing force is mainly represented by a rectangular pressure pad. Starting from the outer diameter so-called gas grooves (5) are incorporated in one face. These are usually incorporated in the rotational face and reach into the middle of the sealing surface and are only a few micrometers deep. The main function of these grooves is to achieve an additional pressure pad and thus liftoff during dynamic operation. The rotation creates a shear flow in the grooves which in turn induces pressure generation (Fig. 3.14). The available groove geometries can be divided into unidirectional and bidirectional ones. Above all, due to the hardly avoidable possibility of reverse

t inle Ga s

Di

rec tio

no fr ot

a ti

on

Gas in

Compression

Direction of rotation

FIG. 3.14 Grooves in a DGS (top) and resulting pressure dam action (bottom).

Centrifugal Compressors Chapter

3

45

TABLE 3.1 DGS Materials Part

Material

Housing, shaft sleeve

Stainless steel or nickel basis alloy

Rotational face

Silicon carbide, silicon nitride, or tungsten carbide

Stationary face

Silicon carbide or carbon

Secondary seals

Elastomer o-rings or spring energized PTFE lib-seals

Springs

Nickel basis alloy

rotation of the compressor shaft and the smaller number of necessary spare seals, bidirectional grooves are gaining more and more popularity in the oil and gas industry. Table 3.1 gives an overview of the typical built-in materials. At the beginning of the use of DGSs, especially hard-soft pairings with the rotational face made of tungsten carbide and the stationary face made of carbon were common. However, with increasing pressure and speed requirements, hard-hard pairings have proven to be more advantageous. DGSs require clean and dry seal gas for reliable operation in order to avoid particles or condensates between the seal faces. The seal gas is typically taken from the compressor discharge and then throttled, cooled, and filtered as part of a seal supporting system that will be discussed later. This conditioned seal gas is injected between the process-side seal and the seal faces. The process-side seal is typically a single labyrinth seal and is located between the DGS and the compressor internals as shown in Fig 3.15. When seal gas is injected, it prevents flow of process gas to the DGS. The majority of the seal gas flows into the compressor and a slight amount flows across the seal faces. The seal gas pressure is reduced across the seal faces to vent pressure. The gas leakage exits the compressor through piping, where it is then either sent to a flare system or to some other recovery system. Flow or pressure measurement in this line is used as an indicator of the health condition of the DGS. Static Seals Static seals are responsible for sealing pressure between close clearance stationary elements. These seals are often composed o-ring of various elastomers in lowto medium-pressure applications and Teflon spring energized seals in higher pressure applications. Stationary seals will be discussed further in the design of casings; however, with proper seal design and proper material compatibility for temperature and gas composition, this is not a major technical hurdle.

46 SECTION

II Types of Equipment

FIG. 3.15 Dry gas seal supply and vent ports.

Impellers

Radial bearing

Hub attachment

Radial bearing

Thrust bearing runner

FIG. 3.16 Impeller shaft assembly. (Based on TRIZ, 2018, http://triz-ltd.com/en/produktsiya/ rotora/.)

Shaft The rotating assembly will typically consist of a shaft with mounted impellers, sleeves, seal components, thrust disk, and a coupling hub as shown in Fig. 3.16. The shaft is sized to safely transmit torque from the coupling to the impellers, and if necessary, to additional bodies attached at the nondrive end. Shaft design further considers torsional and lateral rotordynamics, as well as material selection and processing for environmental conditions. Most shafts are manufactured from a single piece of forged alloy steel or bar stock that is heat treated and stress relieved. Typical materials include 4140 and 4340 steels. Unique applications require different grades of steel. Ultralow

Centrifugal Compressors Chapter

3

47

temperatures below 115°C (175°F) require the use of high nickel steels for their low-temperature toughness. Highly corrosive environments require the use of stainless steels. Another design method uses a stacked rotor assembly, where the impeller disks form part of the rotor that is held together to shaft ends by a tie bolt. Shaft journals and shaft probe areas have special requirements with respect to their manufacture in order to limit mechanical and electrical runout. The surface finish at the sensing areas observed by radial vibration probes is held within 1.0 μm (32 μin.) Ra by honing or burnishing. The probe sensing area is then demagnetized to required levels.

Bearings Proper rotor position is maintained by journal and thrust bearings. Two journal bearings are used for radial positioning of the compressor rotor supporting the gravitational load of the shaft as well as various dynamic forces caused by rotor unbalance, misalignment, and other sources. Thrust bearings are used for axial positioning of the compressor rotor supporting thrust loads that arise from gas forces within the compressor case. The arrangement of the bearings is as shown in Fig. 3.17 such that the bearings are external to the gas flow path outboard of the DGSs. This keeps lubricants out of the process gas and process gas out of the lubricants. This arrangement also allows for relatively easy access for maintenance such as bearing and DGS replacement. Bearings used in centrifugal compressors can be hydrodynamic, magnetic, or rolling element. Hydrodynamic bearings are the most prevalent, whereas magnetic bearings are popular in niche applications such as pipeline compressors and hermetically sealed subsea compressors. Hydrodynamic bearings are highly advantageous because they suffer little or no wear due to the formation of a hydrodynamic wedge of oil that separates

FIG. 3.17 Bearing arrangement with bottom halves installed. (Courtesy of Elliott Group.)

48 SECTION

II Types of Equipment

the rotating journal from the stationary bearing. They have exceptionally long life thereby enabling long periods of continuous operation, often in excess of 5 years. In addition, these bearings possess dynamic characteristics that allow for vibration control thereby enabling high-speed operation and traverse of rotor critical speeds. They offer very good load capacity in a reasonably compact design envelope, and they accommodate transient loads very well. This leads to a robust rotordynamic characteristics. The primary disadvantages are relatively high-power loss compared to other bearings and need for a reliable oil supply system. Hydrodynamic journal bearings can be either tilting pad or fixed geometry design. Almost all centrifugal compressors are equipped with tilting pad journal bearings. Fixed geometry designs (plain, elliptical, lobed, pressure dam, and others) are seldom encountered in compressors constructed after about 1970. Fig. 3.18 shows a tilting pad journal bearing with self-aligning pivots and chromium copper pads for temperature reduction. Tilting pad journal bearing designs consist of several pads arranged in an annular ring around the shaft with the pads free to tilt about their respective pivots. Tilting pad journal bearings may include several design variations such as self-aligning features to compensate for misalignment, specialized materials, and special oil feed and drain configurations for reducing temperature and power loss. Fig. 3.19 shows a tilting pad journal bearing with directed lube features with oil introduced close to the pad inlet. One particular advantage of tilting pad journal bearings is their dynamic characteristics and inherent resistance to rotordynamic instability, which enables passive control of vibration while traversing the rotor’s critical speed and further enables reliable stable operation at speeds well above the first critical speed. Hydrodynamic thrust bearings are usually of the self-equalizing tilting pad design. Fixed geometry designs (plain, tapered land, and compound tapered

FIG. 3.18 Tilting pad journal bearing with self-aligning spherical seat. (Courtesy of Elliott Group.)

Centrifugal Compressors Chapter

3

49

FIG. 3.19 Tilting pad journal bearing with directed lubrication. (Courtesy of Waukesha, Barry Blair.)

FIG. 3.20 Self-equalizing tilting pad thrust bearing. (Courtesy of Kingsbury, Scan DeCamillo.)

land) are seldom encountered in centrifugal compressors constructed after about 1970. Fig. 3.20 shows a self-equalizing tilting pad thrust bearing. This type of thrust bearing has a linkage construction that will compensate for misalignment. The bearing shown operates in an oil-flooded cavity with oil exiting primarily through a top tangential drain and secondarily through shaft oil seals. Fig. 3.21 shows a bearing for the same envelope with directed lube features consisting of oil feeds going directly to the pads. These bearings are operated within an evacuated cavity with oil drained from the bottom. The directed lube design with evacuated cavity reduces thrust bearing temperatures and power consumption.

50 SECTION

II Types of Equipment

FIG. 3.21 Self-equalizing tilting pad thrust bearing with directed lubrication. (Courtesy of Kingsbury, Scan DeCamillo.)

Thrust bearings are positioned in a double-acting arrangement as shown in Fig. 3.22 to control thrust loads in both directions. When considering a compressor’s entire operational map, the thrust direction can reverse. The most severe condition would typically involve the loss of either a balance piston seal or center seal, at which point the bearing must survive until the unit is shutdown. Bearing surfaces consist of a soft metal bonded to a hard metal backing. The soft metal surface is most often an ASTM B23 grade 2 babbitt comprised of 89% tin alloyed with other metals. Babbitt provides a good bearing surface that protects the shaft from damage. It is good for embedding hard contaminant particles and for resistance to seizure and galling. The main disadvantage of babbitt bearing materials is their relatively low compressive, tensile, and fatigue strengths especially at high temperature. The babbitt surface material is cast and bonded as a thin layer to a hard metal backing. The thin babbitt layer is typically less than 1 mm thick and the hard metal backing is typically steel or chromium copper. Steel is the most prevalent and least expensive backing material. Chromium copper is used for its superior thermal conductivity enabling reduced bearing metal temperature.

Aerothermodynamics Euler’s Equations Earlier in this chapter, we talked about the impeller which “impart[s] mechanical energy to the gas” and the diffuser, where “part of the velocity is converted into static pressure.” In this section, we describe in more detail how this works. The Bernoulli’s law (which is strictly true only for incompressible flows, but which can be modified for the subsonic compressible flows we find in gas

Centrifugal Compressors Chapter

Oil inlet

3

51

Oil inlet X

2

X

Filler plate or shim optional

80% of X FIG. 3.22 Self-equalizing tilting pad thrust bearing with directed lubrication. (Courtesy of Kingsbury, page 50 of “LEG Bearings, Thrust and Journal,” Kingsbury, Inc.)

compressors) describes the interchangeability of two forms of energy: static pressure and velocity. The incompressible formulation of the Bernoulli’s law for a frictionless, stationary, and adiabatic flow without any work input is: ρ pt ¼ p + c2 ¼ const 2 For compressible flows, the equation becomes ht ¼ h +

c2 ¼ const 2

52 SECTION

II Types of Equipment

Another requirement is, that mass cannot appear or disappear, thus for any flow from a point 1 to a point 2: m_ 1 ¼ ρ1 Q1 ¼ m_ 2 ¼ ρ2 Q2 ρQ¼ρcA This requirement is valid for compressible and incompressible flows, with the caveat that for compressible flows the density is a function of pressure and temperatures, and thus ultimately a function of the velocity. These two concepts explain the working principles of the vanes and diffusers used. Due to the requirement for mass conservation, any flow channel that has a wider flow area at its inlet and a smaller flow area at its exit will require a velocity increase from inlet to exit. If no energy is introduced to the system, the Bernoulli’s law requires a drop in static pressure (Fig. 3.23A). Examples for flow channels like this are turbine blades and nozzles, inlet vanes in compressors, and others (Fig. 3.23B). Conversely, any flow channel that has a smaller flow area A at its inlet and a larger flow area at its exit will require a velocity decrease from inlet to exit. If no energy is introduced to the system, the Bernoulli’s law requires an increase in static pressure (Fig. 3.23C). Examples for flow channels like this are vaned or vaneless diffusers, flow channels in impellers, rotor and stator blades of axial compressors volutes, and others (Fig. 3.23D).

Velocity Pressure Mass flow

Velocity Pressure Mass flow

(A)

Compressor blades

(B)

Turbine blades

Velocity Pressure Mass flow

(C) FIG. 3.23 Acceleration and diffusion [3].

(D)

Velocity Pressure Mass flow

Centrifugal Compressors Chapter

3

53

If these flow channels are in a rotating system (e.g., in an impeller), mechanical energy is added to or removed from the system. Nevertheless, if the velocities are considered in a rotating system of coordinates, above principles are applicable as well. Another important concept is the conservation of momentum. The change in momentum M of gas flowing from a point 1 to a point 2 is its mass times its velocity (m c), and is also the sum of all forces F acting. The change in momentum is thus: !


 ! dM ! ! ¼ m_ c 2  c 1 ¼F dt To change the momentum of this gas, either by changing the velocity or the direction of the gas (or both), a force is necessary. Fig. 3.24 outlines this concept for the case of a bent, conical pipe. The gas flows in through the area A1 with w1, p1, and out through the flow area A2 with w2, p2. The differences in the force is due to the pressure (p1A1 and p2A2, respectively), and the fact that a certain mass flow of gas is forced to change its direction generates a reaction force FR. Split into x and y coordinates, and considering that m_ ¼ ρ1 A1 w1 ¼ ρ2 A2 w2 we get (due to the choice of coordinates, w1y ¼ 0) x : ρA1 w1 ðw2x  w1 Þ ¼ p1 A1  ðp2 A2 Þx + FRx   y : ρA1 w1 w2y ¼ ðp2 A2 Þy + FRy

p

2

W2 2

y x

A2

1

W1

U1 p1

A1

2

Fy 2

(p2 + QW 2)A2 FIG. 3.24 Conservation of momentum [3].

Fx F

(p1 + QW 1) A1

54 SECTION

II Types of Equipment

It should also be noted that this formulation is also valid for viscous flows, because the friction forces become internal forces. All these concepts are applied in a very similar way in pipeline flows. For a rotating row of vanes in order to change the velocity of the gas, the vanes have to exert a force on the gas. This is fundamentally the same force that FRy that acts in the previous example for the pipe. This force has to act in the direction of the circumferential rotation of the vanes in order to do work on the gas. According to the conservation of momentum, the force that the blades exert is balanced by the change in circumferential velocity times the associated mass of the gas. This relationship is often referred to as the Euler’s law: P ¼ m_  Δh ¼ m_  ðu2 cu2  u1 cu1 Þ where u is the circumferential blade velocity at the inlet (1) and exit (2) of the impeller, and cu is the circumferential component of the gas velocity, taken in an absolute reference frame at the inlet (1) and exit (2) (Fig. 3.25). At this point, one of the advantages of centrifugal compressors over axial compressors becomes apparent. In the axial compressor, the entire energy transfer has to come from the turning of the flow imposed by the blade (cu2  cu1), while the centrifugal compressor has added support from the centrifugal forces on

FIG. 3.25 Velocity vectors in a centrifugal impeller [3].

Centrifugal Compressors Chapter

3

55

the gas while flowing from the diameter at the impeller inlet (u1 ¼ πDiN) to the higher diameter at the impeller exit (u2 ¼ πDtipN). The importance of the Euler’s law lies in the fact that it connects aerodynamic considerations (i.e., the velocities involved) with the thermodynamics of the compression process.

Operating Regimes of a Centrifugal Compressor The general behavior of any gas compressor can be gauged by some additional, fundamental relationships: the vanes of the rotating impeller “see” the gas in a coordinate system that rotates with the impeller. The transformation of velocity coordinates from an absolute frame of reference (c) to the frame of reference rotating with a velocity u is by !

!

!

w¼ c  u

where, for any diameter D of the impeller u ¼ πDN. The impeller exit geometry (“backsweep”) determines the direction of the relative velocity w2 at the impeller exit. The basic “ideal“ slope of head versus flow is dictated by the kinematic flow relationship of the compressor, in particular the amount of backsweep of the impeller. Any (Fig. 3.25) increase in flow at constant speed causes a reduction of the circumferential component of the absolute exit velocity (cu2). It follows from the Euler’s equation above, that this causes a reduction in head. Adding the influence of various losses to this basic relationship shape the head-flow-efficiency characteristic of a compressor (Fig. 3.26): whenever the flow deviates from the flow the stage was designed for, the components of the stage operate less efficiently. This is the reason for Head and loss Ideal head

Incidence loss

Best efficiency point

Isentropic head A

Friction loss Flow FIG. 3.26 Head versus flow relationship at constant speed [3].

56 SECTION

II Types of Equipment

(A)

(C)

(B)

(D)

FIG. 3.27 Unseparated (A, B), partially separated (C), and fully separated (D) flow over an airfoil at increasing angle of attack [33a].

incidence losses. Fig. 3.27 illustrates this, using an airfoil as an example: at the “design flow,“ the air follows the contours of the airfoil. If we change the direction of the incoming air, we see increasing zones where the airflow ceases to follow the contours of the airfoil, and create increasing losses. Furthermore, the higher the flow, the higher the velocities, and thus the friction losses. A compressor, operated at constant speed, is operated at its best efficiency point (Fig. 3.26). If we reduce the flow through the compressor (e.g., because the discharge pressure that the compressor has to overcome is increased), then the compressor efficiency will be gradually reduced. At a certain flow, stall, probably in the form of rotating stall, in one or more of the compressor components will occur. At further flow reduction, the compressor will eventually reach its stability limit, and go into surge. If, again starting from the best efficiency point, the flow is increased, then we also see a reduction in efficiency, accompanied by a reduction in head. Eventually, the head and efficiency will drop steeply, until the compressor will not produce any head at all. This operating scenario is called choke. (For practical applications, the compressor is usually considered to be in choke when the head falls below a certain percentage of the head at the best efficiency point.) Surge At flows lower than the flow at the stability limit, practical operation of the compressor is not possible. At flows to the left of the stability limit, the

Centrifugal Compressors Chapter

3

57

compressor cannot produce the same head as at the stability limit. It is therefore no longer able to overcome the pressure differential between suction and discharge side. Because the gas volumes upstream (at discharge pressure) is now at a higher pressure than the compressor can achieve, the gas will follow its natural tendency to flow from the higher to the lower pressure. The flow through the compressor is reversed. Due to the flow reversal, the system pressure at the discharge side will be reduced overtime, and eventually the compressor will be able to overcome the pressure on the discharge side again. If no corrective action is taken, the compressor will again operate to the left of the stability limit and the above described cycle is repeated: the compressor is in surge. The observer will detect strong oscillations of pressure and flow in the compression system. It must be noted that the violence and the onset of surge are a function of the interaction between the compressor and the piping system. Stall If the flow through a compressor at constant speed is reduced, the losses in all aerodynamic components will increase. Eventually, the flow in one of the aerodynamic components, usually in the diffuser, but sometimes in the impeller inlet, will separate (the last picture in Fig. 3.27 shows such a flow separation for an airfoil). It should be noted that stall usually appears in one stage of a compressor first. Flow separation in a vaneless diffuser means, that all or parts of the flow will not exit the diffuser on its discharge end, but will form areas where the flow stagnates or reverses its direction back to the inlet of the diffuser (i.e., the impeller exit; Fig. 3.27). Stall in the impeller inlet or a vaned diffuser is due to the fact, that the direction of the incoming flow (relative to the rotating impeller changes with the flow rate through the compressor). Usually, vanes in the diffuser reduce the operating range of a stage compared to a vaneless diffuser. Therefore, a reduction in flow will lead to an increased mismatch between the direction of the incoming flow the impeller was designed for and the actual direction of the incoming flow. At one point, this mismatch becomes so significant that the flow through the impeller breaks down. Flow separation can take on the characteristics of a rotating stall. When the flow through the compressor stage is reduced, parts of the diffuser experience flow separations. Rotating stall occurs if the regions of flow separation are not stationary, but move in the direction of the rotating impeller (typically at 15%– 30% of the impeller speed). Rotating stall can often be detected from increasing vibration signatures in the subsynchronous region. Onset of stall does not necessarily constitute an operating limit of the compressor. In fact, in many cases, the flow can be reduced further before the actual stability limit is reached.

58 SECTION

II Types of Equipment

Choke At high flow, the head and efficiency will drop steeply, until the compressor will not produce any head at all. This operating scenario is called choke. However, for practical applications, the compressor is usually considered to be in choke when the head falls below a certain percentage of the head at the best efficiency point. Some compressor manufacturers do not allow operation of their machines in deep choke. In these cases, the compressor map has a distinct high flow limit for each speed line. The efficiency starts to drop off at higher flows, because a higher flow causes higher internal velocities, and thus higher friction losses. The head reduction is a result of both the increased losses and the basic kinematic relationships in a centrifugal compressor: even without any losses, a compressor with backwards bent blades (as they are used in virtually every industrial centrifugal compressor) will experience a reduction in head with increased flow (Fig. 3.26). “Choke” and “Stonewall” are different terms for the same phenomenon.

Compressor Control The interaction between a compressor and a compression system, in conjunction with control mechanisms and the compressor characteristic determine the operating point of the compressor in a given situation. Control mechanisms can be: l l l l l l

power input from driver, compressor speed, guide vane settings, compressor suction or discharge pressure set by a throttle, compressor throughput modified by recycling gas, and multiple units.

External process objectives can be minimum suction pressure, maximum discharge pressure, or delivered flow. Compressor operational boundaries include surge, minimum speed, maximum speed, and in some instances minimum pressure rise (choke). The operating envelope of a centrifugal compressor is limited by the maximum allowable speed (or, for other control means, the maximum guide vane angle), the minimum flow (surge flow), and the maximum flow (choke or stonewall), and the minimum speed (Fig. 3.28). Another limiting factor may be the available driver power. Only the minimum flow requires special attention, because it is defined by an aerodynamic stability limit of the compressor. Crossing this limit to lower flows will cause a flow reversal in the compressor, which can damage the compressor. Modern control systems prevent this situation by automatically

Centrifugal Compressors Chapter

3

59

FIG. 3.28 Typical pipeline operating points plotted on a typical centrifugal compressor performance map.

opening a recycle valve. For this reason, virtually all modern compressor installations use a recycle line with a control valve that allows the flow through the compressor to increase if it comes near the stability limit. The control system constantly monitors the operating point of the compressor in relation to its surge line, and automatically opens or closes the recycle valve if necessary. For most applications, the operating mode with open, or partially open recycle valve is only used for start-up and shutdown, or for brief periods during upset operating conditions. Of importance for compressor control is the fact that the operating map is continuous, that is, there are no areas within the defined boundaries that are off-limits for operation. The only exception is found in some electric motordriven trains, where the variable frequency drives (VFDs) may induce unacceptably high torsional vibrations at certain speeds [4].

Variable Speed Compressor drivers that can operate at variable speed (two-shaft gas turbines, steam turbines, turbo expanders, and electric motors with VFDs or variable speed gearboxes) allow the compressor to operate over a range of different speeds. The faster the compressor runs, the more head and flow it generates, and the more power it consumes. The efficiency characteristics of the compressor are retained for different speeds, so this is a very efficient way of adjusting the compressor to a wide range of different operating conditions. Figs. 3.28 and 3.29 show the resulting map.

60 SECTION

II Types of Equipment

FIG. 3.29 Control methods for centrifugal compressors: throttling, variable speed, and adjustable guide vanes. (P.C. Rasmussen, R. Kurz, Centrifugal compressor applications-upstream and midstream, in: Proc. 38th Turbomachinery Symposium, Houston, TX, 2009.)

Centrifugal Compressors Chapter

3

61

Adjustable Inlet Vanes Modifying the swirl of the flow into the impeller allows to modify the operating characteristics of the stage (Fig. 3.29). This can be accomplished by adjustable vanes upstream of the impeller. Increasing the swirl against the rotation of the impeller increases the head and flow through the stage. Increasing the swirl with the rotation of the impeller reduces the head and flow through the stage. This is very effective to increase the range for a single stage. In multistage compressors, the range increase is limited if only the first stage has adjustable vanes. The technical difficulty for high-pressure compressors lies in the fact that complicated mechanical linkage has to be actuated from outside the pressure containing body. Adjustable Diffusor Vanes Vaned diffusers tend to limit the operating range of the compressor because the vanes are subject to increased incidence at off-design conditions, thus eventually causing stall. Adjustable diffuser vanes allow adjustments for the changing flow conditions, thus effectively allowing for operation at much lower flows by delaying the onset of diffuser stall (Fig. 3.29). They will not increase the head or flow capability of the stage. In multistage compressors, the range increase is limited if only one stage has adjustable vanes. The technical difficulty for high-pressure compressors lies in the fact that complicated mechanical linkage has to be actuated from outside the pressure containing body. Another issue is that for the vanes to operate, small gaps between the vanes and the diffuser walls have to exist. Ubiquitous leakage through these gaps causes efficiency and range penalties, in particular in machines with narrow diffusers. Throttling (Suction and Discharge) A throttle valve on the suction or discharge side of the compressor increases the pressure ratio the compressor sees, and therefore, moves the operating point to lower flows on the constant speed map. It is a very effective, but inefficient way of controlling compressors (Fig. 3.29). Recycling A controlled recycle loop allows a certain amount of the process flow to go from the compressor discharge back to compressor suction. The compressor therefore sees a flow that is higher than the process flow. This is a very effective, but inefficient way to allow the compressor system to operate at a low flow. Process Control With Centrifugal Compressors Driven by Two-Shaft Gas Turbines The following is a description of a typical control scenario, in this particular case for compressors with a gas turbine driver that can operate at variable speeds.

62 SECTION

II Types of Equipment

Centrifugal compressors, when driven by two-shaft gas turbines, are usually adapted to varying process conditions by means of speed control. This is a very elegant way of controlling a system, because both the centrifugal compressor and the power turbine of a two-shaft gas turbine can operate over a wide range of speeds without any adverse effects. A typical configuration can operate down to 50% of its maximum continuous speed, and in many cases even lower. Reaction times are very fast, thus allowing a continuous load following using modern, Programmable Logic Control (PLC)-based controllers. A simple case is flow control: the flow into the machine is sensed by a flow metering element (such as a flow orifice, a Venturi nozzle, or an ultrasonic device). A flow set point is selected by the operator. If the discharge pressure increases due to process changes, the controller will increase the fuel flow into the gas turbine. As a result, the power turbine will produce more power and cause the power turbine, together with the driven compressor, to accelerate. Thus, the compressor flow is kept constant. Both the power turbine speed and the power increase in that situation. If the discharge pressure is reduced, or the suction pressure is increased due to process changes, the controller will reduce the fuel flow into the gas turbine. As a result, the power turbine will produce less power and cause the power turbine, together with the driven compressor, to decelerate. Consequently, the compressor flow is kept constant. Similar control mechanisms are available to keep the discharge pressure constant, or to keep the suction pressure constant. Another possible control mode is to run the unit at maximum available driver power (or any other constant driver output). In this case, the operating points are all on a line of constant power, but the speed will vary. The control scheme works for one or more compressors, and can be set up for machines operating in series, as well as, in parallel. If speed control is not available, the compressor can be equipped with a suction throttle, or with variable guide vanes. The latter, if available in front of each impeller is rather effective, but the mechanical complexity proves usually to be prohibitive in higher-pressure applications. The former is a mechanically simple means of control, but it has a detrimental effect on the overall efficiency.

Multiple Units If a compressor station operates multiple compressors, either identical units or dissimilar units, turning these units on or off provides a powerful means of control. For particular applications that see large swings in load, this control method allows to operate the drivers near their best efficiency conditions for most of time. It also has a positive impact on maintenance costs, since the drivers are only operated when needed. The method works well if the selected equipment has a high starting reliability. It is also advantageous if the compressors can

Centrifugal Compressors Chapter

3

63

stay pressurized and can be restarted without having to vent gas into the atmosphere. Modern centrifugal compressors with DGSs are capable of this “pressurized hold.”

Structural Pressure Vessel The compressor casing along with its nozzles acts as a pressure vessel that is designed to seal in the process gas without fugitive emissions. Controlled leakage consists of properly working end seals where the emissions are captured. Fugitive emissions consist of leakage across joints, flanges, endwalls, and uncontrolled emissions across end seals. The compressor casing serves many additional purposes acting to support all other components of the compressor including diaphragms and endwalls. Bearing housings are attached to either the endwall or the casing. The bearing housings in turn support the bearings and the rotor. The casing main inlet and discharge connections are flanged. These main inlet and discharge flange connections typically coincide with ASME/ANSI B16.1, B16.5, B16.42, or B16.47 series A or B as specified within the API 617 standard. Raised face (RF) flange designs are commonly used with either a flat gasket Fig. 3.30 or spiral wound gasket typically with a solid inner ring to prevent spiral wound gasket extrusion. Ring type joints (RTJs) flanges are also used and are often preferred for their ease of assembly. Pressures in excess of the standard ratings require more specialized piping connections such as API 6A design or Grayloc fittings. All flanges are designed to handle the hydrotest pressure. Many different casing materials have been applied throughout the years including ductile iron, cast iron, and cast, forged or fabricated steels. Carbon steel is commonly used for most simple applications as fabrication expenses are less than for stainless varieties. On the other hand, corrosive environments

FIG. 3.30 Flange and gasket [5].

64 SECTION

II Types of Equipment

require the use of stainless steels and sometimes exotic claddings. Lowtemperature service requires the use of metals with good low-temperature toughness such as nickel alloy steels, or in the case of ultralow-temperature service, austenitic stainless steels. API 617 defines low-temperature service as having metal temperature below 29°C (20°F), impact testing is required by the API 617 standard for materials used in such low-temperature pressure containing components including nozzles, flanges, weldments, and casing. The material selection and processing are based on meeting the service requirements and economics. FEA is often used to analyze and design the casing. In this case, the analysis is performed in accordance with Section VIII, Division 2 of the ASME Code. The casing is hydrotested at 150% of the maximum allowable working pressure, which is based on the customer provided relief valve setting or is determined as 125% of the maximum discharge pressure. The limiting factor for horizontally split casing designs is typically deflection that could result in a failed hydrotest (Fig. 3.31). The limiting factor for vertically split casing designs is typically stress under hydrotest conditions, not just at the barrel, but also at the endwall and in the DGS cavity. This is also true for extrusion gaps in barrel-style compressors of very high pressure. In these applications, meeting the required extrusion gap requires substantially thicker casing walls than the allowable stresses in the cross sections.

Lateral Rotordynamics Lateral rotordynamics of centrifugal compressors must be considered to ensure acceptable lateral vibration levels and long-term machine reliability. High levels of lateral rotor vibration can cause operational issues, unplanned shutdowns, or even a lack of operability. The most common lateral rotordynamic issues are often related to critical speeds (synchronous excitation of a natural frequency) and rotordynamic stability (subsynchronous vibration). This section discusses lateral rotordynamic modeling and analysis for a centrifugal compressor along with acceptance criteria. The discussion here is limited to a basic discussion of lateral rotordynamics for centrifugal compressors. Numerous books and other material have been written on the topic of rotordynamics and machinery vibration [6, 7, 7a].

FIG. 3.31 Casing hydrotest FEA showing potential split-line leakage.

Centrifugal Compressors Chapter

3

65

Rotordynamic Modeling and Analysis Rotordynamic modeling begins by modeling the rotor with beam or threedimensional (3D) finite elements. The compressor rotor, or shaft, is modeled as a series of smaller segments, or elements, that appropriately capture the diameter steps and attachment points for all added mass components. A node or station is the start or end of each element. Components like impellers, couplings half-weights, DGSs, and thrust disks generally do not add shaft stiffness, and are modeled as added mass components. Fig. 3.32 shows an example centrifugal compressor (10 stage, back-to-back design) and the corresponding beam element rotor model. The fluid-film journal bearings support the rotor through hydrodynamic lubrication. As a shaft rotates, the viscous lubricant creates fluid pressure between the bearing surface and rotor that provides lift, thereby supporting the rotor. Computerized bearing codes are used to predict this fluid film pressure profile and further provide the rotordynamic coefficients, or stiffness and damping. In summary, the fluid film forces acting between a bearing and rotor can be modeled as a system of springs (representing stiffness) and dampers (representing damping). These rotordynamic coefficients are a function of several parameters, including bearing type (plain journal, multilobe, pressure dam, or tilting pad), bearing geometry (including clearance and pad preload), oil properties, loading, and rotational speed. The stiffness and damping forces from the journal bearings are calculated and applied to the rotor model in the lateral rotordynamic analysis. The appropriate ranges for bearing clearance and oil inlet temperature (not just the nominal values) should also be considered in the bearing analysis. Casing or foundation support stiffness is often considerably higher than the journal bearings stiffness for most centrifugal compressors, thus allowing casing or foundation to be ignored. The casing and foundation are often important on industrial gas turbines, which may require a finite element casing model coupled to the rotor-bearing model. Casing and foundation flexibility can have a great effect on the rotor response, decreasing the critical speeds and increasing

FIG. 3.32 Example of centrifugal compressor and corresponding rotor model.

66 SECTION

II Types of Equipment

amplification factors (AFs). According to the API 617 [8], if the foundation flexibility is less than 3.5 times the bearing stiffness, then a foundation model should be included. An unbalance response analysis is used to determine the rotor vibration amplitude from start-up through the maximum continuous speed and all the way up to 150% of the maximum continuous speed. The magnitude and placement (location) of the unbalance will affect the rotor response. For centrifugal compressors, the common unbalance amount is typically the amount described by API, 4*W/N oz-in, where W is the rotor weight in pounds and N is the rotational speed in rpm. In general, the unbalance placement is oriented with respect to the points of maximum amplitude for the subject mode shape. API 617 and 684 offer recommendations for unbalance configurations for various example mode shapes. The unbalance response calculation produces amplitude and phase-response plots. Critical speeds are identified where the system’s sensitivity to unbalance excitation is maximized. Fig. 3.33 shows an example unbalance amplitude response plots for the first and second critical speeds for the subject example rotor. The first critical speed is well excited by mid-span unbalance, meeting API separation margin (SM) requirements. The second critical speed for this example is excited by quarter-span unbalance, with the unbalance amounts placed 180 degrees out-of-phase. Damping notably increases the second critical speed as compared to an undamped analysis. In addition, no SM is required for the second critical speed since the AF is below 2.5. Fig. 3.34 describes the API SMs and AF calculations for rotor unbalance response. It is noted that SM requirements defined by API are a function of AF, and are provided in equation form in 617 and 684 [7]. For a lightly damped response, the AF will be higher and require a larger SM. When damping increases, the AF is lowered and the SM requirement decreases. API considers any AF below 2.5 to be well damped and does not require a SM. As part of the unbalance response calculation, the response amplitude at the seals (in addition to the bearings) should be considered. As per API 617 [8], the major axis of the rotor response orbit cannot exceed 75% of the design seal clearances. In order to verify that a rotor design meets these requirements, the rotor orbit should be calculated at the annular seal locations, at the speed of maximum displacement (i.e., the first critical speed). Bearing eccentricity should be included for this seal clearance check. In addition, rotor sag should be considered, especially for larger compressors, such as those in liquefied natural gas (LNG) applications, where the rotor static deflections due to gravity can become notable. Rotordynamic stability is assessed through the calculation of damped eigenvalues (or damped natural frequencies) and associated log dec values. For centrifugal compressors, it is necessary to include potential destabilizing forces from all close-clearance components, including journal bearings, annular seals (including labyrinth and damper seals), and impellers.

40

Operating speed

30

NC1 = 4060 rpm AF1 = 5.84 First mode

20 10 0 0

2000

4000

6000

8000 10,000 12,000 14,000 16,000 18,000 20,000

Rotor speed (rpm)

Rotordynamic response plot Response (microns pk-pk)

First critical speed

50 45

Second critical speed

40

Operating speed

35 30

NC2 = 15,000 rpm AF2 = 2.05

Second mode

25 20 15 10 5 0 0

10,000

5000

Rotor speed (rpm)

FIG. 3.33 Unbalance amplitude response plot examples for the first and second mode unbalance configurations.

15,000

20,000

Centrifugal Compressors Chapter

Response (microns pk-pk)

Rotordynamic response plot

50

3

67

68 SECTION

II Types of Equipment

FIG. 3.34 API response showing separation margins and amplification factor calculation for centrifugal compressors [8].

In regards to the bearings, the plain journal bearings are the least stable design. Elliptical, axial groove, and multilobe journal bearings introduce “preload” that improves the stability. Tilting-pad journal bearings possess essentially no cross-coupling and are the most stable bearing design, often used for high-speed centrifugal compressors. Titling-pad journal bearings are necessary when operating at speeds well above the first critical speed. In addition, tilting-pad bearings have several parameters, including clearance, preload, pad offset, pad orientation, etc. that can be adjusted to achieve the desired stiffness and damping. The various annular seals within the compressor are a source of potential destabilization. Much like the bearings, seal forces are modeled as a system

Centrifugal Compressors Chapter

3

69

of springs (representing stiffness) and dampers (representing damping). Seal rotordynamic characteristics are impacted by a number of parameters, including seal type, geometry, location, and gas conditions. Labyrinth hub seals generally have a small pressure differential and can generally be neglected from a rotordynamic analysis. Impeller shroud or eye seals are generally labyrinth seals, and impact rotordynamic stability. Balance piston or division-wall seals generally have the largest rotordynamic impact because of the large pressure differential and long sealing length. For centrifugal compressors, the balance piston or division wall seal can be a labyrinth or a honeycomb/hole-pattern seal. The destabilizing effect of annular seals is largely a function of the swirl or bulk circumferential velocity of the flow entering the seal. Centrifugal impellers by design impose high circumferential swirl at their exit. The flow field in this region communicates with the secondary passages feeding swirl to the various annular seals throughout the machine. Fluid swirl in the direction of rotation is the primary destabilizing mechanism in annular seals. This circumferential flow entering the seal can be reduced with de-swirl methods, including swirl brakes or shunt injection [9]. Swirl brakes (or antiswirl vanes) include geometry, typically consisting of radial vanes oriented in the axial direction to impede circumferential flow entering the seals. Swirl brakes have little effect on aerodynamic efficiency and can be designed integral to the seal material; however, they can increase leakage by reducing the swirl in secondary flow cavities. Swirl brakes can be applied to interstage and balance piston seals, as well as shroud cavities. Shunt injection, in contrast, bleeds off higher pressure gas and feeds a reverse swirl annulus near the entrance to the seal, hindering the swirling flow. Shunt injection is limited to the balance piston seal and increases the secondary leakage, thus lowering compressor efficiency. Successfully deswirling the inlet flow turns annular seals into stabilizing components, increasing the system log dec values. Labyrinth seal rotordynamic coefficients can be calculated via bulk flow methods (such as assuming a single control volume between labyrinth teeth) or with CFDs methods. It is noted that the bulk flow seal calculations are the most common for centrifugal compressors compared to CFD analyses. Damper seals such as honeycomb or hole-pattern seals can be used as the balance piston or division-wall seals to provide significant stiffness and damping to improve rotor stability [10]. These seals provide substantial damping while generally sacrificing leakage compared to traditional labyrinth seals. Damping increases with increasing seal length as well as increasing pressure differential. The rotordynamic coefficients of honeycomb or hole-pattern seals are frequency dependent and should be properly accounted for in the stability analysis. In addition to notable damping, the direct stiffness of damper seals can become large enough to influence critical speeds. In addition, the stabilizing effect of honeycomb or hole-pattern seals is significantly influenced by the seal clearance taper (inlet seal clearance relative to exit clearance). While more damping can be achieved with a slightly diverging taper compared to a

70 SECTION

II Types of Equipment

converging taper, a converging taper is recommended as the conservative approach. While the direct damping increases as seal clearance divergence increases (more diverging, more damping), a cliff is reached where the damping suddenly becomes negative and the seal becomes destabilizing. In addition, the seal clearance taper can be largely influenced by the seal carrier structure. Deflection of the seal carrier or support structure can impact the taper, thus influencing the available damping and resulting stability. Overall, seal taper and carrier deflection should be considered in a rotordynamic analysis for all honeycomb or hole-pattern seals. Centrifugal impellers generate destabilizing aerodynamic forces. A widely applied approach for predicting the bulk impeller aerodynamic cross-coupling coefficients including eye and shaft seal effects is the Wachel method [11]. This method was derived when the phenomenon of instability in high-pressure centrifugal compressors first started to cause problems for oil and gas applications. The original Wachel equation for cross-coupling is shown in Fig. 3.35. For compressors with a single suction and no side streams, this formula may be collapsed to a single stream. API 617 provides a slightly different formulation of the Wachel number known as a modified Wachel’s equation. It differs from the original Wachel formulation in two ways. First, the mole weight variable is replaced with a constant value of 30. Second, the density ratio is based on a particular impeller stage, not the density ratio across a section. This density ratio difference can make a large difference in beam-style compressors as opposed to an overhung design. In addition, the CFD-based methods have been shown in literature for the calculation of impeller cross-coupling, including Moore et al. [12a]. While not as widely used as the Wachel-based equations, CFD methods offer a more

FIG. 3.35 Wachel equation and input parameters [12].

Centrifugal Compressors Chapter

3

71

physics-based approach to model the aerodynamic excitation. CFD-based methods generally require a CFD analysis for a representative impeller wheel type, which can then be applied to similar impeller designs. Rotordynamic stability is assessed by calculating the logarithmic decrement (log dec), and the sensitivity of the applied cross-coupling. It is noted that API 617 provides two levels of stability analysis: level I and level II. The level I method is considered a screening criterion that is valid within a certain zone of average gas density and critical speed ratios. Level II is required elsewhere. In a level I analysis, the API Wachel equation is used as a global destabilizing force value in the log dec calculation. Many, if not most, of today’s centrifugal compressors require an API level II stability analysis, where all destabilizing forces are to be included, including annular seals and impeller excitations. The API requirement is a minimum log dec of 0.1 when all destabilizing forces are considered. An example log dec stability plot is shown in Fig. 3.36. The log dec is plotted against the applied impeller cross-coupling for various levels of increased cross-coupling to show sensitivity. For this compressor example, the bearings-only rotor configuration was predicted to be unstable (negative log dec) at the nominal cross-coupling value (API Kxy). When including the seal effects, the log dec of this example increases substantially. Also, for the rotor with seal effects, a good stability margin (relative slope of the log dec line) is noted. This means that the rotor stability for this example is not very sensitive to cross-coupling with properly operating seals. It is noted that the seals are stabilizing for the example shown, and this is not always the case. Labyrinth seals, without any de-swirl mechanisms, are generally destabilizing and commonly used in centrifugal compressors. A proper rotordynamic analysis should account for the seal configuration for that particular rotor.

FIG. 3.36 Rotor stability plot.

72 SECTION

II Types of Equipment

Other Related Phenomena While unbalance response and rotordynamic stability are commonly covered in a lateral analysis, other phenomena can cause lateral vibration problems. These additional topics are listed here for reference and are not discussed in detail. Lateral response or vibration can be influenced by other bearing-related issues, including oil whirl within fixed arc journal bearings, and wiped/damaged journal bearing surfaces. Off-design compressor operation can also impact lateral vibration, including rotating stall and surge. In addition, turbulence in the flow field can cause a lateral vibration response through an unsteady aerodynamic excitation. Mechanical issues, such as looseness on the rotor, looseness on stator components, misalignment, rubs, and shaft cracks can also cause lateral vibration issues. In summary, lateral rotordynamics must be considered for all centrifugal compressors operating in the oil and gas industry, where modern analysis tools and techniques minimize the risk of encountering a critical speed or stability problem on new equipment. It is emphasized that a proper lateral rotordynamic analysis is performed during the design phase where modifications can be easily implemented, compared to costly trouble-shooting and retrofit solutions in the field.

Torsional Rotordynamics The avoidance of torsional dynamic problems can be difficult since nearly all machinery trains lack a torsional measurement system. Proper torsional analysis is therefore critically important. Torsional evaluations generally involve a steady-state analysis, preparation of interference diagram(s), forced response analyses, and transient analyses (if necessary and applicable). These analyses are often conducted for trains involving compressors, pumps, motors, turbines, or engines, and any associated rotating components such as gearbox shafts and couplings. The steady-state torsional analysis involves preparation of a torsional masselastic model derived from manufacturer provided drawings and mass elastic information for the equipment involved. The steady-state torsional natural frequencies (critical speeds) and mode shapes are then calculated. The mode shapes are plotted to graphically as shown in Fig. 3.37 to show the deflection associated with the torsional natural frequencies. This allows for investigation of controlling stiffness, and relative modal participation of the major inertias in the system. An interference diagram (Campbell diagram) similar to that shown in Fig. 3.38 is prepared to graphically display prevalent excitation energy orders versus speed and frequency. The specified operating speed range(s) are superimposed on the interference diagram, and any coincidences between calculated critical speeds and prevalent excitation energy in the system are identified. The likelihood of exciting the critical speeds involved in any such coincidences

Centrifugal Compressors Chapter

3

73

FIG. 3.37 Typical torsional mode shapes.

FIG. 3.38 Representative interference diagram (whole order excitation)—needs upgraded to more of an API avoidance range of 10%.

74 SECTION

II Types of Equipment

is assessed by studying the respective mode shapes and potential excitation mechanisms. The SM for a mode can be calculated as follows: SM ¼ |fe  fn |=fe such that SM is the SM, fe is the frequency of excitation, and fn is the natural frequency of concern. SMs of more than 10% are preferred or a force response analysis is required. A forced response analysis is usually conducted to determine anticipated stress levels in the shafting during normal loaded operation. For turbine drives, whole mechanical excitation orders are usually considered. Trains involving motors, slip, and/or VFD frequencies are normally evaluated in addition to whole mechanical orders. During the forced response calculations, stress concentration factors are developed to account for keyways, major diameter changes, fillet radii, etc. based on shaft geometry. The resultant intensified stress at each station in the model is compared against allowable stress recommendations to determine acceptability. Transient torsional analyses are usually conducted for trains involving electric motors. The transient events most often studied include start-up and short circuit events. Various torsional modes can be excited during these events, which can amplify the forced response stress levels occurring as the critical speeds are traversed, as illustrated in Fig. 3.39. The method of achieving acceptable torsional dynamics is generally limited to coupling selection and tuning, as this is the least costly modification. In rare cases, other modifications may be required such as to the shafting or to the speed range. Torsional vibration analysis demonstration case–synchronous motor startup

Stress (psi)

Shaft Section 7 Amplification Factor = 30 Effective Radius = 8.25 in Inner Radius = 0.00 Length = 27.30 in Speed Ratio = 1.0 Shear Modulus = 1.18 e+7 pai

18,000 14,000 10,000 6000 2000 –2000 –6000 –10,000 –14,000 –18,000

Maximum Shear Stress Low Cycle Failure Stress Shear Endurance Limit

Predicted Number of Tolerable Events: 4361

0

2

4

6 Time (s)

8

10

12

FIG. 3.39 Representative transient torsional results—synchronous motor start-up (SwRI results from the University of Virginia-ROMAC TORTRAN code).

Centrifugal Compressors Chapter

3

75

Blade and Disk Vibrations While impellers are primarily selected based on their aerodynamic performance, they also have mechanical limitations that must be considered in their selection. An impeller will experience centrifugal stress and deflection due to rotation. An impeller will also experience thrust force due to differential pressure, and vibratory force due to disturbances in the flow field. Stresses and deflections must be managed to remain within allowable values, which are determined by the design, material, and the method of manufacture. A review of traditional methods of impeller mechanical design have been presented by numerous authors including Kushner et al. [13, 14] and Singh et al. [15]. The following are typically examined: l l l l

mean stress, disk critical speeds, leading edge impeller blade resonance, and disk interaction resonance

Impeller mechanical design typically begins with an evaluation of the mean stress and bore growth. The maximum mean stress must remain within safe limits during an API 617 impeller overspeed test [8]. In addition, the impeller to shaft junction must safely transmit the required horsepower either through a key, hirth, curvic, or bore-to-shaft frictional contact. Once manufactured, the impeller is overspeed tested at 115% of the maximum continuous speed. After testing, critical dimensions such as the bore, eye seal, and outside diameter are reviewed for deviations that would take the part out of tolerance. Manufacturers have typically avoided disk critical speeds. The term disk critical speed was defined by Wilfred Campbell [16]. Disks vibrate in a pattern defined by nodal lines. The resonant operating speed is equal to the natural frequency divided by the number of diametrical nodal lines. When operating at a disk critical speed, the relative amplitude of the disk around the circumference in stationary coordinates would form a standing wave in stationary space thus being easily excited by pressure variations around the impeller. The modes of concern are from two (2) up to the number of rotating impeller blades divided by two (B/2). Fig. 3.40 provides a specific example of a predicted and tested threenodal diameter for an impeller. In this case, the natural frequency was 774 Hz by prediction and 796 Hz by test. Dividing these frequencies by the number of nodal lines yield the disk critical speeds. In this case, operation near 15,487– 15,920 rpm could be disastrous. Overall, there have not been many disk critical speed problems in the past 50 years because all manufacturers try to avoid them. Blade resonance for the impeller blade’s leading edge can be a concern. Fig. 3.41 illustrates this concern where the inlet guide vane and blade leading edge are close to one another. Fig. 3.42 shows an example of Campbell diagram for a first leading edge blade mode. For this specific impeller, large SM was present from the 16 upstream return channel vanes as well as the 14 sideload guide vanes.

76 SECTION II Types of Equipment

FIG. 3.40 Three-diameter mode modeled (left) and tested (right). (Courtesy of Elliott Group.)

Centrifugal Compressors Chapter

1400 1200

Leading edge fundamental 16E 14E

1000 Frequency (Hz)

3636 rpm

3420 rpm

FIG. 3.41 Guide vane and impeller arrangement. (Courtesy of Elliott Group.)

800 600 6E 5E 4E 3E 2E 1E

400 200 0 0

1000 2000 3000 Operating speed (rpm)

FIG. 3.42 Impeller Campbell diagram [17].

4000

3

77

78 SECTION

II Types of Equipment

In addition to disk critical speeds, significant minor resonances can also be excited due to the interaction of stationary vanes and rotating blades. Parametric equations have been developed for evaluating these interaction resonances. For certain combinations of the two numbers, the coupled blade/disk modes with diametric nodal lines are either phase canceled or excitable when an excitation source coincides with a resonant mode. These equations are represented as follows [13]: When not at a disk critical speed, |y  S|  |z  B| ¼ n yS¼h

(3.1)

fn ¼ y  S  ω Whereas when at a disk critical speed, For B > 1 h¼n

(3.2)

fn ¼ n  ω such that B is the number of rotating blades, S is the number of stationary elements, fn is the natural frequency (Hz), h is the harmonic of speed, n is the number of diameter nodal lines, y and z are the integers >0, and ω is the rotating speed (Hz). An interference diagram as applied to impellers is essentially a graphical representation of these equations. This type of interference diagram is also called a “SAFE diagram” [15]. Such methods are used to determine whether a certain number of stator vanes could excite the impeller to resonant vibrations. As an example, Fig. 3.43 shows a Campbell diagram of an example impeller showing all possible modes of vibration. The Campbell diagram indicates that it is essentially impossible to avoid every mode in this case. However, the equations of Kushner [13] would indicate that only certain modes are of concern as the rest would experience phase cancelation. Fig. 3.44 shows test results confirming phase cancelation [17]. Phase cancelation was noted to reduce the response to excitation by as much as having good SM. On rare occasions, impellers that meet all design criteria may still fail. Konig et al. [18] and Petry et al. [18a] discuss the potential for Tyler/Sofrin modes coinciding with an acoustic eigen-frequency coinciding with an impeller mode shape. These situations are sometimes referred to as “triple coincidence.”

Deflection Temperatures and pressures can change considerably through compressor startup and operation. Thermal deflection of the piping, the casing, and the support must be managed to keep forces within allowable limits hence preserving equipment alignment.

Centrifugal Compressors Chapter

3

79

FIG. 3.43 Impeller Campbell diagram [17].

Maximum response at each measurement location

7

Point on cover

6

Point on hub

5 4 3 2 1 0

13´ 14´ 15´ 16´ 17´ 18´ 19´ Excitation order of the traveling wave force

FIG. 3.44 Maximum response at each measurement location [17].

Hot/Cold Applications, Centerline Mounting Compressor casings are most often supported near the horizontal centerline since sizeable shifts in shaft centerline exist for cases where the compressor is supported at the base. One such example of this shift for a compressor mounted at the feet is shown in Table 3.2. This shift creates a change in both the coupling hub location on the driven end, as well as the angle of the hub. For the case shown, target alignments for the coupling were 2.64 mm; however, the hot alignment was still targeted to be zero for the coupling to ensure smooth operation.

80 SECTION

II Types of Equipment

TABLE 3.2 Horizontal (X) and Vertical (Y) Shift in Foot Mounted Compressor Centerline at a Discharge Temperature of 436 K Compressor Growth

Compressor 436 K

XInlet

mm

0

YInlet

mm

0.12

XDischarge

mm

0

YDischarge

mm

1.05

FIG. 3.45 Compressor casing with wobble foot arrangement.

Centerline mounting, on the other hand, enables casing expansion or contraction with minimal disturbance of the shaft centerline. This in turn preserves the machine alignment with the driver as the compressor transitions from hot to cold alignment. The attachment of the compressor casing to its support feet may allow sliding to accommodate the thermal deflections or they may be of a flexible design commonly known as a “wobble foot.” Fig. 3.45 shows a wobble foot arrangement. The feet closest to the drive end normally accommodate radial growth whereas the feet closest to the nondrive end normally accommodate axial growth.

Piping/Flange loads The main intake and discharge piping connections to and from the compressor must be arranged in such a manner as to minimize the stress transmitted to the compressor casing via the piping due to either weight or expansion. This means

Centrifugal Compressors Chapter

3

81

the pipe flanges adjacent to the unit have to be aligned and parallel, and that the piping must be supported with provisions for thermal expansion without exerting a total strain on the compressor casing of more than just a very small percentage of the weight of the machine. Precautions should be taken if expansion joints are used, since these do not necessarily reduce piping strains to a reasonable magnitude unless the pressure forces exerted by the expansion joint are compensated for and the spring forces are considered. If the piping is not analyzed carefully and stresses relieved to reasonable values, trouble can be expected with casing distortion, bearing, shaft misalignment, vibration, and generally rough operation.

Thrust Loads An axial thrust load is generated on the compressor rotor by gas forces within the machine. The overall thrust force can be determined from the sum of forces acting at each impeller, at each shaft end, and at each seal. The overall axial thrust load will vary over the compressor operating envelope. As shown in Fig. 3.46, the thrust acting on an impeller is broken down into the static forces due to stage inlet pressure, FS1 (eye), and stage discharge pressure, FS2 (hub), the forces acting on the shroud FD and on the hub disk FN and the momentum forces FM (in the impeller eye). In case of stepped labyrinth seals, the thrust resulting from the pressure distribution along the seal is also considered [19]. Summing these forces, the overall thrust force on the impeller is toward the impeller eye. The magnitude of the thrust force is dependent on the inlet and discharge gas pressures, the inlet fluid momentum, geometry of the impeller and cavities, gas density and rotational speed, of the shroud and hub side cavities, and the seals.

Fo FN

Fs1

FM

Fs2

FIG. 3.46 Axial forces acting on an impeller. (Figure 7 of Y. Bidaut, D. Dessibourg, The challenge for the accurate determination of the axial rotor thrust in centrifugal compressors, in: Proceedings of the Asia Turbomachinery and Pump Symposium, Singapore, February 22–25, 2016.)

82 SECTION

II Types of Equipment

On straight thru compressors, a balance piston is placed adjacent to the last stage impeller to partially balance the thrust from the impellers and seals within the compressor. The force acting on the balance piston is opposite of the impellers and is due to the differential pressure across the piston between the highpressure cavity at the last stage impeller and the lower-pressure cavity outboard of the balance piston, which is connected to the compressor suction. The balance piston is sized to reduce the overall thrust load. As compared to the impeller, the thrust at the balance piston and other seals is a more straightforward calculation. The sum of the loads is the thrust. Due to multiple conditions of operation, the total thrust will vary. On determining the expected range of thrust loading, an appropriate thrust bearing can be selected. Thrust loads of centrifugal impellers are a result of a pressure imbalance between the front face and the rear face of the impeller. The sum of these forces over all impellers and the forces created by the balance piston are the resulting load on the compressor thrust bearing [19a]. From the axial momentum equation, which takes into account the change of the axial momentum of the gas, and the forces due to the static gas pressure in the axial direction: ðð þ !
! ! ! ! ρ C C  d A ¼ p  d A + F (3.3) we get the resulting forces on the impeller as (Fig. 3.47)   ! ! ! F impeller ¼ F momentum ðcexit , cinlet Þ  F pressure pcavity, front , pcavity,rear , pinlet , pexit (3.4) The front and rear cavities are formed between the impeller tip and the labyrinth seals at the impeller inlet, and the impeller hub seals, respectively. The force on the thrust bearing is thus (Fig. 3.48)   ! ! ! F thrust bearing ¼ F impeller  F balance piston pdischarge , psuction (3.5) In the simplest approach to calculate the forces on the impeller, one would assume the pressure in the front and rear cavities to be equal to the pressure at the impeller tip. In a shrouded impeller, however, the gas in the cavity is subject to swirl, and as a result, the static pressure at lower radii is lower than at the tip. The amount of swirl is a function of the cavity geometry, and the leakage flows through the labyrinths. The cavity static pressure distribution can be calculated by
 1 2  r2 (3.6) pðr Þ ¼ ptip  ρðqωÞ2 rtip 2 accounting for the cavity characteristics by introducing a cavity swirl coefficient q.

Centrifugal Compressors Chapter

3

83

Pexit

Cexit Pcavityfront Pcavityrear

Pinlet

Cinlet

FIG. 3.47 Forces on the impeller.

F Impeller

Pdischarge

Psuction

FIG. 3.48 Balance piston.

A simple approach would assume constant swirl coefficients for front and rear cavities. This approach is frequently used in the industry, but high-pressure compressors require more accurate estimates. Correlations and CFD analysis (Fig. 3.49) are typically used for these, along with subscale test measurements for validation.

84 SECTION

II Types of Equipment

FIG. 3.49 Swirl ratio in the shroud and the backside cavity.

Of particular importance for the topic of off-design operation is the fact that the swirl coefficient changes when the impeller is operated away from its design point (Fig. 3.50). Also, the magnitude of the swirl coefficient on the impeller backside changes in the opposite direction from the swirl coefficient on the front side of the impeller. This means that the thrust imbalance (for a given pressure level and a given speed) changes not just due to the pressure difference between the impeller eye and the corresponding backside, but also due to different swirl factors in the cavities in the front and back of the impeller. This imbalance, in particular, changes when the compressor moves from the design point to choke. In general, the shroud side swirl is higher than the backside swirl, a result also reported by Koenig et al. [18]. Because the thrust load has a direct impact of the thrust bearing temperature, which can be measured conveniently, Figs. 3.51–3.53 establish the correlation between nondimensional operating point (Fig. 3.51), thrust load at different speeds (Fig. 3.52), the resulting bearing temperature of the loaded and unloaded pads of the thrust bearing (Fig. 3.53), as well as the axial position of the rotor as a result (Fig. 3.54). The inboard bearing shows a significant increase in temperature (albeit not to a level that would cause concern) when the compressor enters the choke region. The outboard bearing shows only a much lower increase in temperature when the operating point moves toward surge. For this particular application, with the particular selection of the balance piston size,

Centrifugal Compressors Chapter

3

85

0.8 Front cavity—Correlation Front cavity—CFD analysis Front cavity—Pressure measurement Rear cavity—Correlation

0.7

Cavity swirl coefficient (q)

Rear cavity—CFD analysis

0.6

0.5

0.4

0.3

p(r) = ptip – —12 r(qw)2 (r 2tip – r 2)

0.2 0.02

0.03

0.04

0.05 0.06 0.07 Inlet flow coefficient (F)

0.08

0.09

FIG. 3.50 Cavity swirl coefficient for a medium flow stage for different operating points.

the thrust load reverses direction, which explains the behavior of the inboard and outboard bearing temperature. Of course, the bearing temperature increases also with speed. As a result of the thrust load changes and the changing load capacity of the thrust bearing with speed, the axial gaps for all speeds are fairly close together, but change significantly when the compressor is operated from design point to surge or into choke. If we compare the magnitude of the forces acting on the impeller (Fig. 3.55), the pressure from the inlet eye and the pressures in the cavities are usually dominant, but act in the opposite direction. In general, they generate a resulting force, much smaller than the pressure forces, in the direction of the compressor inlet, but as can be seen in Fig. 3.55, this is not always the case. The momentum force, generated by deflecting the gas form more or less axial to more or less

FIG. 3.51 Nondimensional map for multistage compressor. 1500

Axial thrust, lbf (+: Outboard direction)

1250 Speed increase

1000 750 500 250 0 0.030 –250

0.035

0.040

0.045

0.050

0.055

–500 –750 –1000 –1250 –1500

Inlet flow coefficient

FIG. 3.52 Change of axial thrust with operating point.

0.060

0.065

0.070

0.075

Centrifugal Compressors Chapter

3

87

Thrust bearing temperature vs. inlet flow coefficient

150 148

Temperature (F)

146 144 142 140 138 136 134 132 130 0.030

Low-speed inboard

Outboard

Design-Speed inboard

Outboard

High-Speed inboard

Outboard

0.035

0.040

0.045

0.050

0.055

0.060

0.065

0.070

0.075

Inlet flow coefficient FIG. 3.53 Thrust bearing temperature as a function of operating point.

Axial gap vs. inlet flow coefficient

7.0

Low speed

Averaged axial gap (mi)

6.0

Design speed High speed

5.0 4.0 3.0 2.0 1.0 0.0 0.030

0.035

0.040

0.045 0.050 0.055 0.060 Inlet flow coefficient

0.065

0.070

0.075

FIG. 3.54 Axial position of the rotor as a function of the operating point.

radial direction, is usually much smaller than the pressure forces. At very high discharge pressures near choke, when the pressure differential over the impeller is rather small, the momentum force can become dominant, and create a net force toward the discharge end of the compressor.

88 SECTION

II Types of Equipment

30

2 Fpressure, inlet 1.5

20

1 10

0.5

F/Fthrust

Fmomentum 0

0 –0.5

–10

–1

Fimpeller –20

–30

–1.5 Fpressure, cavities 1

2

4

3

5

6

–2

Stage FIG. 3.55 Contributing forces to the impeller thrust in a six-stage compressor.

The descriptions in this section are based on shrouded impellers. As opposed to open-faced impellers, that have free-standing blades, the blades in shrouded impellers are covered. Therefore, the pressure distribution on the front face of the impeller is governed by the impeller discharge pressure and the impact of swirl flow. In an open-faced impeller, the pressure distribution would be governed by the pressure buildup in the impeller flow passages.

Foundation The foundation and support system for a centrifugal compressor must support the equipment and keep them in sufficient alignment. The support base is either reinforced concrete with steel sole plates or a fabricated steel baseplate supported by another structure. Concrete foundations with steel sole plates are quite economical. The concrete foundation typically consists of concrete beams on the order of 46–80 cm thick with reinforcing steel throughout. The concrete beams may be elevated as part of an overall concrete structure thereby elevating the equipment as needed. Fabricated steel baseplates are made from structural steel and typically designed for perimeter support (Fig. 3.56). Main longitudinal beams run the length of the string parallel to the shaft centerline. A top deck is normally included. These baseplates are typically designed to be lifted or set down while fully loaded. The baseplates may require grouting to a concrete foundation or

Centrifugal Compressors Chapter

3

89

FIG. 3.56 Baseplate with compressors, and baseplate with gear and motor. (Courtesy of Elliott Group.)

FIG. 3.57 AVM fixed and sliding directions to isolate base package deflection from FPSO deck twist and bending [20].

may allow bolting directly to a steel substructure. In many cases, the baseplate can be made continuous supporting all bodies in the string. Baseplates that are used for offshore floating platform applications are unique. They must be of the continuous design and made especially rigid to allow for three point mounting. Either antivibration mounts (AVMs) or Gimbal mounts are used at each of the three locations. AVMs are the most common due to their high damping and successful application. The top pad of each AVM is bolted to the base and the bottom pad is bolted to the ship deck. The AVMs isolate the base package from the vessel hull and deck in two ways (Fig. 3.57). First, the AVMs are heavily damped, decreasing the amplitude of base package displacement. Second, the AVMs accommodate sliding and twisting, which prevent deck and twisting and motion from being transmitted into the base package [20].

90 SECTION

II Types of Equipment

The API 617 and 686 standards are quite detailed and specific in regards to support structures and alignment provisions and should be followed. Base plates, or individual sole plates for compressor and driver, should be aligned before grouting on a concrete subbase or anchoring to structural steel supporting members in order to provide a proper basis for alignment and leveling of compressor and driver shafts. Mounting surfaces must be flat, have good surface finish, and be parallel to each other. Jack screws are provided to allow shifting the compressor horizontally and axially. Vertical jack screws are provided to raise the compressor for shimming.

Auxiliary Systems/Supporting System Seal Systems The primary purpose of a seal system is safety. Emissions reduction, while an objective, is of secondary importance.

Dry Gas Most centrifugal compressors are equipped with DGSs as shaft end seals, as they are currently the best way to meet the requirements for safety, reliability, and economy. In principle, four different types of DGS arrangement are used, see Fig. 3.58. All arrangements have in common that treated process gas is injected at between the process seal and the primary seal. The process seal is shown as a labyrinth. The majority of the seal gas flows across the process seal and a minor amount flows across the primary seal faces. The amount of seal gas is designed to maintain a minimum velocity in the gap of the labyrinth. This prevents migration of untreated process gas into the area of the DGS. Typical

FIG. 3.58 DGS system.

Centrifugal Compressors Chapter

3

91

values for the gap velocity are 5–10 m/s. Furthermore, all arrangements have a separation seal in common which separates the DGS from the bearing cavity. A single seal arrangement consists of a single set of seal faces separating process gas from the atmosphere. Single arrangements are mainly considered for nonflammable and nontoxic gases. The seal gas leakage is typically vented to the atmosphere. A double seal arrangement consists of two single DGSs arranged in a mirrored configuration. Double seal arrangements are usually used only for process gas pressures smaller than 5 barg. As the common supplied seal gas is typically nitrogen, there is no process leakage to atmosphere but nitrogen will migrate into the process. The vent is typically open to atmosphere. A tandem seal consists of two single DGSs arranged in series. Furthermore, the tandem arrangement differs in one without and one with intermediate labyrinth. In the latter, a labyrinth sits between the two single seals. As the compressed gases in the oil and gas industry are usually flammable or toxic, the tandem seal arrangement is the most considered. In the case that leakage of process gas to the atmosphere is unacceptable the tandem seal arrangement with intermediate labyrinth is the right choice. Fig. 3.59 shows this arrangement in detail. The seal gas pressure is throttled across the primary seal to primary vent pressure. Secondary seal gas, typically nitrogen, is supplied between the intermediate labyrinth and the secondary seal to prevent primary seal leakage from reaching the secondary seal. For this purpose, the amount of secondary seal gas need to be sufficient to maintain a gap velocity at the intermediate labyrinth of about 3–5 m/s. Leakage from the primary seal and the majority of the secondary seal gas flow is routed to the primary vent. Secondary seal leakage is throttled across the secondary seal and routed to the secondary vent, which is typically open to atmosphere.

FIG. 3.59 Tandem arrangement with intermediate labyrinth.

92 SECTION

II Types of Equipment

In case of primary seal failure, the secondary seal is designed to operate at primary seal conditions, which prevents uncontrolled leakage to atmosphere in the event of a failure and should allow for a safe shutdown of the compressor. In case of a total seal failure where both the primary and the secondary seal are broken, the intermediate labyrinth provides a restriction between the primary and secondary seal helping to limit process gas leakage. As a result, the pressure increase in the secondary vent is reduced and the separation seal is less heavily loaded. In addition to the function of preventing migration of process gas into the bearing cavity in the event of a total failure, the task of the separation seal is to keep away the lubricating oil from the secondary seal during normal operation. In Fig. 3.59, a set of floating carbon rings is used as a separation seal. Typically, nitrogen or air is supplied in the middle of the two carbon rings, generating a gap velocity of 10–20 m/s to prevent oil mist and splash oil from entry. To ensure proper function of a DGS, an adequate support system is essential. Accordingly, the support system is designed to provide: l

l l

clean and sufficiently dry seal gas/separation seal gas at the required pressure, flow rate, and temperature, safe venting of leakages, and condition monitoring of the DGS and the separation seal.

The most complex support system is given for the case of a tandem seal with intermediate labyrinth. This arrangement has five connections, all connected to the support system. An example P&ID is shown in Fig. 3.60. Typically, process gas is used as the primary seal gas. In order to generate a positive flow, the pressure of the supply source needs to be higher than the process gas pressure at the process seal in any operating condition. Most often, the seal gas is taken from the discharge or a designated location in the compressor. As during transient or standstill conditions, the supply pressure can equal or even drop below the pressure at the process seal, a seal gas booster might be required. Less often an external source is used. The disadvantage here is that the settle out pressure increases continuously when the compressor is isolated and in standstill/idling condition. In order to achieve the required flow velocity at the process seal either a flow control or a differential pressure control device is used. Flow control is achieved with a needle valve as the most simplistic option, or the use of a self-adjusting flow control valve. The maintained flow is constant and not affected by the clearance of the process seal. This can lead to higher flow velocities as needed and represent a recycle loss. A differential control valve regulates the differential pressure acting on the process seal. The flow rate varies with the clearance of the process seal but the flow velocity is constant. To ensure the seal gas is clean a coalescing filter shall be included in the supply line. The filter should be a duplex design to ensure an online exchange

Centrifugal Compressors Chapter

3

93

FIG. 3.60 Example of a support system for tandem dry gas seals with intermediate labyrinth in a single-shaft compressor.

of the filter cartridges. The purpose of a coalescing filter is to remove fine particles and liquid-phase contaminants. Debris and particles can cause excessive wear of the faces and hang-up the axially displaceable seal components. Typically, the filter shall be able to remove particles that are smaller in size than the range of the operating gap of the seal faces. Typical filter element particle size rating is 1–5 μm and the achieved removal efficiency rate is typically 99.9%. Only for separation seals with a fixed clearance, that is, labyrinths or noncontacting bushing seals, the filtration requirement is less. Typically, 10 μm at a removal efficiency of 99.9% is required. To ensure the supplied seal gas is sufficiently dry, special attention must be given to avoid condensation. Due to the Joule-Thomson effect, depressurization of nonideal gases causes a change in temperature. Typical hydrocarbon gas mixtures cool down. In operation, this cooling takes place on all components in the supply lines. As a result, condensates can form and get into the seal where they cause considerable malfunction, such as hang-up by sticking the axially displaceable seal components. Especially during pressurized standstill condition, condensation may occur in the seal gap between the faces. Usually, the largest pressure drop takes place there and the absence of shear and windage effect cannot compensate the cooling. Formation of condensates sticks the faces, resulting in high torques during start-up, followed by chipping and breakage of the faces.

94 SECTION

II Types of Equipment

1000 900 800

Pressure (psia)

700 600 500

Mixed (Gas/liquid) phase

Gas phase

400 300 200 100 0 50

70

90

110

130

150

170

190

Temperature (°F) Gas dew line

Seal gas supply

FIG. 3.61 Phase map of the seal gas and 20 K margin to the expansion curve.

In order to avoid condensation, the calculation of the expansion curve from the point the seal gas enters the support system all the way through the seal gap and comparison with the condensation curve in the phase map is mandatory. A minimum design margin of 20 K is usually recommended to ensure the seal gas is gaseous, see Fig. 3.61. Additional cooling in the lines and supply channels must be considered. If a proper design margin cannot be maintained, it may be required to heat up the seal gas to maintain the specified design margin from dew point. Electrical heaters with Thyristor control utilizing silicon-controlled rectifiers (SCRs) can precisely control the heating load and hold the seal gas temperature. For mixtures with already small amounts of longer-chain hydrocarbons (C6 +), the condensation limit shifts strongly in the direction of higher temperature. This can result in gas temperatures that exceed the design temperature of individual components downstream of the heater. For these cases, a cooler with a downstream separator is installed. Long-chain hydrocarbons condensate and get removed in the separator. A heater after the separator is then absolutely necessary. In cryogenic applications, increasing the seal gas temperature to provide higher margins from dew point may adversely impact the compressor performance, because of the main portion of warm seal gas (e.g., more than 90%) is circulated back to the compressor flow path through the process labyrinth at the back of the impellers. Fig. 3.62 shows a typical arrangement of the entirety of components that might be required for gas conditioning. After passing the conditioning unit, the seal gas gets throttled by means of the control valve and forwarded to

2

3

4

5

6

7 TE

1

TE LIT

Filter

Filter

PDIT

LIC

TSHH

Heater

TE

RoTech Booster

TIC

PIT LCV

Cooler (optional)

Separator (optional)

Filter (required)

Heater (optional)

1 Seal gas

Shift operation point next to liquid phase in order to be able to deposit more condensates in the separator.

To deposit liquid out of gas, when gas is very wet and filter can’t deposit enough.

To clean gas (condensates and particles).

Increase operating gas To provide gas in temperature. adequate amount and pressure for start up, shut down, and standstill.

supply

2 External seal gas supply

Booster (optional)

2 Cooler

3 Separator 4 Filter unit

5 Heating 6 RoTechBooster

FIG. 3.62 Typical arrangement of the entirety of seal gas conditioning components.

7 Gas OUT - to primary seal gas supply

95

1 Gas IN

3

Diagram of a gas conditioning system

Centrifugal Compressors Chapter

Seal gas supply

96 SECTION

II Types of Equipment

the seal. To monitor proper supply, the gas temperature, flow, or the differential pressure is measured and compared with setpoint values. An alarm is issued by the compressor control system if mismatch occurs. Frequently, orifices are installed in each supply line downstream of a common control valve to limit the amount of seal gas in the event of a seal failure and to avoid harm to other DGSs connected to the same supply. A support system for DGSs with intermediate labyrinth implements two vents, primary and secondary vents. Both vents have the function to safely remove the gas mixture from the leakage of the upstream DGS and the leakage of the intermediate labyrinth or the partial leakage of the bearing seal. Special attention is required for the primary vent as the seal health monitoring functionality for both the primary and the secondary seal is usually hosted there. To protect the necessary instruments, a rupture disk or a pressure relieve valve is installed in parallel with the vent line. This limits the pressure increase in the vent in the event of primary seal damage. Because primary seal leakage is typically flammable or toxic, the primary vent line often discharges into a flare system charged slightly above atmospheric pressure. This requires the installation of a back-pressure valve at the discharge of the vent line, to control the pressure upstream about 0.5–1 bar above the flare pressure. A failure of the primary or secondary seal usually leads to a sharp increase in leakage; this principle is used for monitoring. Failure of the primary seal leads to an increase in flow, while failure of the secondary seal leads to a drop of gas flow in the primary vent. Flow is measured and compared with setpoint values. An alarm (low or high) followed by a shutdown (low low or high high) is triggered by the compressor control system. Alternatively, pressure measurement for monitoring is common. For this purpose, installation of a flow orifice between the pressure measurement and the back-pressure valve is necessary to allow flow-dependent pressure changes. In low-pressure applications, monitoring with pressure measurement reaches its limits. To avoid reverse pressure on the primary seal, the primary vent pressure must be correspondingly low. Pressure fluctuations due to a DGS failure can practically no longer be measured reliably.

Wet Wet seals are generally not selected for new equipment due to their higher leakage rates and cost of ownership as compared to DGSs. However, a very large population of wet seals has been installed in the past, and these applications continue to operate reliably. As the name implies, wet seals require a liquid; and for oil and gas applications, that liquid is typically oil such that these seals are often called oil seals. Two types of wet seals are discussed: l l

mechanical (contact) seal and liquid ring seal

Centrifugal Compressors Chapter

3

97

FIG. 3.63 Mechanical (contact) face seal [21]. (Courtesy of Kaydon Ring and Seal.)

A typical example of a mechanical (contact) face seal is shown in Fig. 3.63. These seals are typically used to seal against suction pressures up to about 3.8 MPa. The requirement is to seal against the maximum pressure which will be encountered at the compressor suction, and this is typically the settle out pressure, which is encountered during start-up and shutdown. These seals are comprised of several components including a stationary seat, a rotating seat, and, in this example, a carbon ring sandwiched in between, each with a high degree of contact to form sealing faces between them. The seal surfaces are held together by a combination of hydraulic and mechanical forces. The hydraulic forces are produced by the seal liquid, whereas the mechanical forces are normally produced by some type of spring. Oil acts to lubricate and cool the seal, and also creates a positive fluid seal at the sealing faces. The oil is supplied at 241–345 kPa above the compressor suction pressure; thus a slight flow of seal oil takes place across the faces toward the process gas, which prevents the outward flow of gas to atmosphere. Oil leakage on the inboard side of the seal, which is exposed to process gas, is considered dirty (sour) and is drained to a separate area from that of the clean (sweet) seal oil and lubrication oil. The sour oil is kept from entering the process by an inboard labyrinth seal. A buffer gas supply may be applied to further

98 SECTION

II Types of Equipment

prevent the seal oil from entering the process stream and to reduce the amount of oil contamination. Little to no gas leakage occurs across the actual seal. Most emissions are actually from seal oil degassing due to the oil absorbing gas on the inboard side. The sweet uncontaminated oil is drained back directly to the reservoir. The contaminated oil is sent to a separate drain cavity and is either discarded or degassed. If the contaminated oil is reconditioned, it must first be sent to a degassing tank to release the entrained gas by heating the oil. This eliminates the possibility of contaminating the fresh lube and seal oil reservoirs. Then, the degassed oil is sent directly to the reservoir to be reused. The released gas may be vented, flared, returned to the compressor inlet, or used as fuel in some cases. The main purpose of the seal oil supply system is to supply clean, cool oil to the seals at the required pressures and flows. The system generally consists of an oil reservoir, oil pumps, coolers, filters, and regulators to precisely control the pressures and temperatures required for the compressor seals. Fig. 3.64 is a typical oil supply system for a mechanical face seal. For cost and space savings the seal system may be incorporated with the lubrication system to allow for additional oil capacity and flows.

FIG. 3.64 Seal oil supply system for mechanical face seal [21].

Centrifugal Compressors Chapter

A

B

C

3

99

D

E

F

G

H

FIG. 3.65 Oil sleeve seal [21].

A typical example of a liquid film seal is shown in Fig. 3.65. These seals are typically used to seal against suction pressures that are too high for mechanical seals. The basic principles of sealing are similar to the contact seal with the primary difference being the restriction has changed from an axial face to close clearance sleeves. Oil is supplied between these floating bushings that run at close clearance to the rotor. By maintaining the oil pressure 34–70 kPa above the gas pressure, a positive seal is formed and the process gas is prevented from leaking to atmosphere. The majority of the oil supplied exits the seal by flowing across one or more atmospheric side floating bushings. The film of oil between the bushings and rotating sleeve provide hydrodynamic separation. The seals can influence the machine’s rotordynamic characteristics. The sour oil is kept from entering the process by an inboard labyrinth seal. A buffer gas supply may be applied to further prevent the seal oil from entering the process stream and to reduce the amount of oil contamination. Liquid Film Seal Oil Supply System The seal oil supply system for a liquid film seal is essentially the same as for a wet mechanical seal. The major difference is the amount of pressure difference between the oil and gas, and how the pressure difference is maintained. A

100 SECTION

II Types of Equipment

FIG. 3.66 Seal oil system for liquid film seal [21].

mechanical seal needs to provide oil at 0.24–0.34 MPa above a gas pressure on the order of 2.76 MPa. This can easily be done with a pressure-regulating valve. A liquid film seal may be required to provide oil at 30–70 kPa above a gas pressure of 17.24 MPa. The general practice to maintain this type of pressure difference is to use an overhead tank as shown in Fig. 3.66. Here, a seal oil level is maintained by a level control valve. With gas pressure on top of the seal oil, the static oil head automatically provides the correct pressure differential [21].

Bearing/Lube Oil Centrifugal compressor strings with hydrodynamic bearings require a pressurized lubrication system. The lubrication system provides an uninterrupted supply of cooled and filtered oil to the bearings located at the compressor, driver, and any gear. Lubrication may also be required for compressor wet seals if they are present, as well as any turbine hydraulic control system. A simple system is shown schematically in Fig. 3.67 and pictorially in Fig. 3.68. In this system, oil is taken from a reservoir by a low-pressure pump and sent through a cooler and filter. After the filter, the oil will reach a common header, which has takeoff legs to provide lubrication for the bearings and any gears that may be present. The common header is typically supplied at a

High pressure control oil supply To oil reservoir

To oil reservoir

Cooling water out

Oil return from units

A

A Fill line

Cooling water in

Oil coolers

PCV

Vent

Vent B

B

Clean oil drain

Drain

Cooling water in

PDIT

Main oil filters

PCV

Exhauster vent

TI

Dirty oil drain

PI

PSL

Start secondary pump @ setpoint (PSIG) falling press.

PI

Purge Fill conn w/strainer Oil from clairifier

RV

Purge

RV PI

1/2” per foot minimum Oil to clairifier Tank drain

TI

TE

PI

LS

LI Reservoir

Heater

Primary oil pump and driver

NC

LIT

Secondary oil pump and driver

Legend Transfer valve 6–way Check valve Globe valve

Relief/safety valve

Oil filter

LI LIT

Pump suction strainer Cooling fan

LS

PSH

Level indicator transmit

PSL

Pressure switch, low

Level switch

RV

Relief valve

Pressure control valve

Concentric reducer

TE

Temperature sensor

PI

Pressure indicator

TI

Temperature indicator

3

Double pass, shell and tube heat exchanger

PCV

Pressure switch, high

Level indicator

Orifice

Gate valve Ball valve

Electric motor driver AC or DC power

Centrifugal Compressors Chapter

Slope

PCV

TE

TI

Cooling water out

Coolers and filter vents

Lube oil supply

101

FIG. 3.67 Lube oil console P&ID. (Figure 8.10 from B.C. Pettinato, Steam turbines, in: George E. Totten (Ed.), Handbook of Lubrication, Application and Maintenance, vol. I, second ed., CRC Press, Inc., 2006.)

102 SECTION

II Types of Equipment

FIG. 3.68 Lube oil console with motor, gear, and compressor on a common baseplate. (Courtesy of Elliott Group.)

pressure of 100–125 kPa, but the pressure may range from as low as 55 kPa in some systems to 345 kPa in others. Oil flow is metered from the common header to each bearing by an orifice or other flow-controlling device. Observation of drain oil flow through sight boxes can be taken as an indication of at least partial flow through the bearing and is often used as a quick indication that the oil pump is running and that the oil supply is probably sufficient. Oil supplied to the bearings functions as both a lubricant and as a coolant to counteract the heat generated by shearing of the oil during operation. Oil from the bearings will drain back to the reservoir through sloping pipes. Fig. 3.68 shows the following major components: l l l l l l

oil reservoir, pumps and drivers, filters, coolers, control valves, and piping

Additional accessories may include relief valves, transfer valves, accumulators, and instrumentation all are shown in Fig. 3.68. Additional lube oil circuits such as for governor or other control mechanisms may be added to a lube oil console consisting of a high-pressure pump, filters, and coolers running off the main reservoir. These additional lube oil circuits can also be provided from an entirely separate lube oil console. Lube oil systems must be cleaned and flushed during their commissioning prior to the start-up of turbomachinery. Most manufacturers provide special instructions for the oil flush. In the absence of such instructions, industry recommendations should be consulted such as those detailed in ASTM D 6439 [22] or API RP 686 [22a]. Prior to start of the flush, temporary suction strainers are

Centrifugal Compressors Chapter

3

103

FIG. 3.69 Temporary bypass piping with mesh screen used to flush contaminants from the lube oil system [23].

required on the lube oil pumps for protection. The bearing and seal areas are bypassed with the use of jumpers until the system is proven to be clean (Fig. 3.69). Temporary bypass piping to the reservoir is installed along with a mesh screen at the end of the bypass. The key to a successful oil flush is a high velocity to ensure turbulent flow. This may require the use of an external pump and such preparations should be made in advance. However, simultaneous use of the main pump and auxiliary pumps may be sufficient to reach the required flow rate. Filter differential pressure can be quite high during flushing. Effective flushing will typically involve variation of oil temperature, vibrating the piping, and many hours of flushing. The flushing should continue until the required cleanliness is achieved based on inspection of the flushing filters or strainers, patch test, particle counters, or ISO 4406 cleanliness level. Once the system is clean, the flush continues through the bearing housings until they too are clean. Lube systems must be periodically inspected and maintained to ensure their proper operation. As a minimum, the following regular checks should be performed: l l l

check filter pressure drop and replace elements as recommended, check the oil reservoir level and add oil as required, and check operation of auxiliary oil pump by operating pump periodically.

Special Considerations Interaction With Reciprocating Compressors and Aerodynamic Stability Strong pressure pulsations into the suction or discharge of a centrifugal compressor can move its operating point into operational instability regions such

104 SECTION

II Types of Equipment

as surge, rotating stall, or choke. This is of special operational and safety concern in mixed pipeline compressor stations where many centrifugal compressors operate in series or parallel with reciprocating compressors. Over the last 30 years, several authors have discussed the impact of piping flow pulsations on centrifugal compressor stability and specifically, on the impact on surge margin and performance. For example, Sparks [23a], Kurz et al. [23b], and Brun et al. [23c] provided analysis and numerical predictions on the impact of discrete and periodic pressure pulsation on the behavior of a centrifugal compressor. This interaction came to be known as the “Compressor Dynamic Response (CDR) theory.” CDR theory explains how pulsations are amplified or attenuated by a compression system’s acoustic response characteristic superimposed on the compressor headflow map. Although the CDR Theory describes the impact of the nearby piping system on the compressor surge and pulsation amplification, it provides only limited usefulness as a quantitative analysis tool, primarily due to the lack of numerical prediction tools and test data for comparison. In their work, Brun et al. utilized an efficient 1-D transient Navier-Stokes flow solver to predict CDR in real life compression systems. Numerical results showed that acoustic resonances in the piping system can have a profound impact on a centrifugal compressor’s surge margin. However, although interesting, the fundamental problem with both Spark’s and Brun’s approach was that no experimental data was available to validate the analytical and numerical predictions. In 2014, laboratory testing of reciprocating and centrifugal compressor mixed operation was performed in an air loop at Southwest Research Institute’s (SwRI) compressor laboratory. Results provided clear evidence that suction pulsations can significantly reduce the surge margin of a centrifugal compressors and that the geometry of the piping system immediately upstream and downstream of a centrifugal compressor will have an impact on the surge margin reduction. In severe cases, surge margin reductions of over 30% were observed for high centrifugal compressor inlet suction pulsation. Pulsation impact results are presented as both flow versus surge margin and operating map ellipses. Some basic design rules were developed from the test results to relate predicted flow pulsation amplitudes to corresponding reductions in surge margin.

Erosion/Fouling/Plugging of Components and Mitigation Strategies Erosion and fouling both cause degradation of compressor performance and can result in reduced operating life. Erosion describes the wearing of components due to abrasives in the gas path. Typical culprits include dust, scale, catalyst fines, and other material, which cause a type of wear known as solid-particle erosion. In addition, liquid streams or droplets will cause a type wear known as liquid droplet erosion. In essence, abrasives and liquids carried through the compressor by the high gas velocity are quite capable of eroding impeller blades and stationary vanes.

Centrifugal Compressors Chapter

3

105

Solid-particle erosion is managed by ensuring cleanliness during commissioning and operation. Liquid droplet erosion is managed by several techniques. Liquids should be removed at a properly sized knockout drum prior to entering the compressor. Compressor operation must keep the inlet condition above the dew point. Wash systems must be operated and maintained such that the compressor is not over loaded with liquids. Fouling describes a buildup of solids on aerodynamic surfaces and other locations. Deposit buildup is often in the form of salts (chlorides), polymers, or particulates. Of these foulants, salts are perhaps the most common and difficult to deal with in the refinery whereas polymers are most common in the chemical industry. The causes of fouling are not always known (Fig. 3.70). Foulant buildup causes a number of potential problems. Aerodynamic pathways can become congested adversely affecting the compressor performance. Clearances can decrease altering the thrust balance on the compressor. Labyrinth seal cavities can get plugged, increasing leakage. The rotor can become imbalanced if the foulant is unevenly distributed or if a piece of foulant breaks off. Further, barrel compressors can be difficult to open and clean. Numerous problems may also be experienced outside of the compressor. Foulant can build up at flowmeters and other instrument locations causing erroneous readings. Check valves can collect fouling deposits and leak. Antisurge valves and lines can be a concern. Heat exchangers are a location that is particularly prone to fouling.

FIG. 3.70 Fouled compressor rotor due to salt buildup. (Photo courtesy of Elliot Group.)

106 SECTION

II Types of Equipment

Several methods have been developed for dealing with fouling situations. The surest method remains taking the compressor offline and cleaning out the foulant to restore efficiency. This can be done during an overhaul by mechanical means or done by performing an offline wash. However, this requires shutting down the unit and losing production. Less costly methods that focus on extending the time between cleanings have been developed such as online washing and nonstick coatings. An online wash method involves the use of a solvent spray at the inlet and interstage locations to remove or prevent the accumulation of deposits in the impeller and gas passages. The composition, amount, and frequency of the spraying depend on the nature and rate of buildup of deposits. Any spray should be thoroughly atomized to prevent solid streams of liquids from entering the unit. The introduction of liquids into a compressor has the potential to create problems of drainage and erosion. Drain provisions must be made to prevent the buildup of liquid level inside of the casing as this may result in the collected liquid passing through the unit in slugs. A compressor with these provisions is shown in Fig. 3.71. Erosion is a function of the liquid particle size, the quantity of liquid injected and time. Erosion can be managed by limiting the spray quantity and frequency of spraying to a minimum consistent with the cleaning requirements. Finding the right solvent for online spraying can be tricky. Steam has been successfully used as a solvent for some situations such as coke oven gas. Naphtha has also been used successfully in cracked gas service and other refinery applications as a solvent. Coating the rotor and diaphragms with very slick coatings creates a surface that foulants have difficulty adhering to (see Fig. 3.72). Two common types of coatings are organic coatings such as polytetrafluoroethylene (PTFE) and electroless nickel. The organic coatings tend to have superior corrosion resistance

FIG. 3.71 Compressor with nozzles for online wash. (Photo courtesy of Elliot Group.)

Centrifugal Compressors Chapter

3

107

FIG. 3.72 Organic coating (left) and electroless nickel coating (right). (Photo courtesy of Elliot Group.)

and fouling resistance, but they have low durability. Electroless nickel has good corrosion and fouling resistance, and good durability, which offers additional advantages where aggressive wash oil and water injections are utilized.

Driver-Specific Concerns With Centrifugal Compressors Gas Turbine Driven Gas turbine driven compressor systems are often employed because they can often employ process gas as fuel, neglecting the need for high wattage electrical requirements from the grid. The downside to this approach is that two turbomachines are coupled and must operate at the speed, thus changes in speed affect the performance in both the driver and the driven machines. When a compressor system does not require substantial speed variation, a gas turbine drive often makes economic sense. VFD Motor Driven VFDs are utilized when an electric motor must operate over a speed range, as opposed to a fixed speed. These devices can generate significant dynamic torques, which have the potential to excite torsional critical speeds. The frequency content of the dynamic torques can vary substantially, and can occur at either integer or noninteger multiples of running speed. Please see the section on Torsional Rotordynamics or Chapter 7 (drivers) for further information related to VFDs. Multitrain Configuration Parallel/Serial Centrifugal compressors are often laid out in parallel or serial configurations. Parallel systems provide extra capacity as well as inherent redundancy (or sparing), while serial applications provide greater range of pressure ratios.

108 SECTION

II Types of Equipment

The economics of which system makes the most sense often depends on the application. When multiple compressors are installed at the same site, the complexity of operation increases greatly as well, leading to transient startup and shutdown challenges that can result in surge, stall, or excess vibration in adjacent machines.

HP Inline Centrifugal Compressor Applications Rotordynamic Implications Seal Considerations The balance piston seal type and geometry can have a significant impact on the rotordynamic stability of a centrifugal compressor. Labyrinth seals are generally preferred from a compressor performance standpoint, due to the tighter clearances (possible because of the use of abradable materials) and lower leakage rates involved. However, the destabilizing cross-coupling generated by these seals (and a general lack of damping) can be detrimental. As such, when low stability margins are evident, hole-pattern seals are often incorporated at the balance piston. These can provide significant damping, thus resulting in a strong benefit to rotordynamic stability. However, care must be taken to ensure that the seal clearance is carefully managed, since the behavior of these seals has been shown to be very sensitive to both mean clearance and the taper of the clearance (difference between inlet and exit clearance). In some cases, these seals have been known to be destabilizing if the clearance is not properly managed. This is especially true in high-pressure back-to-back machines, where significant division wall deflection is possible. The effects of differential pressure, tolerance stack-up, and temperature gradients should also be well understood in such cases. An initial cold clearance offset may be necessary to ensure that the seal experiences an optimal clearance environment during the planned operating conditions.

Typical Compressor Types and Their Market Space Once an oil or gas reservoir is discovered and assessed, the task is to maximize the amount of oil or gas that can ultimately be recovered. Oil and gas are contained in the pore spaces of the reservoir rock. Some types of reservoirs allow the oil and gas to move freely, making it easier to recover. Other reservoirs restrict the flow of oil and gas and require special techniques to move the oil or gas from the pores to a producing well. Very tight rock may even require fracking to yield economical amounts of oil or gas. To prepare the well for production, the bore hole is stabilized with a casing (lengths of pipe cemented in place), and a small-diameter tubing string is centered in the wellbore and held in place with packers. This tubing will carry the hydrocarbons from the reservoir to the surface. Many oil and gas wells are on the ocean floor, and production requires an offshore platform.

Centrifugal Compressors Chapter

3

109

Most oil wells produce oil, gas, and water. This mixture is separated at the surface. Initially, the oil well may produce mostly oil with a small amount of water. Overtime, the percentage of water increases. This produced water varies in quality from very briny to relatively fresh. Where this water cannot be used for other purposes, it may be reinjected into the reservoir—either as part of a waterflooding project or for disposal (returning it to the subsurface). Gas or condensate reservoirs produce lighter hydrocarbons than oil wells. In all cases, the hydrocarbons produced are hydrocarbon mixtures, with a high proportion of methane. Often other gases, like nitrogen or CO2 are present, and the gas is frequently saturated with water. These natural gas wells do not produce oil but usually some amount of heavier, liquid hydrocarbons, which are called condensate. In addition to condensate, natural gas liquids (ethane, propane, and butane) are removed at a gas processing plant, along with other impurities, such as hydrogen sulfide and carbon dioxide as well. Natural gas liquids often have significant value as petrochemical feedstock. Natural gas wells also often produce water, but the volumes are much lower than is typical for oil wells. Gas from gas wells is compressed and either fed to a gas plant or a pipeline. Usually, more than one well in a geographic area are piped via “flow lines” to a booster compressor or even multiple booster compressors in an inlet compression station. The compressor station is usually close to the well head, and upstream of the gas plant. The gas usually comes from a number of wells, which often produce at different pressure levels. The applications usually have low suction pressures (0.3–2 MPa), and the gas is compressed to about 7–10 MPa. Therefore, compression is accomplished in stages, with cooling of the gas between stages. A typical scenario involves small compressors close to the wellhead feeding to centrally located larger compressor stations. The inlet pressures vary with the dynamics of the reservoirs. A gas gathering system usually starts off with a relatively low ratio (1.25–1.5 range) and high flows. The reservoirs usually decline in ability to produce and require lower pressures to maintain volume flow rates. Eventually, the losses of the gathering system well tubing and flow line piping dominate and no amount of compression will maintain the flow rate. Many reservoirs under compression will draw down to near vacuum at the wellhead prior to abandonment. This leads to booster compression requirements of high rotation and low flow, usually utilizing all the possible horsepower installed in the initial operating case. Ideally, this is following a constant horsepower, rising pressure ratio and falling volume scenario. The usual strategy for designing booster compression facilities is to consider the reservoir pressure versus flow decline curve and to plan an economically optimized life of field compression scheme. This may involve justifying to invest in compressor configurations that allow higher pressure ratio (thus usually a larger number of impellers), and lower flow future aerodynamic components.

110 SECTION

II Types of Equipment

In some gas fields, the incoming gas is cascaded to a lower-pressure level to separate the liquids from the gas. This also leads to gas available at different pressure levels. In other instances, the gas compression is straight through, without any sidestreams. Natural gas is thus either a byproduct of oil production (“associated gas”), or the main product to be produced. In oil production, well fluid is processed in a gas-oil separation system where its pressure is reduced in several stages. In each decompression stage, the associated gas (also called flash gas) is released in a separator until the pressure is ultimately reduced to slightly above atmospheric pressure. Flash gas is compressed to about 8–10 MPa. The compression is in multiple stages, with sidestreams added the crude oil is then sent to a stabilizer column where it is heated and cascaded through a series of bubble trays spaced throughout the column. Hydrogen sulfide (if present) and remaining light hydrocarbons boil off in this process and are collected at the top of the column, while the sweetened heavy crude is drawn off from the bottom. The stabilized oil is then cooled and stored. The streams collected from the top of the stabilizer unit are treated in accordance with environmental regulations. The associated gas produced can be used in three different ways: it can be exported, that is, sent via a pipeline, which runs subsea in case of offshore production to a gas treatment facility—the gas plant. Gas export compressors will compress the associated gas to pressures of 10–24 MPa, and then feed a pipeline to transport the gas to the gas plant. Another use is for efforts to enhance the oil recovery. Even today, with advanced technologies, in some reservoirs, more than two-thirds of the oil present may not be recoverable. For enhanced oil recovery (EOR), the gas can either be compressed, or injected in the well (gas lift), or it can be compressed into the reservoir (gas reinjection). Gas lift compressors bring the gas to pressures of 6–12 MPa, although sometimes 20 MPa may be required. It is then injected into the oil well, where it increases the downhole pressure, thus enhancing the flow from the reservoir into the well. Gas reinjection compressors have to deliver gas at pressures high enough to overcome the pressure in the reservoir, thus requiring discharge pressures ranging from 10 to over 80 MPa discharge pressure. Recent material technology advances allow associated sour gases containing high percentages of H2S and/or CO2 to be reinjected without the need for sweetening. The pressure ratio in flash gas compression, gas lift, gas export, and reinjection applications is usually so high that the gas has to be intercooled during the compression process. For offshore applications, the services described above are installed on platforms (fixed leg for shallow waters, floating for deeper waters). Potentially, these services can be located on the seafloor instead. The development of subsea compressor packages is a current challenge, and there are no subsea compressors in commercial production at this time. Challenges involve the issue of separation of gas solids and liquids close to the well (or, conversely, the capability

Centrifugal Compressors Chapter

3

111

of the compressor to compress gas with significant amounts of entrained liquids), as well the design of compressor motors and drive systems to survive for extended periods of time on the seafloor [24]. The production pattern for most wells follows a pattern where production will increase for a short period, then peak and follow a long, slow decline. The shape of this decline curve, how high the production peaks, and the length of the decline are all driven by reservoir conditions. The decline curve can be influenced by cleaning out the wellbore to help oil or gas move more easily to the surface. This is done by fracturing or treating the reservoir rock with acid around the bottom of the wellbore to create better pathways for the oil and gas to move through the subsurface to the producing well, or by drilling additional wells or by employing EOR techniques. This also means that the operating conditions of the compression equipment will change overtime. The raw natural gas is treated in a gas plant to created marketable products. Gas plants are designed to produce dry export gas (i.e., gas with very little water, a low hydrocarbon dewpoint, limited amounts of CO2, and other contaminants) and liquefied petroleum gas (LPG) products (ethane, propane, and butane). For the range of gas compositions at the inlet, the plants have specified recovery targets for the heavier hydrocarbons. The process steps inside the plant include: primary separation, front-end compression (boost compression, inlet compression), carbon dioxide removal, mercury/chloride removal, gas dehydration, gas expansion (turboexpander), LPG/condensate fractionation, dry (sales) gas compression, storage, and utilities. In a gas plant, several compression duties have to be covered: - Boost compression (inlet compression) to bring the gas from delivery pressure (from the gas gathering system) to plant pressure. - Recompression (sales gas compressor) to bring the natural gas from plant pressure to pipeline pressure. This duty may also be referred to as pipeline head station (essentially depending on whether the compressor is operated by the gas plant or the pipeline operator). - Turboexpander/compressor for the low-temperature cryogenic cycle. After treatment in the gas plant, and recompression in a sales gas compressor, or in a head station, the gas is fed into a transmission pipeline. Natural gas is almost always transported through pipelines, except in cases where a pipeline cannot economically be built. In that case, the gas can be liquefied (LNG) and transported on a ship.

112 SECTION

II Types of Equipment

Natural gas can be transported over large distances in pipelines. Many transmission pipelines require periodic booster compression, increasing pressure to overcome pressure drop caused by friction in the pipeline. Optimal pipeline pressures, depending on the length of the pipe, as well as the cost of steel, are in the range of 4–16 MPa balancing the amount of power required to pump the gas with the investment in pipe. Most interstate or intercontinental pipeline systems operate at pressures between 6 and 10 MPa, although the pressures for older systems might be lower. The gas usually has to be compressed to pipeline pressure in a head station (usually coming from a gas plant). This head station often sees pressure ratios near 3. The pipeline compressors are arranged in regular distances along the pipeline, usually spaced for pressure ratios between 1.2 and 1.8. The distinction is sometimes made between mainline stations (that basically operate continuously) and booster stations that are only in operation sporadically to assist mainline compression. Subsea pipelines often only have a headstation, but no stations along the line. They are either used to transport gas to shore from an offshore platform (see export compression), or to transport gas through large bodies of water. In either case, relatively high pressures (10–25 MPa) are common. A few onshore pipelines worldwide make use of the added super compressibility of the gas at pressures above 14 MPa (depending on gas composition) and operate as “dense phase” pipelines at pressures between 12.5 and 18 MPa. Not only is natural gas transported in pipelines, but also CO2. CO2 is noncorrosive as long as it is dehydrated. Most applications transport CO2 in its dense phase at pressures above 14 MPa, in particular to avoid two-phase flows when ambient temperatures drop. If the throughput of a pipeline has to be increased, two possible concepts can be used: building a parallel pipe (looping), or adding power to the compressor station (i.e., adding one or more compressors to the station), or a combination of both. These means can also be combined. If power is added to the station, the discharge pressure can be increased (assuming this is not already limited by the pipeline maximum operating pressure). The station will therefore operate at a higher pressure ratio. The added compressors can either be installed in parallel, or in series with the existing machines. If the pipeline is looped, the pressure ratio for the station typically is reduced, and the amount of gas that can be pumped with a given amount of power is increased. In either scenario, the existing machines may have to be restaged (for more pressure ratio and less flow per unit in the case of added power, for more flow and less pressure ratio in the other case). In general, pipelines that have many takeoffs, and interconnects (like in the Unites States) tend to operate more frequently in transient, nonsteady-state conditions. Long, transcontinental pipelines that basically transport gas from points A to B, tend to operate closer to steady-state conditions. In most cases, compressors experience a significant range of operating conditions (Fig. 3.73) [24].

Centrifugal Compressors Chapter

3

113

FIG. 3.73 Operating points collected over a 6-month period in a gas compression station with three units in parallel. (Figure 23 from R. Kurz, S. Ohanian, M. Lubomirsky, On compressor station layout, ASME GT2003-38019, 2003.)

As part of the transportation process in pipelines, gas can be stored in storage facilities, which often use former gas fields, or salt caverns. This allows to balance differences in supply and demand on a seasonal or daily basis. The compressors for storage applications have to be capable to compress high gas flows at low-pressure ratios when the facility is empty, and the cavity pressure is close to the gas pipeline supply pressure; in addition, it must compress low gas flows at high-pressure ratios when the facility is filled, and thus the storage pressure is much higher than the pipeline supply pressure. This requirement is often met by the capability to operate compressors either in parallel or in series. Usually, all applications upstream and including a gas plant are considered “upstream” applications, while the applications related to bringing gas to the ultimate users are referred to as “midstream.” Applications in refineries, chemical, and processing plants are considered “downstream” applications [24].

Gas Transmission Compressor Natural gas is transported over large distances via pipelines. This pipelines are typically operated at pressures of about 100 bar, but there also pipelines at higher pressures (up to about 240 bar in the case of subsea pipelines), and many older pipelines at lower pressures (40 bar). In any case the gas will experience a loss in pressure due to friction effects in the pipe. Therefore, we find gas compressor stations every 50 to 100 miles along the pipeline which recompress the gas. In order to minimize energy consumption along the pipeline, these compressor stations operate at relatively low pressure ratios, typically in the range of 1.2–1.7 [24a]. In many instances, the fluctuation in gas demand (on an hourly, daily or seasonal basis) requires the compressors to operate over a wide

114 SECTION

II Types of Equipment

range of conditions with a high efficiency (Fig. 3.73). The machines are typically variable speed machines. In many instances, multiple trains are installed in compressor station.

High Pressure Injection Gas injection is used to improve oil recovery (Fig 3.74). This can, in some cases, more than double the amount of recoverable oil. The gas used for injection is typically natural gas, N2, or CO2. In the case of gas reinjection, well gas accompanies the oil to the surface, gets separated from the oil and is reinjected back into the reservoir. The composition of the well gas is mostly made up of natural gas along with other gases and impurities. Under the best circumstances, the other gases and impurities are trace amounts and reasonably inert. Under the worst circumstances, the other gases and impurities are a significant fraction and both highly toxic and corrosive. On top of that, both the reservoir conditions and the gas composition can change overtime and the compressor operation and control need to adjust accordingly. In 1974, the highest pressure gas reinjection application was at the Phillips Ekofisk oil field [25]. The Ekofisk compressor was designed for 634 bar (9200 psia) discharge including a discharge gas density of 318.8 kg/m3 (19.9 lbm/ft3). The maximum pressure achieved in the field for this application was 614 bar (8900 psia). Such opportunities do not come along very often, and the Ekofisk compressor would remain the highest pressure application for over two decades. That milestone has been eclipsed a few times, most recently on the Kashagan sour gas project, which achieved a discharge pressure of 828 bar

FIG. 3.74 Reinjection compressor used for oil recovery. (From P.C. Rasmussen, R. Kurz, Centrifugal compressor applications-upstream and midstream, in: Proc. 38th Turbomachinery Symposium, Houston, TX, 2009, Figure 16.)

Centrifugal Compressors Chapter

3

115

(12,000 psig) [25a]. The Petrobras Tupi III project also deserves mention due to achieving 8120 psi (560 bar) during compressor test with an extremely highdensity gas [25b]. The Tupi III compressor is said to have the highest discharge gas density ever achieved in a centrifugal compressor at 556.1 kg/m3 (34.72 lbm/ft3). Interestingly, all three of these projects were by different original equipment manufacturers (OEMs). Most projects do not set new records for pressure or gas density. Fig. 3.75 shows a modern gas reinjection centrifugal compressor. A cross section of this compressor is shown in Fig. 3.76. This compressor was used for gas reinjection at 560 bar (8100 psia). It was tested out to a pressure of 648 bar (9400 psia) with associated discharge density of 426 kg/m3. There are several items worth noting for this compressor. The compressor shown in Fig. 3.74 has the following features: l l

l

Thickened barrel design for high pressure, rated at 700 bar (10,100 psi). Thrust brakes at stages 2–5 for thrust management and rotordynamic stability improvement (Fig. 3.77). Hole-pattern balance piston seal with swirl brakes in front to improve rotordynamic stability (Fig. 3.78).

End users manage their risk by requiring class 1 testing for certain applications. For one end user, applications above roughly 30 MPa get special attention in terms of requiring a class 1 test [26]. Still, other end users may manage their risk differently by requiring rotordynamic stability testing in lieu of a class 1 test [27], rotordynamic stability test in addition to a class 1 test [27a], or

FIG. 3.75 Gas reinjection compressor. (Figure 8 from Y. Bidaut, U. Baumann, S. M. H. Al-Harthy, Rotordynamic stability of a 9500 psi reinjection centrifugal compressor equipped with a hole pattern seal—measurement versus prediction taking into account the operational boundary conditions, in: Proceedings of the Thirty-Eighth Turbomachinery Symposium, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 2009, pp. 251–259.)

FIG. 3.76 Compressor cross section. (Figure 5 from Y. Bidaut, U. Baumann, S.M.H. Al-Harthy, Rotordynamic stability of a 9500 psi reinjection centrifugal compressor equipped with a hole pattern seal—measurement versus prediction taking into account the operational boundary conditions, in: Proceedings of the Thirty-Eighth Turbomachinery Symposium, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 2009, pp. 251–259.)

FIG. 3.77 Thrust brakes.

FIG. 3.78 Hole-pattern balance piston seal with swirl brakes.

Centrifugal Compressors Chapter

3

117

something in between. The compressor shown in Figs. 3.75 and 3.76 had a rotordynamic stability test performed in addition to a full-load, full-pressure, full-density, and full-speed test [28].

Refrigeration Low-temperature service is described in API 617 eighth edition as a service where the specified minimum design metal temperature is below 244 K. Low-temperature service typically requires stringent material certifications including impact testing. One of the colder applications is LNG refrigeration.

LNG All LNG processes run one or more compressed refrigerants through heat exchangers to chill the treated gas stream down to a temperature of roughly 111 K, which results in liquefaction and some subcooling. One such commonly used process is the propane-precooled mixed-refrigerant (C3MR) cycle shown in Fig. 3.79. This process consists of two closed-loop refrigeration streams: a propane refrigeration cycle and a mixed refrigerant (MR) cycle. The propane refrigeration cycle cools the gas to around 238 K through a heat exchanger. It also acts to precool the MR. The MR refrigeration cycle further cools the natural gas from 238 to 111 K in the main cryogenic heat exchanger (MCHE). The multiple streams across the heat exchanger are used to increase the thermal efficiency of the process. These multiple streams of the propane refrigeration cycle can be handled by a single propane compressor arranged with multiple sections with incoming sideloads. Per design, the sideload compressor is well suited for the C3MR refrigeration application where the refrigerant is introduced at progressively lower temperatures to the lower-pressure sections. LNG

Propane MR vapor Feed

Propane precooling

MR liquid Mixed refrigerant FIG. 3.79 C3MR process for LNG refrigeration [29].

118 SECTION

II Types of Equipment

Gas temperature increases in a given section are reduced by the mixing of the sidestream flow into the suction of the following section of the compressor. This maintains gas temperature throughout the machine at reasonable levels [30]. The LNG propane compressor is one of the more difficult compressor applications, typically accommodating a wide operational map with multiple sideload inlets, a high molecular weight gas and low temperatures. The process map in Fig. 3.79 requires a propane compressor that can accommodate four inlets. Fig. 3.80 shows a ¾ cutout of just such a compressor that is used for the C3MR process in a 4.7 MMTPA plant (one inlet is not pictured due to the cutout). The compressor shown in Fig. 3.80 has the following features: l

l l l l

l

Drive-thru shaft with gas turbine main drive at one end and starter motor at other end. Horizontal split casing for simplified maintenance. Five nozzles: one main inlet, three side load, and one discharge. Internal sideload mixing. Unconventional up nozzle requirement necessitates piping removal for maintenance. Not pictured, recycle lines.

Figs. 3.80 and 3.81 show the compressor on its baseplate getting installed and indicate the scale of this large compressor.

FIG. 3.80 C3MR propane compressor. (Courtesy of Elliott Group.)

Centrifugal Compressors Chapter

3

119

FIG. 3.81 Installation of a large propane compressor with three sidestreams. (Courtesy of Omatick and Carpenter, LNG Industry, September 2014, Figure 6.)

Control systems can be particularly difficult for this arrangement. The aerodynamics between each nozzle is known as a section and this compressor has four sections. Each section must be monitored and controlled. Flow monitoring is critical and is performed by at least four flowmeters: one for each section. A fifth flowmeter would be costly, but would enable confirmation of the flows thus determining calibration errors and drift, and this can be quite valuable in maximizing the range. Each stage has an antisurge controller with each controlling one of four recycle valves.

Hydrogen Recycle Hydrogen (H2) recycle—hydrogen recycle compressors are used in a variety of refinery processes including catalytic reforming (platforming), hydrocracking, and hydrotreating. While these different processes have unique operating pressures, temperatures, and catalysts, many basic elements are common. A simple hydrotreating process is shown schematically in Fig. 3.82. This process is used for fuel desulfurization. The process starts with liquid feedstock sent to a heater by a charge pump. The liquid is heated and combined with hydrogen-rich recycle gas coming from the recycle gas compressor then fed to a reactor. Within the reactor, in the presence of a catalyst, sulfur compounds are decomposed to form a hydrocarbon and hydrogen sulfide. Olefins and other unsaturated hydrocarbons are saturated by the additional hydrogen, producing stable hydrocarbons. Hydrogen will further react with nitrogen, oxygen, and chlorine impurities in the feed to form

120 SECTION

II Types of Equipment

FIG. 3.82 Simplified hydrodesulferization process [31].

ammonia, water, and hydrogen chloride. Unfortunately, these products can have further reactions to form various salt compounds that can condense and deposit onto surfaces. The reactor effluent is cooled and flows through a series of gas/ liquid separators. In this example, a steam stripper removes gas and naphtha leaving low sulfur product. The gas from the final separator flows to a hydrogen sulfide (H2S) scrubber where liquid amines are used to remove the H2S from the recycle gas. Gas from the scrubber enters the inlet of the recycle compressor. Since hydrogen is consumed by the chemical reactions, a makeup compressor supplies additional hydrogen to the loop. This compressor is typically a reciprocating unit. The gas entering the compressor is primarily hydrogen plus additional hydrocarbons. Some design challenges for a hydrogen compressor include the following: l l

l

High pressure—up to 17.3 MPa. Rotordynamic considerations—the low molecular weight of hydrogen recycle gas demands many stages (a long rotor) and high rotational speeds for just a moderate pressure rise. Fouling from salts.

A hydrogen recycle compressor is shown in Fig. 3.83. This compressor has 11 stages of compression. The rated discharge pressure is 11.00 MPa. The maximum continuous speed is 14,051 rpm.

Air/Nitrogen Centrifugal compressors are also used for plant air compression and in air separation applications. In air separation, nitrogen and oxygen are separated from

Centrifugal Compressors Chapter

3

121

FIG. 3.83 Elliott 15MB11 hydrogen recycle compressor (DWG 1046291 suggested).

compressed air. Therefore we find air compressors, Nitrogen compressors and Oxygen compressors. Compressing air or nitrogen usually does not put special requirements on the seals, since the compressed gas is nontoxic. The amount of leakage is therefore strictly based on economic and performance considerations. However, being relatively heavy gases, the Mach number levels are often in the high subsonic range. Machines either use adjustable inlet guide vanes or speed control to adapt to the required process conditions. In all applications, keeping the process gas oil free is of great importance Nitrogen compressors, besides in air separation, are also used in small scale LNG applications, where Nitrogen is used as the refrigerant in a reverse Brayton cycle. In other words, Nitrogen is compressed, cooled and expanded through turboexpanders to reach the low temperatures necessary for liquefying natural gas.

Maintenance Practices Design for Maintenance A centrifugal compressor is a long-term capital investment that is generally depreciated over a 30-year span. As such, a compressor and its installation are designed not just for long-term operation, but also for long-term maintenance. Serviceability is highly dependent on being able to access the compressor. Compressor accessibility starts with the plant layout and installation plan. An

122 SECTION

II Types of Equipment

ideal installation for a compressor is to use a “down nozzle” arrangement with inlet and discharge lines coming up from the bottom and a clear aisle all around the equipment string. In this type of arrangement, an overhead crane can be made available ranging the entire footprint of the equipment for maintenance purposes without concern of hitting any piping. A further advantage of the down nozzle arrangement is that the top half of a horizontally split compressor can be easily removed without requiring any piping disassembly. If the compressor is vertically split, sufficient axial space to remove the inner bundle from the compressor is a necessity. In many cases, ideal arrangements are not possible, and some sacrifices are required such as having an up nozzle arrangement or having reduced space around the equipment. Per design, most compressor maintenance only requires rudimentary disassembly. Alignment can be performed by adjustments at the compressor feet. Couplings are accessible by simply removing the coupling guard, and from there can be easily replaced. Journal and thrust bearings are accessible through their own covers without requiring large-scale disassembly. Bearings are horizontally split to allow for easy replacement. End seals tend to require a higher degree of disassembly, but here too, they are designed to be accessible from the ends of the compressor without removing either the top half lid of a horizontally split compressors or inner bundle of a vertically split compressor. Critical auxiliaries such as the lubrication system and DGS support system are also designed for easy maintenance. Most maintenance activities can be performed without taking the compressor offline. The lubrication systems are equipped with duplex filters and twin coolers that can be switched over and isolated via continuous-flow transfer valves. Filters can then be replaced and cooler bundles can be removed and cleaned. The lube oil supply pumps are also spared and can be switched over as well. Likewise, the DGS systems are capable of draining and replacing filter elements concurrent with compressor operation. Maintenance practices can be divided into corrective maintenance, predictive maintenance, and preventive maintenance. Corrective maintenance, also known as breakdown maintenance, occurs due to an unplanned event such as a trip that requires the equipment being taken offline in a forced or unplanned outage to correct a problem. Predictive maintenance, also known as conditionbased maintenance, occurs due to monitoring and trending activities that indicate a potential problem that allows for a planned action to occur. Preventive maintenance occurs at scheduled maintenance intervals such as every 7 years for example. All three types of maintenance have taken important steps forward due to monitoring and inspection technology. Regardless of the maintenance practice, critical spare parts are normally kept on hand including couplings, journal bearings, thrust bearings, seals, and a spare rotor. Spare filters are normally kept on hand for lubrication systems and seal systems.

Centrifugal Compressors Chapter

3

123

Condition Monitoring Compressors and their associated auxiliary systems are instrumented to ensure safety and reliability. Standard alarms, shutdowns, and control systems are discussed in detail within API 614 and API 670 standards. Typical parameters that are monitored online are listed below. Many of these measurements are fed into a machinery protection system. Trips related to overspeed and bearing oil pressure are generally required. Trips related to vibration level and axial position are often specified as well. Typical Instrumentation l l l l

l

l l

Speed Rotor vibration and position (radial and axial) Bearing metal temperatures (journal bearings and thrust bearings) Lube oil console  Lube oil pressure, inlet temperature, and drain temperature  Lube oil differential pressure across the filter  Lube oil level  Lube oil cooling water temperature Dry gas seal  Seal-gas supply  Vent-gas pressure(s) or flow(s)  Separation-gas pressure of flow  Differential pressure at each filter set Lube oil cooling water temperature Compressor performance  Suction and discharge pressure  Suction and discharge temperature  Flow rates

Manual inspections are also of importance. The compressor and the auxiliary equipment should be inspected for leaks at piping connections. Lubricating oil should be periodically checked for the presence of water and for degradation.

Predictive Maintenance Data collection and measurements can be used for predictive maintenance. The concept is to base maintenance decisions, and in some cases, operating decisions, on the measured condition of the machine. Trending information overtime can provide advanced warning of a deteriorating compressor thereby enabling preparation for maintenance and a shortened turnaround cycle. Better, more accurate, production planning can also be achieved. Rotor vibration is measured using noncontacting eddy-current displacement probes that sense the motion of the rotor relative to the casing and bearings. The probes are typically located immediately adjacent to each radial bearing arranged 45 degrees from top dead center and sensing the burnished probe

124 SECTION

II Types of Equipment

area on the shaft. Increasing vibration can be indicative of a number of problems. Rotor unbalance and rotor misalignment are two of the most common causes of vibration. However, there are many other causes of high vibration. Excessive radial vibration can lead to bearing distress and may cause babbitt surfaces to fatigue or internal seal rubs to occur. Axial position is measured using noncontacting eddy-current displacement probes. In this case, the probes are arranged to monitor the rotor shaft end. Changes to axial position can occur over a compressor’s normal operating range and thrust is even known to switch directions. Increased movement beyond a thrust bearing’s clearance envelope and into the alarm region is indicative of a highly loaded thrust bearing where some elastic deflection is occurring. Further movement could result in wiping of babbitted surfaces. Bearing pads are instrumented for the purpose of health monitoring. Temperature transducers are embedded in the bearing backing metal near the babbitt bond line and bearing surface as specified in API 670. High bearing metal temperature can be indicative of potential bearing failure. Bearing metal temperature that rises in an upward trend without corresponding change to load or speed is also indicative of potential bearing failure. Bearing metal temperature at the instrumented location may range from less than 328 K for an unloaded inactive thrust bearing up to 403 K for a bearing operating near its design limits. There is varying opinion with respect to metal temperature limitation. In general, the manufacturer’s recommendation should be followed, especially for new equipment lacking in historical data. Monitoring of auxiliaries is also performed such as lube oil system pressures and temperatures, seal gas differential pressure, and suction scrubber liquid level. Lube oil supply pressure must be maintained within allowable levels to ensure the required oil delivery to bearings. This supply pressure is not a standard value and will differ between vendors. Performance monitoring can also be used for surge control, but can also be used for predictive maintenance as well. Polytropic efficiency deviation is often associated with worn labyrinth seals. Polytropic head deviation can be due to changes in molecular weight or fouling of the compressor rotor. Power deviation can be due to worn seals especially a worn balance piston seal or due to molecular weight changes. Discharge temperature deviation can be due to worn seals especially a worn balance piston seal. Such seal wear as described here can occur when the compressor experiences high vibration excursions from process upsets or operation in a surge condition.

Preventive Maintenance Most maintenance and inspections on the actual compressor are carried out at regular maintenance intervals during planned turnarounds also known as planned outages. In clean services such as closed-loop refrigeration cycles, a centrifugal compressor may operate continuously for 7–14 years between turnarounds provided the compressor and related systems are well designed,

Centrifugal Compressors Chapter

3

125

manufactured, installed, and operated. However, many centrifugal compressors operate in less than ideal services with less than optimal installation and operation. In such cases, more frequent turnarounds are required. Some causes of a shortened turnaround cycle may involve corrosion, fouling, and the maintenance of other machinery or systems. As previously discussed, maintenance that is conducted prior to a planned turnaround is typically centered on predictive maintenance based on condition monitoring. During a planned turnaround, the centrifugal compressor unit should undergo a general and thorough inspection. At a minimum, bearings, couplings, and compressor shaft endseals are inspected during the turnaround. Auxiliary lube and seal systems also require maintenance of miscellaneous items including filters, coolers, pumps, etc. Bearing inspection starts with removing the journal and thrust bearings from the compressor. In almost all compressor designs, these are externally accessible through their own covers. During visual inspection, the bearings are examined for signs of wear or distress. The journal and thrust areas of the rotor will be accessible and also need to be examined for signs of distress such as scoring. The coupling should be inspected for wear and cleanliness. Coupling inspection is typically centered on the flexible elements. Any broken parts, yielding, buckling, corrosion, or scratches should lead to replacement. Couplings can be returned to the manufacturer for recertification. Inspection of the compressor internals requires further disassembly. During a major outage, the rotor should be taken out of the casing to have all deposits removed as well as to determine the degree and rate of corrosion or erosion. The casing should be examined and cleaned if necessary. Labyrinth seals should be examined and replaced if necessary. The auxiliary equipment frequently requires as much or more maintenance work than the main compressor and driver. Maintenance of oil pumps, oil coolers, and filters are often required. Oil pumps are typically arranged to be serviced. Oil cooler tube bundles are designed to be pulled in one direction to facilitate cleaning. And, oil filters are arranged such that they are easily drained and replaced.

Sparing and Availability Critical equipment is defined as one that would shutdown a process unit in case of failure. For critical pieces, it is common that entire units, or possibly part spares be dedicated for that unit to reduce downtime. As such, it is obvious that a cost-benefit analysis be performed in to weigh the risk of equipment failure on lost process time to determine the level of sparing or redundancy. Of course, this analysis is performed more readily in light of operational experience on a given piece of equipment to evaluate the likelihood of a failure. In some circumstances, critical units may be placed in parallel such that they can be brought offline without any disruption to the process.

126 SECTION

II Types of Equipment

Equipment availability is standardized as the ability of an item to be in a state to perform a required function under given conditions at a given instant of time, or in average over a given time interval, assuming that the required external resources are provided [32]. For more information on availability and optimization in oil and gas operation, Brissaud et al. [33] should be consulted.

Construction Options for Maintainability and Upgrades—Optimization and Restage Gas compressors are typically sized in a high-efficiency zone to meet the initial pressure, temperature, gas composition, flow, and other initial conditions. However, the changing of operating conditions, such as gas field depletion and natural gas demand increasing, is the nature for either production or pipeline compressor applications. While centrifugal compressors provide a tremendous flexibility, it is often economic to restage the existing compressor to optimize for new process conditions, gains in compressor efficiency, flow capability, or reduction in fuel or electricity consumption.

Principles of Gas Compressor Restage Typical compressor has to adopt a wide range of operation. If conditions, inlet/ discharge temperature/pressure, flow, speed, and gas properties, make compressor performing the best efficiency point, no restaging is needed. When conditions change in one direction away from the best efficiency point, compressor restaging should be considered (Fig. 3.84). The change of temperature, pressure, and gas composition mainly moves the operating point in the T (speed Topping) or D (speed Decreasing) direction. When suction temperature is increased, the head is increased for the same pressure ratio and higher speed is required to move the flow point to the T direction. Higher temperature also tends to tilt the map in the counterclockwise direction. Suction pressure has similar effect too. For example, in a declining field, the suction pressure reduces overtime. The pressure ratio increases as the same discharge pressure typically is kept the same. More pressure ratio requires higher speed. Meanwhile, as filed pressure drops, the density drops, thus more volume flow passes through the compressor. The operating point moves in the T direction. In the opposite, there are also cases where the suction pressure is increased. The design point moves to the D direction, as the required head reduces. For discharge pressure, when it increases, the pressure ratio increases. The head is increased. The flow point moves in the T direction. When discharge pressure deceases and the pressure ratio decreases, the flow point moves in the D direction. For pipeline transmission, gas composition stays relatively constant. But for gas production applications, gas composition changes. Heavier gas requires less power to reach the same pressure ratio, the compressor speed needs to be

Centrifugal Compressors Chapter

s

50,000

lle r im pe

di

flo

Head or pressure ratio

35,000 30,000 25,000 20,000 15,000 10,000

ers

D: Decreased speed R: Recycle

75% 73% 71%

C

D

Re m

ov ei

mp

ell

ers

100

150

200

lo

rf

e igh

H

Actual inlet flow 50

15,500 rpm

14,500 13,500 s 12,500 er 11,500 ell p 10,500 im 9500 rpm w

5000 0

ell

T

R

C: Choke

mp

Lo

we r

40,000

127

T: Speed topping

Ad

w

45,000

3

250

300

350

400

FIG. 3.84 A typical compressor flow-head map. (Courtesy of Solar Turbines Inc.)

slower. The flow point moves vertically down. The map is also tilted in the clockwise direction. Therefore, the flow point moves in the D direction. Lighter gas behaves opposite. Volumetric flow change can have a very straight forward effect. More flow moves the flow point in the C direction to the Choke side. If at constant power consumption, the flow point moves downward the C direction; if at constant head, the flow point moves horizontally. Compressor efficiency is mainly a function of flow. The efficiency drops fast from the best efficiency point to choke. Lower efficiency increases discharge temperature. More power loss happens due to a less-efficient compressor. On the other side, less flow moves to the recycle side of the map. The efficiency reduction is less rapid than in the C direction. In this case, insufficient flow moves the flow point to the left of the surge line, requiring the antisurge valve open to protect the compressor from surge. In this situation, power is wasted by recycling the gas through the compressor. Restage Applications Power ¼ C 

SQ Hisen ¼ Fuel Energyηengine ηisen ηmech

where Power is driver (engine) output power C is a constant

(3.7)

128 SECTION

II Types of Equipment

SQ is standard flow ηmech is mechanical efficiency ηengine is engine efficiency The energy balance of the whole turbine/engine driven compressor (Fig. 3.85) train can be expressed in Eq. (3.7). For a given head and flow, the power is a function of driver thermal efficiency, compressor efficiency, and mechanical efficiency. When the operating point is away from the compressor best efficiency zone for an extended time, the compressor or engine is running less efficient. The purpose of a gas compressor restage is to optimize the compressor staging in order to maximize efficiency at the new conditions to use less power and deliver higher head and more flow. Speed Topping Increasing discharge pressure for gas injection and decreasing suction pressure for gas gathering are two typical scenarios in which the flow point moves in the T direction. The speed of the compressor needs to be increased to keep up with the increasing pressure ratio. But the driver and compressor have their maximum speeds. When the maximum speed is reached, the demand could not met by simply increasing compressor speed. This is called speed toping. In this case, if there is still room in the compressor, more stages can be added to deliver higher head without increasing speed. For gas gathering in a declining field, that means extended field life. For gas injection, higher pressure means more oil production. Running Slow Normally, the compressor speed is designed to require the power turbine to run over 90% of max speed, in order to reach highest efficiency levels. When the operating point consistently requires turbine speeds lower than

FIG. 3.85 A typical turbine driving compressor package. (Courtesy of Solar Turbines Inc.)

Centrifugal Compressors Chapter

3

129

FIG. 3.86 Gas compressor restage principles.

optimum levels, the compressor needs to be speedup by removing stages to improve turbine efficiency. This type of restaging reduces engine fuel consumption. Compressor Choking Compressor choking is a phenomena happening often due to trend of more and more demand on natural gas. The compressor keeps running at the choke region to deliver more gas. The compressor efficiency is low. Although the compressor may not be physically choked, the available power can limit the capacity throughput due to the low efficiency. When this happens for long time, larger flow stages should be used to replace smaller flow stages. Fig. 3.86 shows a typical restage due to choke using larger flow stages to change the performance map to better match the conditions. The flow capacity is dramatically improved. Gas Recycling Opposite the choke, when the compressor does not get enough flow, the antisurge valve opens to avoid surge. Some amount of discharge flow is cooled to feed back to the compressor. The energy of recycling gas is wasted. Replacing higher flow stages with smaller stages is an effective way to accommodate the lower volumetric flows. After restage, the wasted power can be used to increase pressure ratio for more capacity or more oil production. Besides economic reasons, running in recycle mode causes high discharge temperatures if insufficient cooling is supplied. DGSs, balance piston babbitt, and antisurge valves can be damaged during long period of recycling. Restage Criteria The economic study of payback period requires interaction between the user and the OEMs. Four parameters stood out as good indicators for restaging: inlet flow coefficient (Φ), isentropic head coefficient (Ψ ), inlet pressure (P1), and the required power. The detailed criteria for each parameter are shown in Table 3.3. Generally, for power, suction pressure, and head

130 SECTION

II Types of Equipment

TABLE 3.3 Trigger Points for Restage Parameters Percent Change

50%

Φ1

31%

Ψ1

19%

P1

15%

HP

13%

coefficient, the trigger point for restage consideration for next overhaul is when the parameter drifted 5%–15%. If they drifted more than 15%, that is, the trigger for immediate restage consideration. The flow coefficient trigger points are 25% for next overhaul and 50% for immediate consideration. The other general rule is that a compressor restage is recommended when the efficiency is less than 6% of peak efficiency and power is a limiting factor. Q1 is the inlet flow coefficient, for compressors, using the first Φ1 ¼ ðD 2 Þ3 N compressor inlet flow coefficient. Hisen is the isentropic head coefficient for single body compresΨ isen ¼ ðD2 N Þ2 P k1 2 T1 k  1 for compressors using the total pressure sor, Ψ isen ¼ Cp ðD NÞ2 2 P1 ratio and the first compressor speed and impeller tip diameter.

References [1] T. Allison, J. Moore, A. Rimpel, J. Wilkes, R. Pelton, K. Wygant, Manufacturing and testing experience with direct metal laser sintering for closed centrifugal compressor impellers, in: Proc. of 43rd Turbomachinery & 30th Pump Users Symposia, September 23–25, Houston, TX, 2014. [2] J. Sorokes, Selecting a Centrifugal Compressor, 2013. [3] R. Kurz, The Physics of Centrifugal Compressor Performance, PSIG, Palm Springs, CA, 2004. [4] V. Huetten, T. Krause, V.A. Ganesan, C. Beer, S. Demming, VSDS motor inverter design concept for compressor trains avoiding interharmonics in operating speed range and verification, in: Turbosymposium 2013, Tutorial 10, 2013. [5] Fluid Sealing Association Gasket Division Members, A guide to gasketing principles and best practices, in: Proceedings of the Forty-First Turbomachinery Symposium, September 24–27, 2012, Houston, TX, 2012. [6] J.M. Vance, B. Murphy, F. Zeidan, Machinery Vibration and Rotordynamics, Wiley, Hoboken, NJ, 2010. [7] API Recommended Practice 684, API Standard Paragraphs Rotordynamic Tutorial: Lateral Critical Speeds, Unbalance Response, Stability, Train Torsional and Rotor Balancing, second ed., American Petroleum Institute, Washington, DC, 2005.

Centrifugal Compressors Chapter

3

131

[7a] D.W. Childs, Turbomachinery Rotordynamics, Phenomena, Modeling, and Analysis, John Wiley & Sons, 1993. [8] API Standard 617, Axial and Centrifugal Compressor and Expander-Compressors, eighth ed., American Petroleum Institute, Washington, DC, 2014. [9] B. Venkataraman, D. Moulton, C. Cave, J. Wilkes, J. Moore, Eldridge, Design and implementation of swirl brakes for enhanced rotordynamic stability in an off-shore centrifugal compressor, in: Proc. Asia Turbomachinery and Pump Symposia, May 12–15, Singapore, 2018. [10] J.J. Moore, T.S. Soulas, Damper seal comparison in a high-pressure re-injection centrifugal compressor during full-load, full-pressure factory testing using direct rotordynamic stability measurement, in: Proceedings of the DETC’03 ASME 2003 Design Engineering Technical Conference, Chicago, IL, Sept. 2–6, 2003, 2003. [11] J.C. Wachel, W. von Nimitz, Assuring the reliability of offshore gas compression systems, in: European Offshore Petroleum Conference & Exhibition, October 21–24, 1980, London, England, 1980. EUR 205. [12] J.J. Moore, D. Ransom, Refinement of physics based approach used in the prediction of impeller rotordynamic forces for centrifugal compressors, in: Proc. ASME Turbo Expo 2008, Power for Land, Sea, and Air, GT2008-51484, 2008. [12a] J.J. Moore, D.L. Ransom, Centrifugal Compressor Stability Prediction Using a New Physics Based Approach, in: ASME Turbo Expo 2009: Power for Land, Sea, and Air Volume 6: Structures and Dynamics, Parts A and B Orlando, Florida, USA, June 8–12, 2009. [13] F. Kushner, Disc vibration—rotating blade and stationary vane interaction, ASME J. Mech. Des. 102 (1980) 579–584. [14] F. Kushner, S.J. Richard, R.A. Strickland, Critical review of compressor impeller vibration parameters for failure prevention, in: Proceedings of the Twenty-Ninth Turbomachinery Symposium, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 2000, pp. 143–161. [15] M.P. Singh, B.K. Thakur, W.E. Sullivan, G. Donald, Resonance identification for impellers, in: Proceedings of the Thirty-Second Turbomachinery Symposium, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 2003, pp. 59–70. [16] W. Campbell, The protection of steam turbine disc wheels from axial vibration, Trans. ASME 46 (1924) 31–160. [17] B. Pettinato, J. Griffin, Y. Wang, D. Feiner, B. Echols, M. Cushman, Lecture 10: Experimentally based statistical forced response analysis for purpose of impeller mistuning identification, in: Proceedings of the 2nd Middle East Turbomachinery Symposium, Presented in Doha, Qatar, March, 2013, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 2013. [18] S. Konig, N. Petry, N.G. Wagner, Aeroacoustics phenomenon in high-pressure centrifugal compressors—a possible root cause for impeller failures, in: Proceedings of the Thirty-Eighth Turbomachinery Symposium, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 2009, pp. 103–121. [18a] N. Petry, F.K. Benra, S. Koenig, Experimental study of acoustic resonances in the side cavities of a high-pressure centrifugal compressor excited by rotor/stator interaction, in: Proc. ASME Turbo Expo 2010: Power for Land, Sea and Air (GT2010-22054). [19] Y. Bidaut, D. Dessibourg, The challenge for the accurate determination of the axial rotor thrust in centrifugal compressors, in: Proceedings of the Asia Turbomachinery and Pump Symposium, Singapore, February 22–25, 2016, 2016. [19a] R. Kurz, E.J. Fowler, R.K. Marechale, M. Ji, M.J. Cave, Operation of centrifugal compressors in choke conditions, in: Proc. 40th Turbosymposium, Houston, TX, 2011.

132 SECTION

II Types of Equipment

[20] E. Abraham, H. Miller, Meeting compression train base package design requirements for service on floating production storage and offloading vessels, in: Proceedings of the Asia Turbomachinery and Pump Symposium, Singapore, February 22–25, 2016, 2016. [21] S.L. Ross, R.F. Beckinger, Compressor seal selection and justification, in: Proceedings of the Thirty-Second Turbomachinery Symposium, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 2003, pp. 167–177. [22] ASTM D6439-11, Standard Guide for Cleaning, Flushing, and Purification of Steam, Gas, and Hydroelectric Turbine Lubrication Systems, ASTM International, West Conshohocken, PA, 2017. [22a] API RP 686, Recommended Practice for Machinery Installation and Installation Design, second ed, (December 2009). [23] C. Meher-Homji, P. Pillai, R. Kurz, P. Rasmussen, Short course 7: LNG liquefaction plants— overview, design & operation, in: Proceedings of the Forty-Sixth Turbomachinery Symposium, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 2017. [23a] C.R. Sparks, On the transient interaction of centrifugal compressors and their piping systems, ASME 83-GT-236(1983). [23b] R. Kurz, R. McKee, K. Brun, Pulsations in centrifugal compressor installations, ASME Paper No. GT2006-90700. (2006). [23c] K. Brun, M. Nored, R. Kurz, Impact of the piping impedance and acoustic characteristics on centrifugal compressor surge and operating range, ASME Paper No. 2014-2504. (2014). [24] P.C. Rasmussen, R. Kurz, Centrifugal compressor applications-upstream and midstream, in: Proc. 38th Turbomachinery Symposium, Houston, TX, 2009. [24a] R. Kurz, S. Ohanian, K. Brun, Compressors in high pressure pipeline applications, ASME GT2010-22018. (2010). [25] C. Geary, L. Damratowski, Evolution of high-pressure gas-injection centrifugal compressors, J. Pet. Technol. (1977) 630–638. [25a] TMI Staff and Contributors, A new high in compression pressures, Turbomachinery Magazine, (December 31, 2017). [25b] G.M. Colby, T.R. Griffin, M.K. Gupta, H.F. Miller, S.E. Nove, N.H. Sehlstedt, High pressure CO2 compressor testing for Tupi 1, Tupi 2, and Tupi 3, Paper No. GT2012-70137, pp. 1025–1032, https://doi.org/10.1115/GT2012-70137. [26] M. Sandberg, Discussion group: advanced centrifugal compressors, in: 40th Turbomachinery & 27th Pump Users Symposia, 2011, Houston, TX, 2011. [27] B.C. Pettinato, C.H. Cloud, R.S. Campos, Shop acceptance testing of compressor stability and theoretical correlation, in: Proceedings of the Thirty-Ninth Turbomachinery Symposium, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 2010, pp. 31–42. [27a] J.J. Moore, S.T. Walker, M.J. Kuzdal, Rotodynamic stability measurement during full-load full-pressure testing of a 6000 psi reinjection centrifugal compressor, in: Proceedings of the 31st Turbomachinery Symposium 2002, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 29–38. [28] Y. Bidaut, U. Baumann, S.M.H. Al-Harthy, Rotordynamic stability of a 9500 psi reinjection centrifugal compressor equipped with a hole pattern seal—measurement versus prediction taking into account the operational boundary conditions, in: Proceedings of the Thirty-Eighth Turbomachinery Symposium, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 2009, pp. 251–259. [29] M. Pillarella, Y.N. Liu, J. Petrowski, R. Bower, The C3MR Liquefaction Cycle; Versatility for a Fast Growing Ever Changing LNG Industry, Air Products and Chemicals Inc., Pennsylvania, 2007

Centrifugal Compressors Chapter

3

133

[30] M. Sandberg, Centrifugal compressors: matching the configuration to the application, Turbomach. Int. 58 (5) (2017) 20–24. [31] S.L. Ross, M.T. Gresh, R.M. Kranz, Compressor seals for hydrogen recycle service, Seal. Technol. 2003 (5) (2003) 5–8. Elsevier. [32] International Electrotechnical Commission (IEC), Dependability and quality of service, in: IEC 60050-191: International Electrotechnical Vocabulary, second ed., IEC, Geneva, 2001 (Chapter 191). [33] F. Brissaud, H. Varela, B. Declerck, N. Bouvier, Production Availability Analysis for Oil and Gas Facilities, Concepts and Procedure, (2015). HAL Archives, HAL Id: hal-01107310. [33a] A. Nakayama, H.-P. Kreplin, H. Morgan, An experimental investigation of flowfield about a multielement airfoil. In 15th Aerodynamic Testing Conference, San Diego, CA, USA. https://doi.org/10.2514/MADT88. Read More: https://arc.aiaa.org/doi/10.2514/6.1988-2035.

Further Reading [34] API Standard 614, Lubrication, Shaft-Sealing and Oil-Control Systems and Auxiliaries, fifth ed., American Petroleum Institute, Washington, DC, 2008. [35] API Standard 670, Machinery Protection Systems, fifth ed., American Petroleum Institute, Washington, DC, 2014. [36] API Recommended Practice 686, Recommended Practice for Machinery Installation and Installation Design, first ed., American Petroleum Institute, Washington, DC, 1996. [37] M.T. Gresh, Compressor Performance: Aerodynamics for the User, second ed., ButterworthHeinemann, 2001. ISBN-13: 978-0750673426. [38] S. Harvey, Centrifugal compressors in ethylene plants, CEP Mag., American Institute of Chemical Engineers (2017) 28–32. [39] A. Kumar, M. Sabharwal, Save your centrifugal machinery during commissioning, in: 43rd Turbomachinery & 30th Pump Users Symposia, September 23–25, 2014, Houston, TX, 2014. [40] R. Kurz, Natural gas compression, in: S. Mokhatab, W.A. Poe, J.G. Speight (Eds.), Handbook of Natural Gas Transmission and Processing, Gulf Professional Publishing, Elsevier Inc, 2006. [41] J.J. Moore, S.T. Walker, M.J. Kuzdal, Rotodynamic stability measurement during full-load full-pressure testing of a 6000 psi reinjection centrifugal compressor, in: Proceedings of the Thirty-First Turbomachinery Symposium, Turbomachinery Laboratory, Texas A&M University, College Station, TX, 2002, pp. 29–38. [42] J.J. Moore, A. Lerche, H. Delgado, T. Allison, J. Pacheco, Development of advanced centrifugal compressors and pumps for carbon capture and sequestration applications, in: Proceedings of the Fortieth Turbomachinery Symposium, September 12–15, 2011, Houston, Texas, 2011. [43] T. Omatick, K. Carpenter, The heart of liquefaction, LNG Ind. (2014) [44] PetroWiki, http://petrowiki.org/Centrifugal_compressor#Horizontally_.28axially.29_split_ case. [45] B.C. Pettinato, J.A. Kocur, E.E. Swanson, Evolution and trend of API 617 compressor Rotordynamic criteria, Turbomach. Trans. Turbomach. Soc. Jpn. 39 (5) (2011) 36–47. [46] B.C. Pettinato, Steam turbines, in: G.E. Totten (Ed.), Handbook of Lubrication, second ed., Application and Maintenance, vol. I, CRC Press, Inc., Boca Raton, FL, 2006 [47] M. Thorat, B. Pettinato, P. De Choudhury, Metal temperature correlations in tilting pad journal bearings, ASME J. Eng. Gas Turbines Power 136 (2014) 11. [48] TRIZ, http://triz-ltd.com/en/produktsiya/rotora/, 2018. [49] TTMI Staff and Contributors, A new high in compression pressures, Turbomach. Int. (2017)

Chapter 4

Integrally Geared Compressors Aoron Rimpel*, Karl Wygant†, Rob Pelton†, Christian Wacker‡ and Kolja Metz‡ *

Southwest Research Institute, San Antonio, TX, United States, †Hanwha Power Systems Americas Inc., Houston, TX, United States, ‡MAN Energy Solutions, Berlin, Germany

Introduction Integrally geared compressors (IGCs) have been in use worldwide since before the 1950s. Early applications were primarily in plant air service, but IGCs have been used increasingly in the process industry since the 1980s [1, 2]. The term “integrally geared” refers to the fact that IGCs enable the rotational speed change from the input device (motor, etc.) to the driven component (compressor impeller) within the frame of the machine—that is, rather than with a separate gearbox unit and high-speed shaft coupling required by other compressor architectures. This typically allows an IGC to have a smaller footprint than other compressor types. In general, integrally geared (IG) architecture can be used for any type of turbomachine—including centrifugal compressors, axial compressors, expanders, etc.—but most applications involve centrifugal compressors. An IGC comprises a bull gear shaft and one or more pinion shafts, with impellers attached to each end of the pinion shafts—see Fig. 4.1. IGCs with up to 5 pinions and 10 compressor stages are not uncommon. Typically, the bull gear shaft is driven by an electric motor at a fixed speed, although different drivers and variable speed operation are possible. For motor-driven IGCs, the bull gear typically runs at 1500–3600 rpm, allowing a standard, low-cost driver to be used. The bull gear then drives the individual pinions at higher relative speeds, depending on their respective gear ratios, which may be over 50,000 rpm in the latter stages of compression. It may also be possible to have more than one bull gear meshed with one another, each driving separate pinions. The basic configuration of the IGC rotating assembly is depicted in Fig. 4.1 and shows the bull gear shaft driving two separate pinion shafts and a total of four impellers. The bull gear is shown with a dedicated thrust bearing on one side of the gear, and each of the pinions utilize thrust collars. Other bearing configurations are possible and are discussed further within the topics of bearings and thrust management later. Compression Machinery for Oil and Gas. https://doi.org/10.1016/B978-0-12-814683-5.00004-3 © 2019 Elsevier Inc. All rights reserved.

135

136 SECTION

II Types of Equipment

FIG. 4.1 Schematic of IGC shaft architecture.

An exploded view of a generic IGC is shown in Fig. 4.2. Note that the individual volute casings attach to the outside of the gearbox housing, and the impellers, diffusers, seals, etc., can all be assembled independently for each stage. This illustrates an advantage of IGCs in terms of modularity and maintenance. The arrangement also allows relatively easy access to the inlet and diffuser sections of the impeller, which permits the easy integration of variable inlet guide vanes (VIGVs) and/or variable diffuser vanes (VDVs) for each stage. The unique design characteristics of IGCs and their advantages and disadvantages will be discussed in detail in comparison to other compressor architectures.

Comparison With Other Compressor Architectures Advantages IGCs are used in a wide range of industrial applications, including some traditionally covered by inline (also referred to as beam style) centrifugal and reciprocating compressors. The unique architecture of an IGC sets it apart from other styles of compressors and gives it certain performance advantages, including: l l

optimal rotational speed can be selected for each stage, ability to intercool between each stage,

Integrally Geared Compressors Chapter

4

137

FIG. 4.2 Exploded view of IGC. (From K. Wygant, J. Bygrave, W. Bosen, R. Pelton, Tutorial on the application and design of integrally geared compressors, in: Proceedings of Asia Turbomachinery & Pump Symposium, February 22–25, Singapore, 2016.)

l l l l l

simple integration of VIGVs and VDVs, modularity and simpler restaging, multiple process streams can be incorporated within a single IGC, smaller footprint, and reduced interstage pressure loss

The following sections elaborate further on these advantages.

Design Speed The fundamental difference between IGCs and traditional inline centrifugal compressors is that the speeds of the compressor stages can be more closely matched to their optimal aerodynamic speeds. In an inline compressor, all of the stages are on the same shaft and operate at the same rotational speed. This makes achieving high efficiency in a multistage inline compressor challenging since the later stages may be required to operate at a very low flow coefficient where high frictional and leakage losses dominate. However, IGCs have at most two stages of compression on a single pinion shaft that operate at a common rotational speed. Other pinions in the machine can be designed to operate at different speeds optimal to the respective compressor stages, which yields better overall efficiency at a reduced stage count. Since the optimal impeller speed increases for each subsequent stage of compression in a typical IGC, consecutive stages are positioned on a common pinion: for example, stages 1 and 2 on pinion 1, stages 3 and 4 on pinion 2, etc. In some applications, only one stage is mounted on a pinion to give even more flexibility on stage speed in the design; however, this results in greater net thrust force that has to be managed. More discussion on thrust management in IGCs is presented in a later section.

138 SECTION

II Types of Equipment

Intercooling In each stage of compression, the increase in fluid density caused by the increase in pressure is partially offset by an increase in temperature. Thus, the temperature rise decreases the average fluid density compared to the isothermal compression case, and it is well known that this increases compression work. Naturally, this problem is magnified in a multistage compression process. Intercooling refers to the installation of a heat exchanger between the exit of one stage and the inlet of the subsequent stage in order to remove the heat of compression from the previous stage or stages, thereby increasing the average fluid density over the compression process and reducing compressor power consumption. For inline compressors, intercooling is done with a back-to-back configuration between the last stage of the first section and the first stage of the second section. This is practical because back-to-back machines have case penetrations and utilize external piping between these two points. However, straight-through inline compressors do not have a means to incorporate a traditional heat exchanger for intercooling because the only case penetrations are for the inlet to the first stage and exit of the last stage. By contrast, IGCs have the ability to be intercooled between every stage since piping exists between the exit/inlet of every stage, which can offer the best thermodynamic efficiency for the total compression process. As an example, Fig. 4.3 shows the compression process for a 13:1 pressure ratio (PR) air compressor. The minimal compression work is represented by the theoretical line of isothermal compression. It can be seen that the actual compression work is minimized by increasing the number of compression and intercooling stages. In general, IGCs can accomplish the same overall compression with fewer stages than an inline compressor due to the cooler inlet temperature and with improved efficiency since the additional intercooling brings the total compression process closer to isothermal. Of course, there is no “free lunch”: heat exchangers introduce pressure losses and require energy to flow cooling fluid through the nonprocess side. Therefore, the thermodynamic benefits of intercooling must be optimized with these penalties in mind.

FIG. 4.3 Different compression processes with intercooling.

Integrally Geared Compressors Chapter

4

139

Effect of intercooling on compression work (Overall PR 13:1 - air) 110% 105%

Relative shaft power

100% Intercooled 95%

No intercooling

90% 85% 80% 75% 70%

1

2

3

4 Stages

5

6

7

FIG. 4.4 Comparison of different number of intercooling stages.

Figure 4.4 shows a comparison of the relative compression work for a notional 13:1 PR air compressor in two configurations: one with intercooling between every stage, and one with no intercooling (although it is noted that in an actual application, at least one stage of intercooling would be necessary to keep the process gas temperature within reasonable limits). A number of stages from one to six is considered, although configurations with just one or two stages are typically not practical due to mechanical and aerodynamic limitations—these are shown with a dotted line for reference only. The minimum compression work is achieved when an adequate number of stages are used to allow for efficient compression, but no more than necessary to avoid additional interstage pressure losses. This trade results in an optimum number of stages where overall compressor power is a minimum—for example, four stages for the intercooled case in Fig. 4.4.

Variable Geometry The typical IGC pinion shaft arrangement, having the impellers outboard of the seals and bearings, allows access for the application of VIGV mechanisms at the inlet of each stage (Fig. 4.5A), if desired, and similar accessibility exists for VDVs (Fig. 4.5B) with similar actuation mechanisms. In both cases, a single actuator rotates a control ring, which subsequently rotates each IGV or VDV in unison via identical cam features. Variable geometry can be used to increase the peak performance or operating range of a compressor, as is demonstrated in a later section. Because of the relative ease of access, VIGVs and VDVs are significantly more cost effective to implement on IGCs than with typical inline

140 SECTION

II Types of Equipment

FIG. 4.5 Model of typical variable IGV assembly (A) and photograph of variable diffuser vanes (B). (Adapted from K. Wygant, J. Bygrave, W. Bosen, R. Pelton, Tutorial on the application and design of integrally geared compressors, in: Proceedings of Asia Turbomachinery & Pump Symposium, February 22–25, Singapore, 2016.)

compressors. However, while this accessibility exists to varying degrees for every stage in an IGC, it is common to only implement VIGVs and/or VDVs for the first stage.

Multiple Process Streams Another unique capability of IGCs is the ability to handle more than one gas process in the same unit. This can be accomplished since the individual stages are independent from each other, so in a single gearbox, there can be multiple compression or expansion processes. Different compression processes are usually, but not always, accomplished on different pinions. Expansion stages can also be added either opposite a compressor stage on a single pinion or on a separate pinion. Incorporating multiple processes into a single unit can reduce both capital and operating expenditures for many types of plants. Size Fig. 4.6 shows a comparison of two compressor units—an IGC and a reciprocating compressor—sized for the same duty for boil-off gas (BOG) applications. Note that the IGC is significantly smaller than the reciprocating compressor package and also requires a much smaller foundation. Compared to inline compressors, an IGC package is also typically smaller since fewer stages are needed for a given compression process. Also, since IGCs integrate directly with the gearbox, the overall package can be smaller than an equivalent capacity inline centrifugal compressor.

Integrally Geared Compressors Chapter

4

141

FIG. 4.6 Size comparison of IGC and reciprocating compressor sized for same duty. (From K. Wygant, J. Bygrave, W. Bosen, R. Pelton, Tutorial on the application and design of integrally geared compressors, in: Proceedings of Asia Turbomachinery & Pump Symposium, February 22–25, Singapore, 2016.)

Pressure Losses The inlets of inline centrifugal compressors must distribute the flow around the circumference of the impeller from radial collector or tangential volute case penetrations. This is because the bearings are outboard of the compressor stages, and the shaft passes through the center of the impeller. In contrast, IGC impellers are mounted outboard of the bearings and seals on the pinion shaft. This permits the use of an axial run of pipe for the inlet, which has relatively lower pressure loss. The return bends and return channels between stages of inline compressors are also a source of pressure losses because of the amount of flow turning that takes place. IGC impellers discharge into collectors or volutes, which have slightly lower losses. Disadvantages While the previous section highlighted many of the advantages that IGCs have over other compressor architectures, they also have notable drawbacks. The following sections discuss the most prominent ones.

Sealing In an IGC, a shaft seal is required for each stage of compression, while only two seals are needed in an inline centrifugal compressor, regardless of the number of stages. The additional seals in an IGC tend to add to the overall cost of the machine and can adversely affect reliability. Although, as IGC technology has matured, the reliability of the machines with many seals has greatly improved.

142 SECTION

II Types of Equipment

Bearings Like with seals, IGCs require many more bearings than inline compressors. The journal bearings in IGCs operate at similar loads and speed to those in inline compressors, although the main contribution to the static load in IGCs is due to power loads transmitted through the gear as opposed to mainly gravity loads in an inline compressor. Since the bearing loads in IGCs vary with aerodynamic power, the rotordynamic analysis must also consider off-design operating loads and bearing characteristics to ensure acceptable stability at all potential running conditions. Service Servicing an IGC machine can be more involved due to the greater number of seals and bearings. In small IGCs, servicing the bearings and seals requires removing the top cover of the gearbox. However, in large IGCs, it may be possible to inspect and replace seals and bearings without removing the top of the gearbox or stage casings. Access to an IGC is typically more challenging than a typical inline compressor since there is interstage piping to disassemble for each stage.

IGC Design Topics Design Process As discussed by Wygant et al. [3], an engineered IGC essentially follows the same development process regardless of manufacturer or application. The first portion of that process optimizes the thermodynamic cycle and the aerodynamic performance. Next, the mechanical implications are considered and are iterated against the aerodynamic components for the optimal configuration. Fig. 4.7 shows the interconnectivity of various constraints that must be iterated upon to achieve a satisfactory design. The importance of the aerodynamic design cannot be overstated, as shown in the “satellite configuration” on all the aspects that are influenced by aerodynamics. Likewise, errors or poor assumptions in other aspects can result in problems that prevent operation of the unit altogether. Some issues that prevent operation are bearing over-temperatures, excessive rotor vibration, impeller fractures, excessive seal leakage, etc. The sizing and selection of a compressor for a particular application depends on matching aerodynamic performance to cycle requirements and integrating mechanical limitations into the aerodynamic design. Wygant et al. [3] provides a general description of the sizing and selection approach, which is as follows: l

Step 1: Determine the number of stages and intercoolers. The number of stages is determined based upon the overall head or PR. Higher head requirements can lead to increased torque requirements per pinion and result in unacceptable bearing and gear loads.

Integrally Geared Compressors Chapter

4

143

FIG. 4.7 Component connectivity for the design process. (Adapted from K. Wygant, J. Bygrave, W. Bosen, R. Pelton, Tutorial on the application and design of integrally geared compressors, in: Proceedings of Asia Turbomachinery & Pump Symposium, February 22–25, Singapore, 2016.)

l

l

l

Step 2: Determine impeller speeds and diameters. Structural limitations of the rotating impellers are considered based on the impeller tip speed and material constraints. Step 4: Determine the aerodynamic power and gear load requirements for each stage. The aerodynamic power of each stage on a pinion sums to the total mechanical loading that the gear teeth must transmit. These gear forces must also have a reactionary component, which is the bearing. Therefore, the load capacity of the bearings is frequently the limiting factor for power transmission of a pinion. Step 5: Size the bearings. To minimize mechanical loss, the pinion bearing journal diameter should be as small as possible, that is, minimizing the journal surface speed and lubricant shear velocity. To handle the maximum load, it is common to maximize the axial length of IGC pinion bearings. The primary issue is to ensure that the bearings have sufficient area to handle the reaction load from the gears as well as offering adequate stiffness and

144 SECTION

l

l

II Types of Equipment

damping to ensure a stable rotor design. For the bull gear bearings, the limiting factors are shaft strength, lateral rotordynamics, and torsional rotordynamic considerations. Step 6: Determine total compressor power and mechanical losses. After preliminary selection of the bearings and gears, it is possible to estimate parasitic losses that will be present. These include losses in the gear meshes and the bearings, which are evident in the increase in oil temperature. The parasitic losses are added to the stage power, plus additional power margin, to determine the required driver power. Step 7: Evaluate secondary gas path leakage. A multistage IGC has more seals than a comparable inline compressor. As previously noted, an inline compressor must only seal at the inlet of the first stage and the discharge of the last stage, while each stage of an IGC requires its own shaft seal. The leakage from the gas path must be assessed to accurately estimate the performance of the unit. The leakage rate will depend upon the type of seal applied and the pressures. The greater the pressure, the greater the sealing challenge. Once the leakage rate is known, this can be compared to original estimates of leakage to determine if additional iteration is required.

Standards When considering the oil and gas industry, the primary design concerns are related to safety and reliability. IGCs and other IG turbomachinery, while having superior performance, do have increased mechanical complexity as related to inline machines. In the design process, the need to balance safety and reliability constraints with capital and operating expenditures drive many design decisions. To achieve an optimal design requires balancing the aerodynamic considerations previously mentioned with mechanical concerns. API 617 [4], Chapter 3 is applied to critical applications. API 672 [5] is a more general design criteria, still relevant to the oil and gas industry, but limited more to air machines as opposed to other process gases. In more detail, the ASME Boiler and Pressure Vessel Code [6] is applied to pressure boundaries and the casing, and AGMA 6011 [7] is applied to high-speed gear units. To a great extent, each manufacturer applies their own design criteria to the impeller static and fatigue life assessments.

Drivers Since the central bull gear of an IGC rotates at relatively low speed, the most common drivers for IGCs are directly coupled electric motors. Depending on the machine size, motors with two, four, or six poles are used. Both induction motors and synchronous motors are chosen as drivers for IGCs, with induction motors preferred up to 12–15 MW and synchronous motors for higher power

Integrally Geared Compressors Chapter

4

145

levels. Regardless of the type of motor, a variable frequency drive (VFD) can be used to achieve compressor control via speed variation. Another important driver family is turbines, especially steam turbines. The relatively high speed of turbines must be adapted to the bull gear speeds of an IGC either by way of an intermediate speed decreasing gearbox or by a dedicated gear pinion in the IGC itself. The advantage of an integrated driver pinion is the reduction of train elements and the compactness of the arrangement. This arrangement is in use for IGCs with up to eight compressor stages. In addition, a compressor stage can be mounted on the other end of the driver pinion. The driver selection has considerable impact on the gear design since each driver type exerts specific loads onto the gearbox, and this is reflected in the recommendation for service factors for the different driver types. An induction motor has the lowest requirements on the gearbox. The oscillating torque amplitudes at start-up are relatively low, and the only other loads that have to be considered are electrical faults within the motor. Since these occur very infrequently, they can be taken into account with a low number of load cycles, which means they may not be critical for the design. A synchronous motor, by contrast, excites the torsional natural frequency of the compressor train during each run-up, which may lead to transient torque amplitudes in the train elements several times higher than the rated torque. All train elements and the gear teeth have to be designed to withstand these torques for the desired number of starts, meaning that start-up is usually the critical design case for an IGC with a synchronous motor driver. This effect can be mitigated by using a soft-starting device. The use of a VFD, either for start-up or for normal operation, leads to high frequency ripple superimposed on the steady-state torque. Torsional natural frequencies of any shaft in the compressor system must either not coincide with any of the excitation frequencies or be proven to be sufficiently damped. Finally, a steam or gas turbine can produce steady-state torque which is greater than the rated torque for a considerable period of time. This, too, has to be considered in the gearbox design.

Aerodynamics A typical stage in an IGC for oil and gas applications consists of an axial flow inlet, IGVs (usually only the first stage), a radial flow impeller, a radial diffuser, and an overhung volute. A wide range of fluids are seen in oil and gas applications. When sizing a compressor stage, both mechanical and aerodynamic limitations must be met. The maximum impeller tip speed, U2, is typically used as a preliminary mechanical guideline to ensure that stresses in the impeller do not exceed the material limitation, and this also depends on the type of impeller. As a rule of thumb, covered impellers are applied up to approximately 360 m/s, and open impellers are used up to 500 m/s. While higher tip speeds are possible, they are not common in oil and gas applications since the reliability and service life decreases with an increasing tip speed. The machine Mach number, MU2,

146 SECTION

II Types of Equipment

100% 90%

Vaneless diffuser

Operating range: (m· stall – m· choke)/m· choke

80% 70% 60% 50% 40% 30% 20% Vaned diffuser

10%

Stage test results

0% 0.4

0.5

0.6

0.7 0.8 0.9 1.0 1.1 Impeller mach number (MU2)

1.2

1.3

1.4

FIG. 4.8 Expected range of single-stage compressor for MU2 ¼ 0.5–1.4.

defined as the tip speed divided by the inlet sonic velocity, is often used as a basic aerodynamic constraint. As MU2 increases, the range and efficiency of a stage fall, as shown in Fig. 4.8. Based on this understanding, an appropriate number of stages for a given application can be selected. As more stages are used, MU2 is reduced and the range and efficiency improve, although additional interstage losses accrue. Once a specific number of stages and the work distribution are determined, the detailed aerodynamic design of each impeller can be defined. The appropriate impeller flowpath design is primarily a function of the flow coefficient, which is a normalized form of the volume flow rate Q: ϕ¼π 4

Q D22 U2

(4.1)

High-flow coefficient applications will use three-dimensional (3D) impeller geometry with an axial inducer and a conventional radial discharge. Splitters may also be used with high-flow designs. As the flow coefficient decreases, the impeller geometry transitions to a more two-dimensional (2D) design, without an axial inducer and blading only controlling the flow through the radial portion of the stage. Table 4.1 shows an example sizing of a single-stage compressor for five different fluids. Each stage is sized at a flow coefficient of 0.1 and limited to either a machine Mach number of 1.15 or a tip speed 400 m/s (as indicated by bold text in the table). The first sizing case, with air, shows that at standard inlet

Integrally Geared Compressors Chapter

4

147

TABLE 4.1 Compressor Sizing Showing Design Limitations for Different Gases Fluid

–

Air

Hydrogen

CO2

Methane

Propane

Inlet pressure

bar

1.0

1.0

1.0

1.0

1.0

Inlet temperature

°C

20

20

20

20

70

Pressure ratio, PR

–

2.58

1.08

2.42

1.76

2.25

Machine Mach number, MU2

–

1.15

0.31

1.15

0.90

1.15

Flow coefficient, ϕ

–

0.100

0.100

0.100

0.100

0.100

Tip speed, U2

m/s

400

400

306

400

306

Rotational speed

rpm

20,900

5550

17,350

15,650

16,100

Tip diameter, D2

mm

366

1377

336

488

363

conditions, a PR of 2.58 can be achieved, and the impeller tip speed and machine Mach number constraints are both near the maximum. At the same conditions, a stage designed for compression of hydrogen could only achieve PR of 1.08, but the machine Mach number would be very low—only 0.31. With relatively low aerodynamic loads, a hydrogen compressor is more limited by mechanical constraints since the high tip speeds would likely reach the limits of the impeller material before the aerodynamic loading reaches a level where performance would be adversely effected. A stage compressing carbon dioxide would achieve a PR of 2.42, similar to air, but it would be limited by the acceptable machine Mach number, and the impeller rotational speed would be about 75% that of an air compressor. Table 4.1 also shows the physical difference in impeller size and speed that would be required for these different applications.

Flow Control IGCs are usually designed to deliver gas at a constant pressure over a range of flow rates. A machine is usually sized to deliver the largest required volume flow specified at either design or off-design conditions. Off-design operation

148 SECTION

II Types of Equipment

FIG. 4.9 Comparison of compressor maps with VIGV and VDV control.

may be a different mass flow, a change in inlet temperature or pressure, and potentially changes in gas properties. Adequate flow control can be achieved by using one or more of techniques, including incorporating stage bypass, inlet throttle valves, and variable geometry (VIGVs or VDVs). VIGVs add swirl (C1) to the inlet flow, which directly reduces the work according to the Euler turbomachinery equation shown in Eq. (4.2). As the IGVs close, the flow of the compressor is reduced while maintaining the required constant discharge pressure: Wx ¼ U2 C2  U1 C1

(4.2)

VDVs are also used to extend low-flow operating range. They function by matching the diffuser to the impeller exit flow state at all operating conditions to minimize pressure loss and to postpone stall in the diffuser inlet and impeller exit. VDVs can potentially offer greater throttle range than IGVs (see Fig. 4.9) but are generally more complex and costly to integrate into the design.

Gears An IGC usually consists of a central bull gear with several surrounding pinion shafts, which may have different functions. The most frequent function, of course, is the mounting of impellers on the shaft end. High-speed drivers, such as turbines, can also be connected to any pinion end via a coupling. A pinion shaft can even carry a driver coupling on one end and an impeller on the other. For large IGCs, where the required shaft center distance between the bull gear and any one of the pinions is too large, idler gears may be used. In more complex trains of several turbomachines, the pinion with the turbine coupling can also be located between the bull gear and a compressor pinion while the low-speed part of the train is connected to the bull gear.

Integrally Geared Compressors Chapter

4

149

The number of split lines in the casing is dependent on the number and position of pinions. The primary split line is usually located on the bull gear centerline, which is also coincident with the locations of the first two pinions. If an IGC has a third pinion, this is usually located in a separate split line above the primary split line. This secondary split line can hold the fourth pinion as well if volute casing sizing permits. Turbine-drive pinions are usually located in a plane beneath the bull gear. The turbine pinion may be inserted axially through a large assembly opening. In this configuration, a separate casing split line is not necessary. Some examples of these gear casing configurations are shown in Fig. 4.10. Gears in an IGC are typically single helical gears. The teeth are designed for all operating loads, including expected malfunctions like short circuits in electric motor drivers. Depending on the inertia of bull gear and pinions, start-up events may be the limiting case for the gear design. Some of the gearbox design parameters—for example, number of teeth, helix angle, and material properties—can be chosen freely in the design process, while others are the result of calculations according to the standards such as API 613 [7], AGMA 6011 [8], and ISO 6336 [9]. These calculations also take into account whether the gear teeth have to take loads on one or both flanks. If one gear only drives or only gets driven, then only one flank of the teeth is loaded (Fig. 4.11A). However, when one gear is driven by a second gear and also drives a third gear, then both tooth flanks are loaded (Fig. 4.11B). For example, this may be the case when a bull gear is driven by a turbine pinion and drives one or more compressor pinions. To obtain sound design, some iteration may be necessary to balance tooth geometry in terms of tooth width and elastic modulus (which determines tooth size and height). When the general design is fixed, additional calculations to determine the final tooth geometry for grinding are performed. These calculations take into account possible shaft misalignment and deflection.

FIG. 4.10 Gear casing configurations.

150 SECTION

II Types of Equipment

FIG. 4.11 Gear tooth loading (A) on one flank and (B) on both flanks.

Thrust Management It is extremely rare to use a double helical gear in an IG turbomachine because the net pinion thrust (i.e., sum of axial impeller gas forces) would act to load one face of the helical gear more than the other. Therefore, almost all IG turbomachinery use a single helical gear. Furthermore, thrust bearings are almost exclusively fluid film-type configurations with the oil and gas industry. To control the axial position of the large bull gear, a double-acting tapered-land thrust bearing is most commonly applied. The relatively low speed of the bull gear means that minimal fluid shear losses are present in the bull gear shaft thrust bearing. To control the axial position of the pinion shafts, either thrust collars or thrust bearings are applied. Fig. 4.12 shows a pinion with thrust collars. Thrust collars transmit the net axial force from the aerodynamics and the gear mesh to the bull gear disk axial surface. The net residual thrust is then reacted against the bull gear thrust bearing (a lower speed and lower loss mechanism). The area of the load transmission is relatively small as it is formed by overlapping sections of the outer diameters from the bull gear and a thrust collar on the pinion gear (Fig. 4.13). A hydrodynamic oil film is established to keep the bull gear face separate from the pinion thrust collar. This is similar to a plain thrust bearing, with the difference being that relative motion between the bull gear face and the thrust collar face must be considered as both surfaces are in rotation about different axes. San Andres et al. [10] show an approach to thermal-mechanical and dynamic assessment of thrust collars to understand their performance characteristics. Fig. 4.14 shows a pinion shaft and combination journal/thrust bearings. The thrust bearings on the pinion shaft transmit the net axial load from the aerodynamics and gear axial force through the thrust bearing to the static bearing housing. The advantage of the high-speed thrust bearing is that large axial loads can be mitigated; however, this comes at the cost of higher fluid shear losses than a thrust collar.

Integrally Geared Compressors Chapter

4

151

FIG. 4.12 Example of pinion gear-shaft with thrust collars. (Courtesy Hanwha Power Systems Americas Inc.)

FIG. 4.13 Thrust collar load area.

FIG. 4.14 Example of pinion gear-shaft with thrust bearings. (Courtesy Waukesha Bearings.)

Bearings The shafts of an IGC are supported by hydrodynamic fluid film journal bearings. The type of hydrodynamic bearing used is mainly dependent on bearing load and bearing velocity. High-speed shafts are usually supported by tilting pad bearings, whereas low-speed shafts, like bull gear shafts or turbine drive pinions, are supported by multilobe fixed pad bearings.

152 SECTION

II Types of Equipment

FIG. 4.15 Diagram of forces on pinion shaft.

The bearing load acting on each bearing is the result of several different forces: not only is there the weight of the rotor to be supported, but, as discussed previously, there are also the reaction forces from the gear tooth interactions and thrust collars. Fig. 4.15 shows how gear tooth forces, thrust forces, and bearing loads are connected. In the upper part of Fig. 4.15, it is shown how the tangential tooth force, FT, driving the pinion has to be counteracted by the bearing reaction forces, FBG. The sum of the bearing forces has to equal the tangential tooth force, and the distribution of the reaction forces depends on the geometry of the pinion. However, both Fig. 4.15 and Eq. (4.3) represent a symmetrically loaded pinion. Note that AB ¼ dBb is the projected area of each pinion bearing, where dB is the journal diameter and b is the axial length, and pB is the unit load: FBG ¼

FT ¼ AB pB ¼ dB bpB 2

(4.3)

Since the allowable bearing reaction forces are limited by the permissible bearing load and the bearing area, there is a maximum tooth force that can be transmitted to the pinion. Together with the pitch line radius and the shaft speed, which yields the pitch line velocity, upl ¼ rplω, the maximum power that can be transmitted to a pinion of a given geometry can be calculated as follows: P ¼ Mω ¼ FT upl ¼ FT 2πrpl n

(4.4)

Due to the fact that journal bearing velocity is also limited, it follows that the allowable tangential tooth force, and thus the allowable power for low-speed pinions (i.e., larger bearings), is larger than for high-speed pinions. This relation, shown in Fig. 4.16, may limit the ability of an IGC to maintain optimum flow coefficients across all pinions:

Integrally Geared Compressors Chapter

4

153

Large gear

Pinion power →

Increased pitch line velocity

Small gear

Pinion rotational speed →

FIG. 4.16 Effect of pinion speed and gear size on pinion power.

P¼

 u 2 b B pB 2upl πn dB

(4.5)

In the lower part of Fig. 4.16, it is shown why thrust forces originating from different pressures acting on the impellers also have an impact on the bearing reactions, FBT. A resulting gas thrust, Fgas, has to be counteracted by a thrust collar force, FTC. Since the contact of the thrust collar with the corresponding surface on the bull gear is not in the pinion shaft centerline, the thrust collar force FTC creates a moment which tries to turn the pinion shaft around an axis perpendicular to the axis of rotation. Additional bearing reaction forces, FBT, inhibit this movement. Because the bearing reactions to weight force, gear tooth force, and thrust collar moment are not necessarily in the same plane, the resulting bearing reactions for any pinion may point in different directions, even upward if the tooth force is large in comparison to the weight of the pinion. As the tooth force and the gas thrust are dependent on the operating point of the compressor, so are the bearing reaction force magnitudes and directions. This changes the bearing stiffness and damping so that the rotordynamic behavior of any pinion of an IGC has to be verified for different loads.

Lubrication In addition to lubricating the journal and axial bearings like in all other compressors, an IGC also requires lubrication and cooling for the gear meshes and thrust collars. Since the bearing load in an IGC is much higher than in inline

154 SECTION

II Types of Equipment

compressors, bearing lubrication is designed for this higher load. Usually, bearings with directed lubrication are used in contrast to bearings with flooded lubrication. This means that special consideration is taken with regard to high oil throughput and minimal oil retention time in the bearing. The gears and thrust collars are lubricated and cooled with spray oil. Spray bars with oil nozzles are located at the different gear mesh locations. It has to be ensured that the oil nozzles are chosen in way as to provide enough oil to the mesh without spraying so hard that erosion can occur.

Seals Theoretically, any type of shaft seal can be used in IGCs. Labyrinths seals, floating carbon rings, and dry gas seals are most common, depending on the pressure and the process gas, and brush seals can be used for high-temperature applications. Labyrinth seals are used for many low-pressure air compressors, especially when robustness is more important than low leakage. They are normally made of metal, mainly aluminum, but thermoplastics can be used for higher corrosion resistance or better rubbing behavior. With special labyrinth teeth design, smaller clearances can be achieved for improved leakage performance. Carbon ring seals are a good compromise between low leakage and simplicity for nonhazardous gases like air, nitrogen, or carbon dioxide. They allow smaller clearances than labyrinths because the floating carbon rings can follow the shaft motion when traversing critical speeds. Due to the longer axial sealing gap, fewer carbon rings are needed than labyrinth teeth for the same pressure difference. The lowest leakage can be achieved with dry gas seals, which are the most popular seals for explosive and toxic gases. They require the most complex seal gas system to maintain operation free of particles, liquids, and back pressure. The complexity of the seal gas systems increases with higher numbers of compressor stages and can turn unfavorable compared to single-shaft compressors, which only have two dry gas seals per casing regardless of the number of stages. Dry gas seals also add mass to the pinion outboard of the bearings, which lowers the natural frequencies and must be considered in the rotordynamics analysis. Besides the shaft seals, internal sealing between discharge and suction side of the impeller is needed in every compressor stage. For closed impellers, normally labyrinths are used as cover disc or eye seals, while brush seals are less common. Open impellers have no specific seal, but a stator contour matching the blade geometry maintains a small gap, which is limited by the amount of radial and axial movement of the rotor relative to the stator.

Rotordynamics The rotordynamic assessment of IG turbomachinery includes the same considerations as inline machines: undamped critical speeds, damped eigenvalue

Integrally Geared Compressors Chapter

4

155

analyses, lateral unbalance response, and torsional rotordynamics. However, as previously discussed, IGC rotordynamic analysis must consider additional influences based upon the loading from the gear reaction forces and the coupling of the pinion and bull gear dynamics. API standards allow the gearing system to be assessed as uncoupled components, but as the industry continues to evolve toward more aggressive performance (higher speeds and higher efficiency), the need to consider coupled dynamics is becoming evident. The gear mesh adds significant stiffness near the center of the pinion rotor and can affect the dynamic characteristics. Gear mesh stiffness is a complex phenomenon, but API 684 [11] Section 2.8.3 provides equations to estimate the stiffness coefficients in meshing gears as follows: K ¼ CðFW Þ cos 2 ðβÞ106

(4.6)

Kxx ¼ K cos 2 ðγ Þ

(4.7)

Kyy ¼ K sin ðγ Þ

(4.8)

Kxy ¼ Kyx ¼ K sin ðγ Þ cos ðγ Þ

(4.9)

γ ¼ Aα + B

(4.10)

2

In the above equations, C ¼ 12,057 N/m is a constant, FW is the net face width of the gear, and β is the helix angle. Also, in Eq. (4.10), α is the normal pressure angle, A ¼ 1 for downloaded rotors or A ¼  1 for uploaded rotors, and B ¼ 90 degrees for clockwise rotation or B ¼ 270 degrees for counter-clockwise rotation (looking into coupling end). The following equations can also be considered. Eq. (4.11) represents the localized deformation for gear tooth contact, which is based on the Hertzian contact and accounts for gear tooth angle (ψ), loading conditions, and material elasticity (E1 and E2). Dividing the normalized load (W/F) by the localized deformation results in an estimate of gear mesh stiffness as provided by Eq. (4.12):   W 1 1 9:000 + F E1 E 2 (4.11) d¼ cos 2 ðψ Þ 2

K∗ ¼

W=F ¼ d

cos 2 ðψ Þ   1 1 + 9:000 E 1 E2

(4.12)

AGMA 2001 [12] suggests mesh stiffness constant values anywhere from 10 to 20 N/mm/μm. Dudley [13] states stiffness constant for mesh deflection are not known with certainty, but stiffness values of 20–25 N/mm/μm are reasonable based on testing for 20 degrees pressure angle gears with low helix angles.

156 SECTION

II Types of Equipment

3.5 – K ayay , a–y

Normalized moment-angle stiffness

3

– K ayay , a–x

2.5 2 – K axay , a–x

1.5 1

– K axay , a–y

– K axax

0.5 0 –0.5 –1

–0.2

– K ayax –0.1

0

0.1

0.2

Normalized TC static misalignment angle a–x, a–y

FIG. 4.17 Influence of alignment on the moment angle stiffness that couples bull gear and pinion dynamics. (From T. Cable, L. San Andres, K. Wygant, On the predicted effect of angular misalignment on the performance of oil lubricated thrust collars in integrally geared compressors, ASME J. Eng. Gas Turbines Power 139(4) (2016) 042503.)

Thrust collars also provide a mechanism that couples the bull gear and pinion dynamics. The most obvious coupling is in the axial thrust within the fluid film of the thrust collar. However, the angled thrust collar also provides a moment coefficient that couples the lateral motion of the bull gear and pinion. The moment dynamic coefficients are generated by the hydrodynamic oil film in reaction to perturbations in the alignment conditions. Dynamic moment coefficients are produced that couple the lateral motion of the bull gear and the pinions. Fig. 4.17 shows how variations in the alignment may affect the moment angle stiffness. Typically, the inertia of the bull gear is order-of-magnitude greater than that of the pinion, meaning that bull gear motion is more apt to be imparted on the pinion than the pinion onto the bull gear.

Casings As opposed to inline compressors, IGCs require separate casings for each impeller stage. Most prominently, the discharge of a radial compressor requires a volume around the perimeter of the impeller to collect the discharge flow and channel it to the discharge piping. This volume is commonly a scroll or volute, which has improved flow characteristics compared to constant-area collectors. Fig. 4.18 shows an image of an IGC with four pinions and with volute casings for each stage. Note that the lower stages have cast volutes, while the last two stages, which would be at significantly higher pressure, have bulkier fabricated casings.

Integrally Geared Compressors Chapter

4

157

FIG. 4.18 IGC with cast and fabricated casings. (Courtesy Mitsubishi Heavy Industries.)

In addition to defining the aerodynamic flow path, the casings are also the structural element that contains pressure and controls operating gaps in closeclearance areas—for example, eye seals (covered impellers) or shroud-to-blade tip clearance (open impellers). This is a distinct difference with barrel-type inline compressors, for example, which are able to separate the functions of the flow path components (bundle diaphragms) and the pressure-containing component (main barrel casing). So, in IGCs, minimizing deflection due to internal pressure and external loads is an important aspect to the case design. A critical area in a volute is the tongue—where the small and large areas of the flow path intersect. For aerodynamic reasons, this region has a relatively sharp corner, which creates a large stress concentration under pressure loads (Fig. 4.19). Usually, there is a trade-off that needs to be made between aerodynamic performance of the volute and allowable stresses related to the radius in the tongue region.

FIG. 4.19 Finite element analysis (FEA) simulation of volute demonstrating stress concentration in tongue region.

158 SECTION

II Types of Equipment

FIG. 4.20 Horizontally split bull gear casing and volute mounting flanges. (Courtesy Voith.)

External loads from pipe connections and even gravity are particularly important to consider with IGCs since the casings are overhung from the main bull gear casing, which can develop significant bending moment loads. Furthermore, it is common that the volute casings are only attached to the bull gear casing at the lower half (180 degrees sector), which reduces the allowed external loads. This is done so that the bull gear casing can be split and the pinion bearings, seals, etc., can be accessed without having to remove the impellers, volutes, and attached piping. Fig. 4.20 shows an image of a horizontally split bull gear casing and the volute mounting flanges only on the lower halves of the pinions.

Mechanical Design The detailed mechanical design of an IGC is optimized for each particular application. In all applications, the gearbox serves as a housing for the bull gear and pinions as well as a mounting for the individual compressor or expander casings. In addition, the gearbox is often designed to integrate with the intercooler casings and/or lube oil system. The intercoolers are often positioned directly below the gearbox to minimize piping, and on small machines, the cooler housing may be integrated with the gearbox casting. In some applications, the gearbox is packaged directly above the lube oil tank. Fig. 4.21 shows a typical small IGC with the intercoolers packaged directly below the gearbox. This configuration allows for compact packaging and minimizes interstage piping lengths. Fig. 4.21 shows the stage 1 volute discharging straight down into the cooler on the right-hand side of the gearbox, then returning on the left vertically out of the cooler before turning 90 degrees into the stage 2 inlet. In cases where the gearbox is elevated, a structural analysis of the mounting system should be conducted to ensure that it has adequate strength and stiffness that would not adversely affect the rotordynamics of the system.

Integrally Geared Compressors Chapter

4

159

FIG. 4.21 IGC package with coolers directly below the gearbox. (Courtesy Hanwha Power Systems Americas Inc.)

For small- to moderate-size machines, the gearboxes are usually cast, or they may be fabricated and welded from steel plate as the size increases and casting is not practical. For large machines, the gearbox typically includes inspection ports above the pinion gears. Inside the gearbox at the pinion-to-bull gear mesh location, oil spray nozzles are installed to cool and lubricate the gears. Oil is routed to these nozzles and to the bearings although supply ports machined into the casing. Used oil drains from the gear meshes and bearings to the bottom of the gear case where it discharges into the lube oil tank. Oil pumps are selected based on the required oil flow to the gears and bearings. Most IGCs use a combination of a mechanically driven pumps that run off the main bull gear for primary oil supply and electrically driven pumps. A geardriven oil pump ensures that oil is being supplied anytime the machine is running. An electric auxiliary pump is required for start-up, shutdown, and emergency protection. The lube oil tank is sized to allow adequate retention time to eliminate air bubbles from the oil supplied to the core compressor. To avoid leaks, the gearbox pressure is maintained slightly below ambient. A key consideration in the mechanical design of an IGC is maintaining reasonable tolerances in the flowpath and between the rotating and stationary components. The key tolerances that must be maintained are: l l l l

eye seal/or tip clearance, diffuser width, impeller axial position, and impeller radial position.

To achieve the target aerodynamic performance, the clearance between the rotating impeller and the stationary casing must be maintained at design values. The axial position of the impeller relative to the casing is controlled by the stack-up between the gearbox, casing, and shroud. The clearance can be set during assembly using shims between the gearbox and casing. The radial clearance

160 SECTION

II Types of Equipment

FIG. 4.22 Shim and tolerance locations in an IGC.

of the impeller should be assembled with uniform circumferential clearance while running. See Fig. 4.22 for examples of typical shim and tolerance locations.

Operation The operation of IGCs varies in a wide range, affecting the start-up time, the number of starts and stops, and the turndown requirements. Several years of uninterrupted operation or multiple starts and stops per day determine different challenges to mechanics, instruments and controls, system components, and train concepts. On the other hand, a chosen compressor configuration leads to limitations during operation. Long-term uninterrupted operation for several years requires certain components, like air and oil filters and seals, to be optimized for low wear. In this case, large single machines are preferred, with electric motor as well as with steam turbine drive. A high number of overall starts demands additional fatigue calculations for many components, like impellers, gears, and pressure vessels. The requirements for frequent starts per day further include calculations of thermal deformation to avoid imbalance of the rotors. Due to long start-up time and the demand for turning gear (to prevent thermal bow), steam turbines are more complex for this application. In the case of electric motor drives, thermal overload due to high currents at start-up must be considered. Short start-up time

Integrally Geared Compressors Chapter

4

161

leads to more complex instrumentation and control with fast actuators for valves and guide vanes and preset values for the start-up procedure. One the most challenging application is fuel gas compression for gas turbine power plants, where fast start-up and load sharing between multiple units has to be realized. Steam turbines can drive big compressors with high-power demand and are well suited for long uninterrupted operation. On the other hand, they need much longer time for start-up due to their warm-up procedure at different speed steps. They also need turning devices and are often limited in the maximum number of starts per day. Electric motors are available in a wide power range and can be used for long uninterrupted operation as well as for frequent starts and stops. In the latter case, the thermal load due to frequent start-up current has to be considered. Depending on the required availability, multiple trains and hot standby machines are used.

IGC Applications IG turbomachinery are an important part of the oil and gas industry. They are present in many different aspects of oil and gas plants, including air separation units (ASUs), midstream petro-chemical processes, and power generation. In recent years, IG machinery have become commonplace in midstream processes.

Air Separation ASU applications have relied on IG machines since the early 1970s. Power ranges on individual units in ASU plants can be above 30 MW. The air separation industry is one of the most important customers for IGC manufacturers since they can easily benefit from most of IGC advantages—like multispeed aero layouts, ease of installing intercoolers between each stage, and relative simplicity to create flexible and efficient compressor designs out of standardized and proven components. This helps to improve both capital expenditures (development, etc.) and operating expenditures (power consumption, maintenance, etc.).

Midstream Liquefied natural gas (LNG) facilities are quite complex with unique demands for the turbomachinery. For methane, BOG compressors are used to recover gas that evaporates while offloading LNG from a carrier vessel to storage at the facility terminal. Once the gas is reinjected back into the system, secondary pumps raise the pressure of the fluid and transfer it to meet the demands of the final application: power plant, city gas, etc. It is also possible to boost the pressure after the BOG compressor with a higher-pressure compressor to meet delivery needs. This is more commonly done with reciprocating compressors, but advanced IGC designs can have the ability to meet the needs depending upon flow and pressure requirements.

162 SECTION

II Types of Equipment

FIG. 4.23 Compact LNG compressor with bulkhead located between motor and compressor.

In a similar manner to onshore LNG applications compressing BOG, the compact nature of IGCs can be adapted for use on LNG carrier vessels. Fig. 4.23 shows a sample configuration of one such unit. The wide range of operating conditions associated with these applications can be realized relatively easily by an overhung IGC compressor stage since its diffuser area is accessible to actuate VDVs from the outside. In this way, a turn-down range of nearly 75% at constant discharge pressure can be achieved. Frequently, the explosion proof and nonexplosion proof sides of the machinery are integrated through the bulkhead of the ship with an integrated bulkhead seal (see Fig. 4.23).

Petrochemical The oil and gas industry requires strict compliance with API 617 [4], and there is limited application for IGCs directly in hydrocarbon processing. IGCs may be used as refrigeration compressors, heat pumps, carbon-monoxide compressors, natural gas compressors, or as combined feed and recycle compressors for methanol plants. If low-pressure carbon dioxide is available from petrochemical plants, IGCs can be used for enhanced oil recovery where high-pressure carbon dioxide is injected in oil wells at pressures above 13 MPa. For smaller flow rate machines, offshore applications are well suited to the compact nature of IGCs. Locating an IGC on an offshore platform has some intrinsic benefits, such as the small footprint size, high efficiency, and low levels of dynamic forces. For an offshore platform, there are strict safety requirements which will translate to unique requirements for each application and a demand for maximizing API 617 [4] compliance with minimal exceptions. Two typical applications gaining prevalence for offshore platform applications are instrument air compressors and vapor recovery compressors.

Integrally Geared Compressors Chapter

4

163

Power Generation Fuel gas compressors are often employed in gas turbine-driven power generation applications. In such plants, low-pressure gas delivered from the pipeline must be pressurized to match the gas turbine requirements. Often, a multistage IGC is used, and dry gas seals are required to minimize the amount of flammable process fluid escaping to the atmosphere. These applications frequently have requirements for explosion-proof motors. Ongoing research is focused on the development of high-pressure/hightemperature IGC expanders for supercritical carbon dioxide power generation systems. These applications have the advantage of extremely high efficiency cycles, but place extreme challenges on sealing technologies.

Basic Industries IGCs are used for different processes in the basic industries. For ammonia plants, IGCs are employed as process air compressors and natural compressors. In urea processes, IGCs are used as high-pressure carbon dioxide compressors at pressures up to 20 MPa. In nitric acid plants, IG turbomachinery can be used as process air compressors, nitrous gas compressors, and waste gas expanders. All of these duties can be combined on a single gearbox, which is referred to as a compander. The largest IGCs to date (driver power up to 60 MW) are used as process air compressors for terephthalic acid plants. These trains consist of a steam turbine driver, the IGC, a motor/generator, and a waste gas expander. The IGC doubles as a reduction gear between the steam turbine and the motor/generator, and the waste gas expander can be a multistage IG expander as well.

Summary An IGC is a compressor architecture that incorporates a gearbox into the main body of the compressor unit. This chapter discussed the differences between IGCs and other compressor architectures. Advantages of IGCs include the ability to operate different compression stages at respectively optimal design speeds, intercool between stages to reduce compressor work, increase peak performance and operating range with VIGVs and VDVs, and the fact that IGCs tend to be more compact than inline centrifugal or reciprocating compressors sized for a similar duty. In addition, IGCs offer a highly modular architecture and can handle multiple process streams in a single unit. Some of the main disadvantages of IGCs are the result of having multiple pinion shafts and separate inlets/exits for each stage. Specifically, each pinion requires its own bearings, and each stage requires its own shaft seal. This can add significant cost and can adversely affect the reliability of IGCs when compared to inline compressors, which only have two bearings and two shaft seals regardless of the number of stages.

164 SECTION

II Types of Equipment

In many respects, the design of an IGC has many similarities with other turbo-compressor designs. This chapter discussed several areas of the design process and highlighted some aspects unique to IGCs. For example, bearing design has to consider gear loads and multiple operating conditions, and rotordynamics analysis can be more complicated, especially if there significant coupling between the bull gear and pinions. Also, the casing design of an IGC is particularly unique and has different loading requirements (e.g., piping loads and overhung weight) and assembly tolerance mitigation schemes than found on other compressor types. Finally, the applications of IGCs were discussed. Beginning with plant air service since before the 1950s, IGC applications have expanded into the process industry and are used in air separation, midstream production with LNG, power generation, etc. Due to the advantages of IGCs, their applications continue to increase.

References [1] A. Almasi, Integrally geared centrifugal compressors, Process. Mag. (2017). April, online source, http://www.processingmagazine.com/integrally-geared-centrifugal-compressors. [2] U. Fingerhut, E. Rothstein, G. Sterz, in: Standardized integrally geared turbomachines—tailor made for the process industry, Proceedings of the Twentieth Turbomachinery Symposium, 1991. [3] K. Wygant, J. Bygrave, W. Bosen, R. Pelton, in: Tutorial on the application and design of integrally geared compressors, Proceedings of Asia Turbomachinery & Pump Symposium, February 22–25, Singapore, 2016. [4] API 617 n.d., “Axial and Centrifugal Compressors and Expander-Compressors for Petroleum, Chemical and Gas Industry Services,” American Petroleum Institute. [5] API 672 n.d., “Packaged, Integrally Geared Centrifugal Air Compressors for Petroleum, Chemical, and Gas Industry Services,” American Petroleum Institute. [6] ASME n.d. Boiler and Pressure Vessel Code, American Society of Mechanical Engineers. [7] AGMA 6011 n.d., “Specification for High Speed Helical Gear Units,” American Gear Manufacturers Association. [8] API 613 n.d., “Special Purpose Gear Units for Petroleum, Chemical, and Gas Industry Services,” American Petroleum Institute. [9] ISO 6336 n.d., “Calculation of Load Capacity of Spur and Helical Gears,” International Organization for Standardization. [10] S.A. San Andres, T.A. Cable, K. Wygant, A. Morton, On the predicted performance of oil lubricated thrust collars in integrally geared compressors, ASME J. Eng. Gas Turbines Power 137 (5) (2015) 052502. [11] API 684 n.d., “Paragraphs Rotodynamic Tutorial: Lateral Critical Speeds, Unbalance Response, Stability, Train Torsionals and Rotor Balancing,” American Petroleum Institute. [12] AGMA, Fundamental Rating Factors and Calculation Methods for Involute Spur and Helical Gear Teeth, American Gear Manufacturers Association, 2001. [13] D.W. Dudley, Handbook of Practical Gear Design, McGraw-Hill, New York, 1984.

Integrally Geared Compressors Chapter

4

165

Further Reading [14] T. Cable, L. San Andres, K. Wygant, On the predicted effect of angular misalignment on the performance of oil lubricated thrust collars in integrally geared compressors, ASME J. Eng. Gas Turbines Power 139 (4) (2016) 042503. [15] C. Wacker, K. Kisor, in: Integrally geared centrifugal compressors for high-pressure CO2, Proceedings of Carbon Management Technology Conference, Orlando, Florida, USA, Feb. 7-9, 2012. Paper No. CMTC 150951. [16] K. Wygant, C. Robertson, F. Li, in: Tutorial: rotor-bearing dynamics of integrally geared turbomachinery, Proceedings of Asia Turbomachinery & Pump Symposium, March 13–15, Singapore, 2018.

Chapter 5

Reciprocating Compressors Justin Hollingsworth*, Greg Phillippi†, Martin Hinchliff‡,a, Chris Kulhanek*, Aoron Rimpel* and Franzisko Maywald§ *

Southwest Research Institute, San Antonio, TX, United States, †Ariel Corporation, Mount Vernon, OH, United States, ‡Dresser-Rand, Painted Post, NY, United States, § Burckhardt Compression AG, Winterthur, Switzerland

Equipment Selection Service Types and Operating Conditions Reciprocating compressors have been widely used for over 200 years ever since gases needed to be compressed. In the last 50 years however their dominance as the compressor type of choice has been eroded as other compressor types have been developed. Reciprocating compressors can be used in almost any compression application, but other compressor types are generally preferred for certain applications. Centrifugal compressors tend to be preferred when the power is above 2 MW, the mole weight (MW) is greater than 10, and the discharge pressure below 100 MPa. Screw compressors are preferred when the power is in the range 10–500 kW and the discharge pressure is below 30 bar. Roots blowers are used for discharge pressures below 0.1 MPa gage. Power levels below 100 kW tend to be the province of diaphragms, vanes, etc.

Common Reciprocating Compressor Types Compressors Without a Crosshead In these designs, the piston is attached directly to the crankshaft using a connecting rod. The issue with these compressors is that the gas that leaks past the piston rings travels directly into the crankcase. As such they are only used in two applications, for air or nitrogen compressors where the small leakage to the atmosphere via the crankcase is not an issue, and secondly on relatively low-pressure application where the gas is compatible with the crankcase and the crankcase can be hermetically sealed. These can be used at pressures up to around 0.7 MPa. This compressor type is not generally considered suitable for flammable and hazardous gas compression and so will not be covered in this book. a. Retired. Compression Machinery for Oil and Gas. https://doi.org/10.1016/B978-0-12-814683-5.00005-5 © 2019 Elsevier Inc. All rights reserved.

167

168 SECTION

II Types of Equipment

Compressors With a Crosshead With this type of compressor the piston connects to a piston rod, to crosshead and then back to the crankshaft via a connecting rod. The crosshead coverts the rotary motion of the crankshaft to purely linear motion in the piston rod. This means that the piston can be double acting, compressing on both outbound and inbound stroke. Additional sealing of the gas at the piston rod packing is facilitated allowing gas compression to take place with zero leakage to the atmosphere. These compressors are widely used in the size range of 50 kW up to about 20 MW and pressures from atmospheric to 340 MPa. They are available in different configurations from 1 throw up to 10 throws. Single-throw compressors are usually small (600 rpm) and low speed (200–600 rpm). Another common arrangement is a single throw between crankshaft bearings. This is also called the variable crank arrangement. This highly flexible arrangement is used on many slow-speed compressors and allows odd no crank throws from 1 to 10 throws in a horizontal configuration and up to six throws with a vertical cylinder configuration, cylinders are usually opposed but not always, it even allows throws to be different stroke although this is unusual. Balance can be equal to the balance opposed geometry above, but as this configuration is exclusively used for slow-speed applications the inertia forces are lower and so balancing of opposing throws is less critical. Other Types of Reciprocating Compressors Other much less common types of reciprocating compressors have the throws oriented in “W,” “V,” “L,” and radial configurations. These have been used historically in both with and without crosshead variants, mostly in air and natural gas compression. Such configurations are very rare today. The vertical reciprocating compressor is available for process and natural gas applications in both lubricated and nonlubricated labyrinth piston to cylinder bore sealing design. The piston has no piston rings or wearbands and relies on a labyrinth to provide sealing between the head-end and crank-end compression chambers. Of course, the cylinder bore has to be oriented vertically for this to work. Vertical compressors are typically used on clean dry gas applications

Reciprocating Compressors Chapter

5

169

where there is little chance of liquids leaking from the cylinder into the frame. Similarly, labyrinth compressors are suited to clean dry gas at 12 or higher MW. They are poorly suited to hydrogen due to the leakage past the labyrinth causing poor compression efficiency. Vertical reciprocating designs use less floor space, however the cylinders are close together with poor maintenance access. Diaphragm compressors originated in the early 1900s. These hermetically sealed positive displacement machines consist of two systems: a gas compression system, and a hydraulic system. The gas compression system includes thin metal membranes, or diaphragms which are clamped between an oil distribution and gas cavity plate, and process gas inlet and discharge check valves. The hydraulic system involves a motor-driven crankshaft connected to a reciprocating piston. The piston pressurizes a hydraulic fluid which in turn causes the diaphragm group to sweep through a contoured cavity, thus moving the gas out of the compressor through the discharge check valve. The diaphragm group completely isolates the hydraulic fluid from the process gas. Diaphragm compressors utilize static seals at the outer circumference of the diaphragm group providing lubrication free, leak free, and contamination free gas compression. This method of compression makes diaphragm compressors ideal for processing hazardous and high-purity gases. In addition, they are suitable for compression ratios up to 15:1 per stage, whereas conventional reciprocating compressors are limited to approximately 3:1 per stage. Diaphragm compressors can also produce discharge pressures to 100 MPa and are capable of compressing corrosive gases with minimal modifications to the materials of construction. The diaphragms are generally manufactured of stainless steel material, although other high alloy materials such as Inconel and Monel can be used where process gas compatibility dictates their need. Typical operating parameters for diaphragm compressors range from laboratory scale (1.1 kW motor) to continuous operation processes in production scale plants (up to 190 kW motor). Diaphragm compressors used in production scale plants are typically manufactured in accordance with API 618 [1]).

Advantages and Disadvantages Compared to Other Compressors Reciprocating compressors are of high efficiency, they have piston rings which provide effective sealing of the gas plus the heat transfer is low so that the compression efficiency approximates the adiabatic compression cycle. An adiabatic compression efficiency of 85%–95% is typical of slow-speed compression and 80%–90% typical of high-speed compression (difference is due to heat transfer is higher in water cooled slow-speed cylinders, lower valve, and gas inertia losses), in addition, the mechanical efficiency due to the friction losses inside the cylinder and in the frame and running gear is typical 97%–98% for slowspeed designs and 95%–97% for high-speed types. Other positive displacement compressor types such as screw, single screw, and vane tend to have greater internal leakage and lower efficiency. All positive displacement compressors

170 SECTION

II Types of Equipment

have the advantage that they can handle a wide range of gas density (gas MW), and even varying gas density. The screw and vane compressors are limited in pressure to about 30 bar. Reciprocating compressors have high adiabatic efficiency because the compression take place within the cylinder and due to the valve arrangement the amount of compression automatically matches the overall compression. For some screw and vane compressors the amount of compression is fixed and so may be greater or less than actually required, this results in a loss. Some screw types have discharge valves which can be adjusted to match the actual compression required but the adjustment is usually manual. Centrifugal compressors are very simple and highly reliable with minimal maintenance requirements. The mechanical efficiency is very high, over 99%, but the internal losses due to gas recirculation between the blades and stages are relatively high with result that the adiabatic efficiency is lower than is typical for a reciprocating compressor. The power density is very high, that is, the size and weight of a centrifugal compressor is only a fraction of the size of a reciprocating compressor of the same power. Centrifugal compressors operate on the principle of speed-induced gas inertia forces generating pressure (velocity head), as this is a function of both gas velocity and gas density changes in gas density (gas MW) will have a major effect on compressor performance. Centrifugal compressors have a difficult time in handling low MW gases as well where their MW varies over time. It is for this reason that large hydrogen compressors are usually reciprocating compressors.

Speed Considerations Rotative Speed The rotative speed for reciprocating compressors used in the oil and gas industries is usually in the range of 200–1800 rpm. Below 600 rpm is considered slow speed and over 600 rpm is considered high speed. Historically speeds lower than 100 rpm have been used but in modern times (since 1980) speeds of 200– 300 rpm are used for hyper compressors (100–350 MPa discharge pressure), 300–600 rpm for slow speed used mainly in process applications and high speed in most other applications with discharge pressure below 40 MPa. One determination is the speed of the driver, a natural gas compressor is often driven by a gas engine and the speed of the engine will vary between 750 and 1800 rpm depending on the power. Average Linear Piston Speed 2  Stroke  Rotating Speed FPM, 12 2  Stroke  Rotating Speedm=s or 1000  60 Linear Piston Speed ¼

Reciprocating Compressors Chapter

Piston speed ¼ Stroke ¼ Rotating speed ¼

Feet per minute (fpm) Inches Revolutions per minute (rpm)

5

171

m/s mm rpm

For example, for a rotating speed of 400 rpm, a stroke of 304.8 mm (12 in.), the average piston speed is Piston Speed ¼

2  12  400 12

Piston Speed ¼ 800fpm

Piston Speed ¼

2  304:8  400 1000  60

Piston Speed ¼ 4:06m=s

At piston speeds greater than about 6.1 m/s then the inertia forces in the reciprocating parts become so high that they approach the rod load limits of the machine, this provides an upper limit on piston speed.

Fixed vs. Variable Speed It is generally lower in cost and simpler to provide efficient capacity control by varying the volumetric efficiency (VE) using devices such as variable timing of the inlet valves, inlet valve unloaders to inactivate a cylinder end or by using clearance pockets to vary the cylinder clearance (and thereby vary VE). The highest efficiency capacity control is by using clearance pockets followed by compressor speed variations. Speed variation is inherent (within limits) using an engine driver, but also by using a variable frequency drive (VFD) using an electric motor drive. Variable speed may involve additional driver cost (for motor applications) plus technical complications in the compressor valves, inlet and discharge bottle and piping pulsation and also in the torsional system (these issue will be discussed in later sections). Driver Options Reciprocating compressors used in the oil and gas industry today are driven by electric motors or natural gas fueled engines. Electric motors are either induction or synchronous. Induction motors are typically chosen for speeds of 400 rpm or higher, synchronous motors are often preferred for speeds less than 400 rpm and power greater than 2 MW. Motors are generally fixed speed but can incorporate a VFD to provide variable speed. Reciprocating compressors impose a cyclic torque and lateral vibration transmitted through the crankshaft to the motor shaft and lateral vibration transmitted to the motor through the foundation. This means that driving a reciprocating compressor needs to be considered severe duty and appropriate considerations are required.

Effect of Torque Pulsations At each revolution the reciprocating components are accelerated and decelerated as the pistons travels out to the outer end and back to the inboard end

172 SECTION

II Types of Equipment

of the stroke, in addition, the gas is compressed discharged and then on the first part of the reversed stroke is expanded. This implies a severe torque pulsation which on a two- and four-throw compressor will include a peak torque that will usually exceed 200% of the mean torque and a torque reversal for part of each revolution. A six throw is not as extreme but still has a severe pulsating torque. This torque pulsation is partially absorbed by the inertia of the flywheel (when present) and/or the motor inertia so as to reduce the torque variation that must be provided by the motor magnetic field.

Electric Motors Motor Sizing API 618 requires that the motor nameplate HP be 110% of the maximum operating condition BHP. In addition, it is good practice to ensure that the motor insulation is class F or better and the temperature rise is limited to class B rise at 1.0 service factor. This will ensure some thermal reserve and ensure long life of the insulation. It should be noted that unlike other driver types, motors will provide good efficiency over a wide load range, offering maximum efficiency over loads of 50%–90% of nameplate, the efficiency starts to drop off over 90% load. So there is no detriment to operating motors at less than nameplate load. Care needs to be taken to ensure electric motors are operating at no more than nameplate load, nameplate amps and within the insulation temperature limits otherwise operating life and reliability will be severely impacted. Induction Motors An induction motor develops torque by inducing current to the rotor, which is proportional to the differential speed of the rotor and the rotating magnetic field in the stator. For NEMA design B motors the differential speed (called slip) is between 1% and 2% at full load. Due to the torque variation at each revolution the instantaneous speed will vary. For example, if the speed variation was 0.8% so the slip would vary between 0.6% and 1.4% giving a torque variation of 60% torque to 140% torque. But the power factor would vary with load so at 60% power it might be 80% amps and at 140% load 140% amps for an average of 110% amps. Note how the power is at 100% but the amps are at 110%, that is, in an overload situation by amps but not by power. The heating effect in the motor is primarily governed by the motor amps so at 110% amps it is likely that the motor winding temperatures would be excessive. In this case the current pulsation is (140  80)/100 ¼ 60% NEMA MG1 limits the current pulsation to 66% API-618 limits the current pulsation to 40%. It is recommended that the API 618 limits be applied to induction motors because the NEMA pulsation limits are not adequate to protect against overloading the motor.

Reciprocating Compressors Chapter

5

173

Synchronous Motors See Fig. 5.1. In a synchronous motor AC power is supplied to the stator which generates a rotating magnetic field. DC power is supplied to the rotor which results in discrete N and S poles. The poles in the rotor then lock onto (synchronize) and follow the opposing rotating magnetic pole (N follows S). At zero load they follow exactly but at load they follow slightly behind by a load angle which varies between 0 electrical degrees at zero load approximately 32 degrees at 100% load and approximately 70 degrees at stall. There are 180 electrical degrees between each adjacent N and S pole. So take the previous example where the torque variation was 40% the torque would vary between 60% and 140% and the magnetic lag would vary between 0.6  32 ¼ 19.2 degrees and 1.4  32 ¼ 44.8 degrees. However, in a synchronous motor the exciting amps are varied to keep the power factor constant with load and so the amps would also vary between 60% and 140% nameplate, the average amps would be 100%, average power 100%. The current pulsation would be (140  60)/100 ¼ 80%. So in this case the NEMA limit of 66% current pulsation is adequate to protect the motor because a synchronous motor is less affected by torque pulsations. API 618 also recommends 66% as a current pulsation limit. Note that the rotor lags the stator magnetic field by an amount proportional to the torque. The magnetic field acts as a spring and the rotor inertia and drive inertia
will have a natural frequency that is equal to f n ¼ 35; 200=n  √ ðPr  f Þ=WK 2 [2] where fn is the undamped natural frequency gnetic field rotation Stator ma D

M

E ∞

∞ ∞

∞

⬘ otor FIG. 5.1 Rotating magnetic field characteristics in synchronous motors. (From The ABC’s of Synchronous Motors—the EM-WEG Synchronizer, Figure 29, http://ecatalog.weg.net/files/ wegnet/WEG-the-abcs-of-synchronous-motors-usaem200syn42-brochure-english.pdf.)

174 SECTION

II Types of Equipment

in cycles/min, n the synchronous speed in rpm, Pr the synchronizing torque coefficient, this is the stiffness of the electrical field applying torque to the rotor to restore it to the neutral position in units of kW/rad, and f the electrical line frequency in Hz, WK2 in units lbs-ft2 (see [2a]). The electrical natural frequency obviously must be separate from the running speed in order for the current pulsation to be below 66%. Usually it is between 0.5 and 0.75 running speed. Also referring to NEMA MG1 21.38 there is a term called the compressor factor C where C ¼ 0.746WK2n4/(Prf108) so the term compressor factor gives a dimensionless measure of the inertia in the motor and driven system. At C ¼ 9.24 the system is in electrical resonance and the natural frequency is proportional to 1/C1/2. Generally for the multiple throw machine such as a six throw, satisfactory current pulsation will be achieved at C ¼ 20 (fn ¼ 0.68 running speed) and for a two throw at C ¼ 30 or greater (fn ¼ 0.55 running speed). It is typically acceptable to have a motor with minimum C ¼ 20 for a two throw and C ¼ 15 for four or more throws (considering the bare motor inertia only). Any additional inertia required will be in the compressor or can be added to the compressor flywheel. The current pulsation for both induction and synchronous motors should always be checked during the engineering phase by the motor manufacturer using the full-load and part-load crank effort and the driven inertia (supplied by the packager). Most motor manufacturers will calculate the current pulsation for both induct and synchronous motors using a simple single degree of freedom model, electrical bus-magnetic field stiffness-inertia of motor and driven equipment lumped as one mass. This is accurate for most systems where the torsional natural frequency is above 4 running speed. However, for applications where the first torsional natural frequency is below 4 then a more accurate result can be obtained using a two degree of freedom model, electrical bus-magnetic field stiffness-motor inertia-equivalent shaft stiffness-flywheel and compressor stiffness lumped as one mass. Not all motor manufacturers can do this calculation, but it will show more accurate and lower current pulsation for torsionally soft systems. The motor manufacturer will determine the minimum amount of system inertia required to achieve 66% for synchronous and 40% for induction motors. The required inertia can be added using a compressor-mounted flywheel if the motor inertia is not adequate. Note for a synchronous motor it is recommended that the driven inertia always be less than the motor inertia. This is because of the strong (up to 40% of nameplate torque) 2 slip frequency pulsating torque that is induced during acceleration. During acceleration this 2 slip frequency torque will inevitably coincide and excite torsional resonance at some brief point during start-up assuming that torsional resonances occur between 0 and 120 Hz (for a 60 Hz electrical frequency). Reciprocating compressor drives are torsionally robust so if the driven inertia is less than the motor inertia then the resulting torsional stresses are unlikely to exceed permissible limits. See this chapter for a further discussion of torsional analysis.

Reciprocating Compressors Chapter

5

175

Advantage and Disadvantage Comparison: Induction vs. Synchronous Motors Synchronous Motors Typically, it is 1%–2% better efficiency than induction. It runs at a power factor of 1.0 or 0.8 leading so can improve the overall power factor of the plant resulting in reduced electrical demand charges. It is less impacted by the pulsating compressor torques. Starting current is lower in synchronous motor than in induction motor. Starting torque is lower in synchronous motor than in induction motor, 40/30/150 is typical but 60/60/175 is available. Care is required for proper compressor start-up unloading especially at the 95%–100% speed point where the motor pulls into synchronization. A synchronous motor is especially useful in slow-speed applications (under 400 rpm) when it is comparable to the cost of an induction motor and is normally used in a single-bearing configuration directly and rigidly bolted to the crankshaft. Induction Motors It is the lowest cost driver. It provides good starting torque, usually 60% at breakaway which continuously rises to a peak at about 95% speed. Higher torques are available but it is usually better to ensure adequate start-up unloading of the compressor. Starting currents are high for full voltage across the line starting and so a strong power supply is required. Power factor is quite low usually in the range of 0.5–0.8 lagging. Lower speed and highefficiency motors tend to have lower power factors. This can result in higher power costs due to demand charges unless separate power factor correction equipment is installed. Induction motors are quite sensitive to pulsating torques especially electrically stiff motors (low slip). The user is cautioned about the use of high-efficiency motors with low slip (1% or lower), as compared to a standard motor with 1.5% or greater slip. Due to the negative effect of pulsating torque on current pulsation and power factor the hoped for improvement in efficiency may not be achieved and much greater inertia is required in the compressor flywheel to achieve acceptable current pulsation [2b]. Motor Mechanical Effects Torque Pulsation The compressor imposes a strong pulsating torque on the motor shaft at 1 and all higher harmonics of the compressor speed, the strongest harmonic is a function of the number of throws and whether the compressor is a full load (all cylinders double acting) or part load (some or all cylinders single acting or unequal load head end to crank end). A two-throw compressor will have a very strong 2 harmonic. Four-throw strongest harmonic is the fourth. Six-throw strongest is usually the sixth. The GMRC torsional guideline [3] notes that for a two-throw compressor the motor should be designed for a torque of 100% of nameplate mean torque plus a pulsation torque of 250%, for a four throw 100% mean plus 200% pulsating, six-throw compressor 100% mean and 150% pulsating. The pulsating torques can be reduced by a compressor flywheel and a flywheel normally will be required on a two- and

176 SECTION

II Types of Equipment

four-throw compressor in order to meet the current pulsation requirement. The motor should be heavy duty (also called severe duty). For heavy duty applications the motor manufacturer will typically provide a large shaft, heavy duty fan, more internal bracing of windings, etc. Keyways are best avoided, but if necessary they should include a radius fillet to minimized torsional stresses [3]. It is best if the motor shaft at the drive end from the core through shaft end (including bearing) be no less than the crankshaft stub shaft [3]. Lateral Pulsation The compressor imposes a strong lateral pulsation on the motor. This comes from the rod loads of the compressor. They are transmitted to the motor in two ways: through the baseframe and foundation and through the crankshaft to the motor shaft. At each revolution the two throws adjacent to the drive apply the rod load to the crankshaft at the instantaneous angle of the connecting rod. As the crankshaft rotates it roughly orbits and because the two drive end throws are in phase but opposed, the crankshaft also angles. The motion of the crankshaft is limited by the bearing clearance which will typically not exceed 0.1% of the shaft diameter as a diametral clearance. The motion at the crankshaft is primarily at 1 but there is some excitation at all higher harmonics. For a two-bearing motor connected via a flexible coupling the crank lateral motion will cause vibration of the flywheel (0.2 mm p.p. for a 200-mm crankshaft assuming nonresonant motion), but only a limited amount is transmitted through the coupling to the motor. However, for rigidly connected motors, for example, a single-bearing synchronous motor, the entire lateral motion of the crankshaft is transmitted directly. The outboard bearing and motor shaft needs to be designed for the crankshaft orbital motion while limiting the vibration at the bearing to acceptable limits. Because of the compressor pulsating loads the vibration at the drive motor will exceed what is normally considered acceptable. For example, ISO 10816-3 list an acceptable vibration at the motor of 4.5 mm/s rms, however this standard specifically excludes motors driving reciprocating compressors. A more realistic limit would be the acceptable value listed for the compressor frame which is 8 mm/s rms as a typical vibration limit for the motor frame and the bearings.

Variable Frequency Drives A VFD allows an otherwise fixed-speed electric motor to run with variable speed. Variable speed allows almost infinite and very simple capacity control. Refer to Chapter 7 for more detail about VFDs. Engines The majority of reciprocating compressors used in the upstream and midstream industry segments today are driven by natural gas fueled engines. The gas being compressed is also used to fuel the engine. The compressor is matched to the engine power and speed connected by a flexible coupling, commonly referred

Reciprocating Compressors Chapter

5

177

to as a separable compressor. Separables were built starting around 1970. Separables are available to match gas engines (less commonly to diesel engines) and are available in the power range of 200 kW up to around 7000 kW. Modern separable gas engine are almost exclusively four-cycle designs. The primary market for engines used to drive a reciprocating compressor is for electric generators and they are therefore designed to run at electric synchronous speeds such as 720, 900, 1200, and 1800 rpm for 60 Hz frequency and 750, 1000, and 1500 rpm for 50 Hz. Natural gas fueled engines are available from 40 to 6000 kW. As engines get larger they run slower.

Integral Engine-Compressors An engine/compressor configuration that is not prevalent today is the integral engine-compressor. This is a design where the compressor cylinders are mounted on the engine frame and driven by the engine crankshaft. Integral engine-compressors power levels were typically 4500 kW or less, running at 360 rpm or less, and very few have been built since 2000. Many thousands of integral engine-compressors are currently operating. Integral enginecompressors were built both as two and four cycle. Turbines and Gearboxes It is possible and maybe advantageous in some cases to drive a reciprocating compressor with a gas or steam turbine. A gas turbine may allow the burning of natural gas with some quantity of hydrogen sulfide which cannot be accomplished in the current regulatory climate with a reciprocating engine. If a facility has excess steam (or heat from which steam may be produced), it can potentially be used to drive a steam turbine. The “fuel” in this case is essentially free or very low cost. With either kind of turbine drive a speed reducing gearbox will need to be utilized because the turbine will normally run at rotating speeds much higher than the driven reciprocating compressor. The addition of a gearbox to the drive train raises significant torsional vibration concerns and in many cases will require the use of a torsionally soft coupling between the gearbox and compressor. In common practice, gas and steam turbines are rarely used as drivers for reciprocating compressors. Steam turbines were fairly common in refinery applications prior to 1980 when the steam supply from within the refinery was available and highly reliable as a source of power. However since 1980 the steam turbine has been rarely used. Steam turbines operate at 3000– 4000 rpm and gas turbines at greater than 4000 rpm. So the turbine is connected to the compressor by a double reduction gearbox. The gearbox will usually have a double helical gear for smooth power transmission and the absence of thrust. A large flywheel and torsional soft coupling are generally used to couple the compressor and gearbox in these cases. The first torsional natural frequency is tuned to approximately 2/3 of the minimum full-load run speed so that the

178 SECTION

II Types of Equipment

dynamic torques at the gear mesh comply with the API 618 max limit of 25% P-P of the transmitted torque. This is necessary to avoid gear mesh chatter that will destroy the gear box if too severe.

Couplings Couplings are used to connect the driver to the compressor. The simplest form is when the crankshaft is direct bolted to the motor shaft, that is, without a coupling. This is the normal arrangement for bolting a slow-speed compressor to a single-bearing synchronous motor. Alignment is critical and is accomplished by aligning the outboard bearing so that the crank web deflections at the drive end throw are close to zero (typical limit is no greater than 0.025 mm runout using a dial indicator in the crank web). This is highly reliable, torsionally and laterally stiff and will accept a high torque pulsation including a torque reversal. All other coupling types will accept some degree of misalignment, lateral, and angular, they are selected basis the torsional analysis based on the torques applied, stiffness, and damping requirements. Flexible disc couplings are economical and torsional stiff, they will accept lateral and angular misalignment, they will accept large torsional pulsations but only limited torque reversal. A flywheel is often used on the crankshaft to absorb torsional pulsations to avoid overloading and to limit torque reversals on the coupling. Elastomeric couplings come in two basic varieties, rubber in compression and rubber in shear, the elastomer has a limited life and needs to be replaced at the original equipment manufacturer (OEM)’s recommended replacement interval. Also properties are temperature dependent and so the torsional analysis should consider maximum and minimum temperature properties they are usually limited in power to approximately 3 MW due to the heat dissipation limits of the elastomer. The rubber in compression is moderately stiff and has only limited misalignment capability. It consists of matching steel or nodular iron halves with male teeth on one side with female teeth on the other, between the teeth are synthetic or elastomer blocks. The blocks provide the torsional stiffness along with torsional damping. They are used when the torsional analyst determines a moderate torsional stiffness and some level of damping is required. Rubber in shear couplings is highly torsionally flexible, but the lateral misalignment is limited. They also provide good damping. They are typically used when it is desired to place the first torsional natural frequency below the running speed, this would normally be on variable speed applications. Steel couplings with coil springs in compression are used when a torsionally soft coupling is required usually to bring the first TNF (torsional natural frequency) below 1 . Stiffness quoted is quite accurate and is generally constant with load and temperature. Damping is possible using damper friction blocks (coulomb damping), however the damping capacity is quite limited. These are used in high kW applications, and below 3 MW elastomeric couplings are usually more economical.

Reciprocating Compressors Chapter

5

179

Steel couplings with leaf springs and oil circulation are available and widely used in medium to high HP applications, and they are available up to the maximum kiloWatt ratings of modern engines and compressors. The oil supply usually comes through the crankshaft either from the engine, gear box, or compressor. The stiffness is accurate and independent of temperature. The damping capacity is high, relatively constant, and predictable over the coupling life. These couplings are selected when damping is required, for example, for variable speed applications using a motor VFD, turbine gear drive, or when coupling a large engine to a compressor. The TNF can be set below 1  (when using a gearbox) or near a higher harmonic where the damping will effectively reduce the resonant stresses.

Engineering Scope Packaged Units vs. Block-Mounted Units Oil and gas industry reciprocating compressors can be packaged or block mounted. Upstream and midstream applications typically use a packaged unit while the downstream (refining and petrochemical) industry uses block mount but is moving toward packaged units for lower powers, say less than 3700 kW. A packaged unit has the compressor with its driver and coupling mounted on structural steel skid. For compressors up to 1500 kW the skid is the foundation. Higher powers may require the skid be attached to a concrete pad to provide additional foundation mass. The fundamental idea of a package is to have a portable compressor that can be moved from well to well, as an example. A package is completely self-contained unit. The only required connections are compressed gas suction (in) and discharge (out) and the package is ready to be used. A block-mounted compressor is typical for the refining industry and has the compressor system built at site rather than in a shop as in the case of a packaged unit. The compressor, driver, and all the auxiliaries are shipped to site where they are assembled into the gas compression system. The compressor is mounted on a large concrete block (hence the term “block mounted”) that is the compressor foundation. In either case the compressor, driver, coupling, and all of the auxiliary systems are engineered specifically for the gas compression application.

Importance of Pulsation Studies The fundamental idea of a pulsation study is to ensure the process gas piping system will not fail due to stresses caused by pulsation (pressure fluctuations). Failure can lead to a release of process gas. Therefore, a pulsation study is an extremely important piece of the engineering design of the overall compression system.

180 SECTION

II Types of Equipment

Frame and Skid Flexibility Designing a skid to be adequately stiff can be a challenge for higher power reciprocating compressors, above about 1500 kW. Up to 1500 kW it is relatively easy to make a skid stiff so that alignment of the compressor to the driver can be maintained during transportation. But even then, alignment is always checked as part of the commissioning activity and will often require adjustment. For powers above 1500 kW alignment of the compressor to the driver will always require adjustment during commissioning.

Testing Options All reciprocating compressors used in the oil and gas industry undergo a mechanical run test as outlined in API Standard 618. Each compressor manufacturer will perform the test a little differently but the fundamental API 618 requirement is for a 4 h no-load mechanical run test. “No-load” means that no gas is compressed as the suction and discharge valves are not installed in the cylinders during the test. The basic idea is to prove the machine is mechanically sound and the running gear and cylinder lubrication systems are working properly. More can be learned about testing in Chapter 12. Some users ask that an inspection be done after the test that requires some dismantling of the compressor. Many manufacturers balk at this request because they hesitate to dismantle a compressor that has passed a successful run test and reintroduce an element of human error during reassembly. It is not common that the manufacturer or packager performs a performance test either in the shop or in the field. Shop performance tests for reciprocating compressors (commonly referred to as “string tests” with centrifugal machinery) can be very expensive to setup and perform and are extremely rare. However, the performance of reciprocating machinery is usually confirmed in-situ during commissioning of the equipment in the field.

Vibration Concerns Reciprocating compressors impose heavy dynamic forces on the foundation. These forces are of two types: Global forces apply due to net unbalance inertia forces from the rotating and reciprocating parts plus unbalance pulsation forces in the pulsation bottles and in the cylinder nozzles. These are typically the only forces reported and act on the entire foundation. Local forces are internally balanced within the compressor and so local forces on foundation result from the differential elastic stretch of the compressor. These include the gas force acting on the piston and head, the inertia forces of the rotating and reciprocating parts acting on the frame and vertically at the

Reciprocating Compressors Chapter

5

181

crosshead, gas forces at the cylinder nozzle due to pressure pulsations, gas forces in the pulsation bottle. Unbalance inertia forces are the vector addition of the acceleration forces of the rotating and reciprocating mass of all throws. The rotating mass are the unbalance portion of each crank throw plus the rotating portion of the connecting rod. The reciprocating mass is the reciprocating portion of the connecting rod, crosshead, piston, and rod. For vertical force F ¼ mrotrw2 sin(ωt) where mrot is the rotating mass, r is crank radius, w is angular velocity in rad/s, ωt is the crank angle (0 ¼ outer dead center). For horizontal force (axis of piston rod) F ¼ (mrot + mrec)rω2cos(ωt) + mrω2r/Lcos(2ωt) where L is the connecting rod length. For the reciprocating mass there are a 1  and a 2  component. Ratio r/L is the crank radius/connecting rod length and is typically about 20%. However the acceleration forces of each throw act at each throw and so are local forces. So for example, consider a six-throw compressor with all cylinders of equal mass. At each throw there will be a local force due to the gas force acting on the piston this will apply as an equal and opposite force at the cylinder head and at the crankshaft. So while the force is in balance it does cause elastic stretch of the frame, distance piece, and cylinder. The elastic stretch will be approximately 0.1 mm p-p per meter of distance from the center of the frame to the center of the cylinder. In addition, the inertia force acting on the frame and crosshead must be accounted for. Over the entire machine the inertia forces will sum to zero for a fully balanced six throw; however, the local forces will cause the frame to distort elastically and shake by an amount that does not exceed the same value of 0.1 mm p-p per meter along the frame. The forces, both local and global will cause cyclic elastic deformation of the compressor, bottles and piping, plus the skid, and foundation. Normal practice is that for small skid-mounted compressor and for foundation-mounted compressors just the global forces, that is, the unbalance forces and moments are reported and considered in the design of the compressor base and foundation (sometimes called the “rigid frame assumption”). However, for certain large skid-mounted compressors it may be appropriate to consider the more complete picture of the local gas forces and shaking forces acting at each throw. This is sometimes called the “flexible frame analysis” and is a much more complicated analysis comprising an finite element analysis (FEA) model of the frame and cylinder train along with the support structure. This allows the skid to be analyzed including the complex elastic distortion of the frame under load and how those loads are transmitted into the skid and foundation. Allowable limits for compressor vibrations are contained in ISO 10816. This provides good, acceptable, and marginal levels of vibration in displacement, velocity, and acceleration units, rms. Velocity units tend to be the most useful

182 SECTION

II Types of Equipment

as the same numerical values are reasonably valid over all machine speeds and vibration frequencies.

Part

Acceptable velocity limit mm/s rms

Marginal velocity limit mm/s rms

3.0 8.0 13.0 16.0 19.0 19.0

4.5 12.0 19.5 24 28.5 28.5

Foundation Frame top Cylinder lateral Cylinder rod Damper Piping

The foundation would be considered to be the top of the foundation at the foundation compressor base interface. Frame top should be considered the solid part of the frame which is directly linked back to the compressor bearings and direction is the piston rod direction. Cylinder lateral is the side of the cylinder measuring in the axis of the crankshaft. Cylinder rod is the end of the cylinder measured in the piston rod direction. Dampers are the pulsation bottles in any direction. Piping means gas piping. Acceptable and lower levels are considered good values for normal running measurements. Above acceptable to marginal levels are considered high and some evaluation needs to be made as to whether these are acceptable or if remediation is necessary. Above marginal values are considered unacceptable and shutdown or remediation is necessary.

Key Guidance From Standards Oil and gas industry reciprocating compressor standards and some comments about each: l

l

API Standard 618 “Reciprocating Compressors for Petroleum, Chemical, and Gas Industry Services,” Fifth Edition published December 2007. A Sixth Edition is in development at this time. Edition history: First Edition: April 1964 (34 pages) Second Edition: July 1974 (39 pages) Third Edition: February 1986 (109 pages) Fourth Edition: June 1995 (165 pages) Fifth Edition: December 2007 (190 pages) Sixth Edition: Pending Publication API Specification 11P “Specification for Packaged Reciprocating Compressors for Oil and Gas Production Services,” Second Edition published November 1989. The First Edition was published February 1975. This Second Edition is no longer available for purchase from API and not supported by API.

Reciprocating Compressors Chapter

l

l

l

l

5

183

ISO 13707:2000 “Petroleum and natural gas industries—Reciprocating compressors,” published in 2000. It is based on API 618 Fourth Edition published June 1995. API Standard 618 Sixth Edition will not be published as an ANSI/ISO standard as was the Fifth Edition. I am not aware of what plans ISO may have for a new edition. ISO 13631:2002 “Petroleum and natural gas industries—Packaged reciprocating gas compressors,” published in 2002. It is based on API Standard 11P Second Edition published November 1989. API Standard 618 Sixth Edition will not be published as an ANSI/ISO standard as was the Fifth Edition. I am not aware of what plans ISO may have for a new edition. Gas Machinery Research Council (GMRC), division of the Southern Gas Association “Guideline for High-Speed Reciprocating Compressor Packages for Natural Gas Transmission & Storage Applications,” published 2016. This is not a standard or specification as GMRC does not have any authority to publish a nationally recognized standard. This guideline is intended to provide guidance in packaging oil and gas industry reciprocating compressors 2000 hp and larger. GMRC “High-Speed Compressor Package Guideline for Field Gas Applications,” published 2017. This is not a standard or specification as GMRC does not have any authority to publish a nationally recognized standard. This guideline is intended to provide guidance in packaging oil and gas industry reciprocating compressors 2000 hp and smaller.

Performance Cylinder Swept Volume Process Overview A reciprocating compressor is a positive displacement machine in that a volume of gas is drawn into a compressor cylinder’s compression chamber where it is trapped, compressed, and pushed out. The P-V diagram is a plot of the pressure of the gas vs. the volume of the gas trapped in the compression chamber. In Fig. 5.2, PS, suction pressure, represents the pressure of the gas at the inlet to the compressor cylinder. PD, discharge pressure, represents the pressure of the gas at the outlet from the compressor cylinder. V1 represents the maximum volume of gas trapped in the compression chamber and V3 the minimum. The difference between V1 and V3 is known as piston displacement, or how much volume is displaced in one stroke length of the piston: Piston Displacement ¼ V1  V3 Referring to Fig. 5.2, the P-V diagram is made up of four basic segments or events, 1-2, 2-2A-3, 3-4, and 4-4A-1, and each is explained. Position 1 has the maximum volume (V1) of gas trapped in the compression chamber and that gas is at suction pressure and temperature. At Position 1 all of the compressor valves are closed and the piston is at rest. All of the compressor valves are simple

184 SECTION

II Types of Equipment

2A 3

PD

Pressure

2

4

PS

1

4A

0

V3

V4

Volume

V1

FIG. 5.2 Example pressure-volume diagram. (Courtesy of Ariel Corporation.)

spring-loaded check valves and are not actuated by any outside means. As the crankshaft rotates the piston will move and the volume inside the compression chamber decreases (moving from Position 1 to 2). From Position 1, as the volume decreases, the gas pressure increases (the gas is trapped in the compression chamber with the volume decreasing). When the pressure inside the compression chamber becomes slightly higher than discharge pressure (PD, at Position 2), the discharge compressor valve will open. This completes the first segment of the diagram and is called the compression event. At Position 2, the discharge valve opens and, as the piston continues to move, gas at discharge pressure and temperature is pushed out of the compression chamber through the discharge compressor valve and into the discharge gas passage of the cylinder. At Position 3, the piston comes to rest, the discharge compressor valve closes and this segment ends. Segment 2-2A-3 is called the discharge event. Moving from Position 1 to 3 the piston has moved through one stroke length and one-half revolution of the crankshaft. Position 3 represents the minimum volume of gas trapped inside the compression chamber (V3). Note this is not zero volume. At Position 3 the piston will reverse and travel in the opposite direction. As it moves the volume inside the compression chamber increases and the trapped gas (V3) will increase in volume and decrease in pressure—the gas will expand. Segment 3-4 is called the expansion event. At Position 4, the pressure inside the compression chamber will be slightly less than suction pressure (PS) causing the suction compressor valve to open. With the suction valve open, the compression chamber is open to the suction gas passage and as the piston continues to move, the volume continues to increase and the compression chamber fills with gas at suction pressure and

Reciprocating Compressors Chapter

5

185

temperature. The piston returns to Position 1, comes to rest, and the process repeats. Segment 4-4A-1 is called the suction event. This completes the basic reciprocating compression process. One cycle around the P-V diagram represents one revolution of the crankshaft and two stroke lengths of the piston, one from Position 1 to 3 and another from Position 3 to 1. Four events make up the P-V diagram—compression, discharge, expansion, and suction. The basic governing equation for the entire process is the first law of thermodynamics. Typically, the operation of the piston can be considered slow enough that the cylinder thermodynamic properties (i.e., pressures, temperatures, etc.) are treated as quasi steady and without any spatial variation. Neglecting kinetic and potential energy effects, the first law of thermodynamics can be written for the control volume encompassing the cylinder as shown in Eq. (5.1). Here, Q_ is the heat transfer rate from the environment to the fluid; W_ ¼ P V is the rate of change of work done by the fluid on the boundary; m_ s hs and m_ d hd are the rates of change of flow energy entering the suction and leaving the discharge valves, respectively; m_ l hl is the rate of change of flow energy lost by leakage _ + mu_ is the rate of change of through the piston seal; and E_ ¼ dðmuÞ=dt ¼ mu internal energy within the control volume. Also, recognize that mass continuity is imposed on the flow rates by m_ ¼ m_ s  m_ d  m_ l . Discussion of the mechanisms for mass exchange through the valves is presented in a later section. 

Q_  W_ + m_ s hs  m_ d hd  m_ l hl ¼ E_

(5.1)

Valves Types Compressor valves are simple spring-loaded check valves. They are not mechanically actuated in any way and open and close due to differential pressure. The fundamental purpose of the spring is to provide a “push” to help the seal element move from the open to the closed position. There are four basic designs of compressor valves, delineated by the form of the seal element: – – – –

ported plate; individual ring; poppet; and reed.

Every compressor valve consists of a seal element, seat, guard, spring, and a method to hold the seat and guard together as the valve is installed and removed from the cylinder The ported plate design, as shown in Fig. 5.3, is the most common. The seal element is a disc (a “plate”) with slots that allow gas to pass through the disc.

186 SECTION

II Types of Equipment

FIG. 5.3 Typical ported plate valve components. (Courtesy of Ariel Corporation.)

FIG. 5.4 Typical ring valve components. (Courtesy of Ariel Corporation.)

The individual ring design is as it sounds, a series of concentric individual rings that form the seal element (Fig. 5.4). The poppet design has each seal element being an individual “poppet” which can look like a bullet or a mushroom. Each poppet has its own spring and acts individually as there is no mechanism to force the poppets to open and close together (Fig. 5.5). The reed design has the seal element opening and closing parallel to the gas flow through the valve as opposed to perpendicular to the flow as in the previous three types (Fig. 5.6).

Dynamics Fig. 5.7 represents a theoretical, one degree of freedom model of a valve. The valve is normally held closed by a spring force and opens when the pressure force on the valve plate (from differential pressure between up and downstream volumes) overcomes the closing force. In the absence of a preload force, the valve opens with any pressure differential, but with a preload force present, a nominal pressure difference is required before the valve can open. The limits

Reciprocating Compressors Chapter

5

187

FIG. 5.5 Typical poppet valve configuration. (Courtesy of Ariel Corporation.)

FIG. 5.6 One degree of freedom valve model. (Courtesy of Ariel Corporation.)

of the valve plate travel are depicted in Fig. 5.7 with compliant structures for the valve seat (closed position) and the governor (maximum lift position), which can be used to simulate impact effects, such as valve bounce. Newton’s Second Law of Motion for the valve of Fig. 5.7 can be written as Eq. (5.2). Here, xv is the valve plate displacement from the free position of the valve seat, mv is the mass of the valve plate, and kv and cv are the stiffness and damping coefficients for the valve spring, respectively. Also, fw, v is the component of the valve plate weight vector acting along the line of action (typically

188 SECTION

II Types of Equipment

P2

Governor Flow xv,max – xv Valve plate xv P1 > P2 Seat

P1

FIG. 5.7 Theoretical model of valve. (Courtesy of Ariel Corporation.)

negligible), fPL, v is a preload force acting in the direction to close the valve, fP, v is the pressure force acting to open the valve (Eq. (5.3)), and fC, v is a contact force against the valve seat or governor (Eq. (5.4)). Furthermore, AP, v is the pressure area of the valve; Ps and Pd are the suction and discharge pressures, respectively; KC, v and CC, v are the contact stiffness and damping coefficients, respectively; and xv, max is the maximum valve travel or lift. mv x€v + cv x_ v + kv xv ¼ fw,v + fPL,v + fP,v + fC,v  AP,v ðPs  PÞ for suction valve fP,v ¼ AP,v ðP  Pd Þ for discharge valve 8 > < KC,v ðxv  xv,max Þ  CC,v x_ v for xv > xv, max fC,v ¼ KC,v xv  CC,v x_ v for xv < 0 > : 0 otherwise

(5.2) (5.3)

(5.4)

The flow rates through the valves (valid only when xv > 0) may be expressed as shown in Eqs. (5.5) and (5.6). Here, Cd, v is the valve flow coefficient, AF, v is the valve flow area, and ρs and ρd are the suction and discharge densities, respectively, while ρ is the density within the cylinder. Note that the flow coefficient and flow area are constants in this model since the equations include a term which accounts for the amount that the valve is open (i.e., Cd, v and AF, v are maximum values). Also note that the sign convention is consistent with positive suction flow into the volume and positive discharge flow out of the volume, but reverse flow is possible depending on the instantaneous cylinder pressure when the valves are open.

Reciprocating Compressors Chapter

8  pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi xv > > > Cd,v AF,v 2ρs ðPs  PÞ < x v,max   m_ s ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi > xv > > Cd,v AF,v 2ρðP  Ps Þ  : xv,max 8   pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi xv > > > Cd,v AF,v 2ρðP  Pd Þ < xv,max   m_ d ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi > xv > > Cd, v AF,v 2ρd ðPd  PÞ : xv, max

5

189

for Ps > P (5.5) for Ps < P for P > Pd (5.6) for P < Pd

Simulation and Example Performance Plots System Model This section implements the governing equations discussed in the previous sections to simulate the performance of a reciprocating compressor and demonstrate some basic concepts. A simple model for a crank-driven reciprocating compressor is depicted in Fig. 5.8. As the crank rotates, the connecting rod actuates the piston in a linear motion, where the travel of the piston from the top dead center position is described by Eq. (5.7). Here, θ is the crank angle, R is the radius of the crank, and L is the length of the connecting rod. The moving piston changes the volume of the cylinder according to Eq. (5.8), where V0 represents the clearance volume, and D is the diameter of the piston. The stroke of the piston is 2R, and the swept volume—the volume displaced by the piston from minimum to maximum displacement—is πD2R/2. For the crank rotating at a constant angular velocity ω ¼ dθ/dt, the respective piston velocity and rate of change of volume are as shown in Eqs. (5.9) and (5.10). 0 sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi1  2 R sin 2 θA (5.7) x ¼ Rð1  cos θÞ + L@1  1  L π V ¼ V0 + D 2 x 4

FIG. 5.8 Simple crank-driven reciprocating compressor system. (Courtesy of SwRI.)

(5.8)

190 SECTION

II Types of Equipment

2

3   R 6 sin θ cos θ 7 6 7 L 6 ffi7 x_ ¼ Rω6 sin θ + sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 7  2 4 5 R 1 sin 2 θ L

(5.9)

π V ¼ D2 x_ 4

(5.10)



Pressure within the cylinder (P) changes with an inverse relationship to the cylinder volume, which causes mass flow to be exchanged through the suction and discharge valves according to the suction pressure (Ps) and discharge pressure (Pd) and the valve dynamics. The physical principles governing these aspects of the reciprocating compressor operation are discussed in the subsequent sections. The following assumptions are considered for the simulation of the system: l

l

l

The suction and discharge volumes are sufficiently large such that the respective pressures are considered constant, and the gas velocities are low enough that dynamic pressure drop can be ignored. The operation of the piston is slow enough that the cylinder thermodynamic properties (i.e., pressures, temperatures, etc.) may be treated as quasi steady and without any spatial variation. The respective motions of the piston and valves are one-dimensional—that is, no misalignments and angular or orthogonal motion—and frictionless.

Recall Eq. (5.1) represented the first law of thermodynamics for the cylinder. As a simplification, heat transfer effects are ignored and piston seal _ + mu_ and leakage is neglected. With this, and also substituting E_ ¼ mu m_ ¼ m_ s  m_ d  m_ l , Eq. (5.1) is rewritten as Eq. (5.11). Other forms of the first law equation can be derived using ideal gas assumptions to relate pressure and temperature. However, depending on the gas species and conditions, this may not be appropriate. Eq. (5.11) is the form used to generate example results. 

P V + m_ s hs  m_ d hd ¼ ðm_ s  m_ d Þu + mu_

(5.11)

Example Calculations Simulations of a reciprocating compressor (Fig. 5.8) are presented in this section to demonstrate some aspects of reciprocating compressor performance. Baseline simulation parameters are listed in Table 5.1. Time integration is carried out with a simple Euler method and a time step providing just under 20 steps per degree of crank angle. The simulation is transient but is executed long enough to achieve the steady-state cycle. The steady cycles are displayed by truncating the transient portions.

Reciprocating Compressors Chapter

5

191

TABLE 5.1 Baseline Simulation Parameters Cylinder Fluid

Hydrogen

Bore, D

0.1 m

Stroke, 2R

0.1 m

L/R

4

Clearance/swept volume, V0/(πD R/2)

0.25

Crank rotational speed, 30 ω/π

400 rpm

Suction and discharge pressure, Ps and Pd

2000 and 7000 kPa

Suction and discharge temperature, Ts and Td

350 and 500 K

Valves

Suction Valve

Discharge Valve

Port area, AP, v

2

1e 4 m

1e4 m2

Flow area, AF, v

1e 4 m2

1e4 m2

Valve plate mass, mv

0.02 kg

0.02 kg

Valve stiffness, kv

2


pffiffiffiffiffiffiffiffiffiffiffi Valve damping ratio, cv = 2 kv mv

500 N/m

500 N/m

0.1

0.1

Preload, fPL, v/kv

0.003 m

0.003 m

Max lift, xv, max

0.003 m

0.003 m

Max flow coefficient, Cd, v

0.29

0.29

Contact stiffness, KC, v

5e6 N/m

5e6 N/m

Contact damping ratio,
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi CC, v = 2 KC, v mv

1.0

1.0

Baseline Results Fig. 5.9 shows the change in piston position and volume over time for two cycles. Note that the motion is not purely sinusoidal due to the effect of the connecting rod. As observed from Eqs. (5.7) and (5.8), the time traces of these quantities would approach a sinusoid as L/R ! ∞. Fig. 5.10 shows the steady-state time trace of the pressure for two cycles and the pressure vs. piston position. Similar looking plots could be made for temperature and density, but those are not shown for brevity. In an ideal compressor, the cylinder pressure would never exceed discharge pressure (red line) or go below suction pressure (blue line), but in a real compressor this occurs due to valve dynamics and

192 SECTION

II Types of Equipment

0.1

Piston position (m)

0.08 0.06 0.04 0.02 0

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.2

0.25

0.3

0.35

0.25

0.3

0.35

Time (s) 10–3

1

Volume (m3)

0.8 0.6 0.4 0.2 0

0

0.05

0.1

0.15 Time (s)

FIG. 5.9 Position and volume vs. time. (Courtesy of SwRI.) 8000

Pressure (kPa)

7000 6000 5000 4000 3000 2000 1000

0

0.05

0.1

0.15

0.2 Time (s)

8000

Pressure (kPa)

7000 6000 5000 4000 3000 2000 1000

0

0.01

0.02

0.03

0.04 0.05 0.06 Piston position (m)

0.07

FIG. 5.10 Pressure vs. time and piston position. (Courtesy of SwRI.)

0.08

0.09

0.1

Reciprocating Compressors Chapter

193

10–3

1.4

Mass in cylinder (kg)

5

1.2

1

0.8

0.6

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

Time (s) 10–3

Mass in cylinder (kg)

1.4

1.2

1

0.8

0.6

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

Piston position (m)

FIG. 5.11 Mass in cylinder vs. time and piston position. (Courtesy of SwRI.)

pressure losses through the valves. These losses represented in the metric called the PV efficiency, the ratio of the area inside the ideal PV loop to the area inside the actual PV loop. Fig. 5.11 tracks the mass in the cylinder over time and as a function of piston position. The increase in mass occurs while the suction valve is open and the discharge valve is closed, and the decrease in mass occurs while the suction valve is closed and the discharge valve is open. You can notice that there is a slight dip and increase in mass at the end of the discharge process, indicating that some reverse flow through the discharge valve occurs due to the valve not being able to close fast enough. This is also shown in Fig. 5.12, which presents the valve position and the mass flow through each valve over time. Finally, Fig. 5.13 presents the data of Fig. 5.10 in a slightly different way— here, pressure is normalized by suction pressure, and the ordinates are crank angle (instead of time) and normalized cylinder volume (instead of piston position). Note that the mass flow rate, compressor power, and PV efficiency are also summarized in this plot. Increased Gas Density Changing the fluid from hydrogen to air shows the effect of increased gas density (from 1.37 to 19.89 kg/m3 at suction conditions) and otherwise using the

194 SECTION

II Types of Equipment

Valve position (%)

150 Suction Discharge

100 50 0 –50

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

Time (s) 0.08 Suction Discharge

Mass flow (kg/s)

0.06 0.04 0.02 0 –0.02

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

Time (s)

FIG. 5.12 Valve position and mass flow vs. time. (Courtesy of SwRI.) 4 3.5

P / Ps

3 2.5 2 1.5 1 0.5

0

45

90

135

180

225

270

315

360

4.5

5

Crank angle (degrees) Mass flow: 4.47 e–03 kg/s, Power: 69.6 kW, PV Eff: 91.3%

4 3.5

P / Ps

3 2.5 2 1.5 1 0.5

1

1.5

2

2.5

3

3.5

4

V / V0

FIG. 5.13 Normalized pressure vs. crank angle and normalized volume. (Courtesy of SwRI.)

Reciprocating Compressors Chapter

5

195

5 4

P / Ps

3 2 1 0

0

45

90

135

180 225 Crank angle (degrees)

270

315

360

4.5

5

Mass flow: 4.95e–02 kg/s, Power: 73.2 kW, PV Eff: 71.0%

5 4

P / Ps

3 2 1 0

1

1.5

2

2.5

3 V / V0

3.5

4

FIG. 5.14 Baseline compressor performance with air. (Courtesy of SwRI.)

same compressor parameters of Table 5.1. Fig. 5.14 shows the basic performance using air, which can be compared to Fig. 5.13. It is clear that the valves are not sized properly for this gas density as the PV efficiency is reduced by 20 points. Also, there is reverse flow with both the suction and discharge valves as shown in Fig. 5.15.

Changing Piston Velocity Changing the piston velocity by increasing or decreasing the crank speed also has an effect on performance. For direct comparison with the baseline case (Table 5.1), only the speed is changed in the following results. Fig. 5.16 shows the compressor performance for 600 rpm, and Fig. 5.17 is for 200 rpm. It is clearly demonstrated that reducing piston velocity improves PV efficiency. In the case of the 200 rpm simulation, the valves are not optimized as the suction valve only remains fully open for about half of the suction process, and the discharge valve also begins closing prematurely (Fig. 5.18). This indicates that valve stiffness is too large.

120 Suction Discharge

Valve position (%)

100 80 60 40 20 0 –20

0

0.05

0.1

0.15

0.2

0.3

0.25

0.35

Time (s) 0.5 Suction Discharge

Mass flow (kg/s)

0.4 0.3 0.2 0.1 0 –0.1

0

0.05

0.1

0.15

0.2

0.3

0.25

0.35

Time (s)

FIG. 5.15 Baseline valve performance with air. (Courtesy of SwRI.) 4 3.5

P / Ps

3 2.5 2 1.5 1 0.5 0

45

90

135

180

225

270

315

360

4.5

5

Crank angle (degrees) Mass flow: 6.24e–03 kg/s, Power: 106.1 kW, PV Eff: 85.0%

4 3.5

P / Ps

3 2.5 2 1.5 1 0.5

1

1.5

2

2.5

3

3.5

4

V / V0

FIG. 5.16 Baseline compressor performance for 600 rpm. (Courtesy of SwRI.)

4 3.5

P / Ps

3 2.5 2 1.5 1 0.5

0

45

90

135

180

225

270

315

360

4.5

5

Crank angle (degrees) Mass flow: 2.30e–03 kg/s, Power: 33.1 kW, PV Eff: 97.2%

4 3.5

P / Ps

3 2.5 2 1.5 1 0.5

1.5

1

2

2.5

3

3.5

4

V / V0

FIG. 5.17 Baseline compressor performance for 200 rpm. (Courtesy of SwRI.)

120 Suction Discharge

Valve position (%)

100 80 60 40 20 0 –20

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Time (s) 0.05 Suction Discharge

Mass flow (kg/s)

0.04 0.03 0.02 0.01 0 –0.01

0

0.1

0.2

0.3

0.4

0.5

Time (s)

FIG. 5.18 Baseline valve performance for 200 rpm. (Courtesy of SwRI.)

0.6

0.7

198 SECTION

II Types of Equipment

Reduced Valve Stiffness Based on the previous result of the baseline compressor for 200 rpm, which indicated that valve stiffness was too high, suction and discharge valve stiffnesses were reduced by 75% and 50%, respectively. This modification only increased PV efficiency by 0.1 point, but the valve dynamic responses no longer displayed oscillation (Fig. 5.19). Baseline compressor and valve performance, utilizing a contact damping ratio of 0.3, is illustrated in Figs. 5.20 and 5.21.

Efficiency Considerations [4] Capacity The volume of gas compressed by this P-V diagram is the difference in volume between Positions 1 and 4: Capacity ¼ V1  V4 This volume is influenced by the compression ratio (RC) and the magnitude of V3. Compression ratio is: PD RC ¼ PS with absolute pressures 120 Suction Discharge

Valve position (%)

100 80 60 40 20 0 –20

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Time (s) 0.05 Suction Discharge

Mass flow (kg/s)

0.04 0.03 0.02 0.01 0 –0.01

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Time (s)

FIG. 5.19 Baseline valve performance for 200 rpm, reduced valve stiffnesses (suction—75%, discharge—50%). (Courtesy of SwRI.)

4 3.5

P / Ps

3 2.5 2 1.5 1 0.5

0

45

90

135

180

225

270

315

360

4.5

5

Crank angle (degrees) Mass flow: 4.40e–03 kg/s, Power: 69.1 kW, PV Eff: 90.5%

4 3.5

P / Ps

3 2.5 2 1.5 1 0.5

1.5

1

2

2.5

3

3.5

4

V / V0

FIG. 5.20 Baseline compressor performance with contact damping ratio ¼ 0.3. (Courtesy of SwRI.)

120 Suction Discharge

Valve position (%)

100 80 60 40 20 0 –20

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

Time (s) 0.08 Suction Discharge

Mass flow (kg/s)

0.06 0.04 0.02 0 –0.02

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

Time (s)

FIG. 5.21 Baseline valve performance with contact damping ratio ¼ 0.3. (Courtesy of SwRI.)

200 SECTION

II Types of Equipment

As compression ratio increases the capacity decreases as less gas is drawn into the compression chamber and compressed. As V3 increases there is more gas that must expand thus reducing capacity. V3 is referred to as the fixed clearance volume.

Fixed Clearance The fixed clearance volume (V3 in Fig. 5.2) is the volume of gas remaining in the compression chamber at the discharge end of the stroke that must expand from discharge pressure (PD) to suction pressure (PS) before the suction compressor valve can open and the suction event begin. Fig. 5.22 is a cross-section of a typical compressor cylinder assembly with some of the fixed clearance volume highlighted. The cavity between the compressor valve and cylinder bore might represent 70% of the total fixed clearance volume. There is also some volume between the head and the bore, the piston and the bore, and the piston and head. All of this volume adds together to form the total fixed clearance volume. Fixed clearance (CL) is expressed as a percentage of the piston displacement (referring to Fig. 5.2): Volumetric Efficiency VE is the percentage of stroke that fills with gas at suction pressure and suction temperature. In equation form (referring to Fig. 5.2): VE ¼

V 1  V4  100% V 1  V3

Some notes about VE: l l

VE represents the capacity. VE is not suction valve open time. The suction valves do not have to be open for the full VE.

FIG. 5.22 Compressor cylinder cross-section highlighting fixed clearance volume.

Reciprocating Compressors Chapter

l

5

201

A higher number for VE does not mean it is “better” as might be the case for energy efficiency. VE simply represents capacity. The influence of VE on energy efficiency is through the relationship of VE to average piston velocity (average velocity of gas through valves).

The equation for VE as derived from thermodynamics (or from the P-V diagram) is:    ZS PD 1 K 1 VE ¼ 100  CL ZD PS where VE ¼ volumetric efficiency, % CL ¼ fixed clearance, % ZS ¼ compressibility factor at PS and TS ZD ¼ compressibility factor at PD and TD PD ¼ discharge pressure, absolute PS ¼ suction pressure, absolute K ¼ adiabatic exponent, “K-value” A VE equation that might be used in compressor selection software might look like this:   ZS
 1 RC K  1 VE ¼ 100  RC  CL ZD Note the term “ RC” that has been added. Instead of subtracting from 100%, this equation subtracts from 100  RC. This term is intended to account for the fact a real-running compressor does not conform to pure thermodynamic theory. For example, the seals surrounding the compression chamber, specifically the compressor valves, piston rings, and packing, are not perfect. There is always some internal gas leakage. This means real VE will never agree with VE from theory. So a “fudge factor” must be used, and “ RC” is just such a factor. Every compressor manufacturer has a unique method for adjusting the VE equation and “ RC” is just one simple method. Compression ratio (RC) typically varies from 1.3 to 3.5 so the  RC term reduces VE (capacity) by 1.3%–3.5%. There can be a concern with VE being too low. Fig. 5.23 shows a P-V diagram (in red) that has low VE. The P-V diagram is very narrow and the discharge event is very short. This raises the possibility that the discharge valves may not have enough time to open and close properly and can cause the discharge valves to fail prematurely.

Power The power required to drive a reciprocating compressor can be divided into three pieces: adiabatic, valve loss, and friction and each will be discussed separately.

202 SECTION

II Types of Equipment

Pressure

PD

PS

0

V3

V4

Volume

V1

FIG. 5.23 Pressure-volume showing “low” volumetric efficiency.

The power required to compress a volume of gas is represented by the area enclosed by the P-V diagram, or: Z Work ¼ PdV The compression and expansion events are modeled thermodynamically as adiabatic processes, meaning it is assumed that no heat is transferred to or from the gas during these events. An adiabatic thermodynamic process is an isentropic (constant entropy) process. The area of the P-V diagram in Fig. 5.24 bounded by 1-2-3-4-1 is the adiabatic power. 2A 3

PD

Pressure

2

Discharge valve loss power

Suction valve loss power

4

PS

1

4A 0

V3

V4

Volume

V1

FIG. 5.24 Pressure-volume diagram highlighting suction and discharge valve loss power. (Courtesy of Ariel Corporation.)

Reciprocating Compressors Chapter

5

203

How valid is the assumption that the compression and expansion events are adiabatic? For a compressor with a rotating speed of 300 rpm (a slow rotating speed) one P-V cycle takes only 0.2 s to complete. Assuming each of the four events of the P-V cycle take equal time, that is 0.05s (or 50ms) per event. That is not much time for any significant amount of heat to transfer, therefore lending credibility to the adiabatic assumption. Yes, the gas does get hot as it is compressed but not from heat being transferred to the gas. That heat is the heat of compression. Inefficiency in the P-V diagram is the pressure drop incurred in moving the gas from the inlet flange of the cylinder into the compression chamber and in moving the gas from the compression chamber to the outlet flange. Overcoming this pressure drop requires energy. This energy is represented by the areas 1-4-4A-1 (suction valve loss power (VLP)) and 2-2A-3-2 (discharge VLP) in Fig. 5.24. It should be noted that in this initial ideal valve loss discussion, the assumption is that the gas at the cylinder flange is at a constant pressure, and that pulsation bottle and orifice plate pressure losses are ignored. These (very real) additional losses are discussed subsequently. This VLP represents the majority of the inefficiency in the P-V diagram. Additional small losses include piston ring and valve leakage, especially on nonlube machines. Other losses can arise if the gas temperature at the start of compression (point 1) is warmer than the incoming gas temperature, or if significant heat transfer between the cylinder walls and the gas occurs. Friction is the remaining inefficiency and is discussed later. The VLP can be expressed by the following relationship: VLP 

ðMWÞðPÞðVEÞðRP ÞðABORE Þ3 ðS  RPMÞ3 ðZÞðT ÞðN  AVLV PKT Þ2

where: MW ¼ gas mole weight P ¼ pressure, suction or discharge VE ¼ volumetric efficiency RP ¼ resistance factor ABORE ¼ cross-sectional area of the cylinder bore S ¼ stroke RPM ¼ rotating speed, rpm Z ¼ compressibility factor at suction or discharge T ¼ temperature, suction or discharge N ¼ number of suction or discharge valves feeding a head-end or crank-end compression chamber AVLV PKT ¼ cross-sectional area of the valve bore S  RPM ¼ piston speed, fpm. As used in this relationship, it is the average piston speed during the valve open time This relationship in an even simpler form V is discussed in some detail.

ðABORE ÞðS  RPMÞ ðN  AVLV PKT Þ

204 SECTION

II Types of Equipment

The first variable on the right side is pressure drop. Pressure drop is: ΔP  ρðV Þ2 where ρ ¼ density V ¼ velocity Density for a gas is. Þ ρ  PðZMW ðT Þ where

P ¼ pressure MW ¼ mole weight Z ¼ compressibility factor T ¼ temperature The velocity as used here is the average velocity of the gas as it moves through the valve bores as if the valves were not installed. That works out to: ΔP ¼

PðMWÞðABORE Þ2 ðS  RPMÞ2 Z ðT ÞðN  AVLV PKT Þ2

Substituting this relationship for velocity into the equation for pressure drop: ΔP ¼

PðMWÞðRP ÞðABORE Þ2 ðS  RPMÞ2 Z ðT ÞðN  AVLV PKT Þ2

This relationship represents the average pressure drop through the compressor valve bores in the cylinder body—as if the valves were not installed and the valve bores were simple orifices. Of course, what is needed is the pressure drop through the compressor valve. Adding the resistance factor term (RP) accomplishes this: ΔP 

PðMWÞðRP ÞðABORE Þ2 ðS  RPMÞ2 ZðT ÞðN ÞðAVLV

PKT Þ

2

Resistance factor is defined as the ratio of measured pressure drop across a compressor valve to the pressure drop that would be predicted in flowing the same quantity of the same gas at identical upstream pressure and temperature conditions through a round hole (an orifice) having a discharge coefficient equal to one and an area equal to the valve pocket opening. Typical resistance factors range from 30 to 200. Meaning a compressor valve could have 30–200 times the pressure drop as an orifice the same diameter as the compressor valve. Note that resistance factor is a dimensionless number as it is pressure divided by pressure. So resistance factor is: RP ¼

Compressor Valve ΔP Orifice ΔP

Another term used in the same manner is valve equivalent area (VEA). VEA has units of area. VEA is the orifice area required to generate the same pressure

Reciprocating Compressors Chapter

5

205

drop as that through a compressor valve when flowing the same quantity of the same gas at the same pressure and temperature. Compressor and compressor valve manufacturers will use either term (resistance factor or VEA) to describe the relative efficiency of a compressor valve. One can be converted into the other: VEA ¼

AVLV PKT ðAVLV PKT Þ2 pffiffiffiffiffiffi or RP ¼ RP ðVEAÞ2

Some further discussion about the “S  RPM” term in the above relationships is required. This term is generally known as piston speed, or the average linear speed at which the piston moves through one stroke length. Average piston speed in feet per minute is calculated by: PS ¼

2  S  RPM S  RPM , or PS ¼ 12 6

where PS ¼ piston speed, feet per minute S ¼ stroke, in RPM ¼ rotating speed, rpm Fig. 5.25 is a plot of instantaneous and average piston speed vs. crank angle: Instantaneous piston speed reaches a maximum near the middle of the stroke but not exactly the middle (90 degrees of rotation). Note that the maximum piston speed is about 60% greater than the average.

180 160 Instant Average

Piston speed (%)

140 120 100 80 60 40 20 0

0

20

40

60 80 100 120 Crank rotation (degrees)

140

160

180

FIG. 5.25 Plot of piston speed (as percent of average) vs. crankshaft rotation. (Courtesy of Ariel Corporation.)

206 SECTION

II Types of Equipment

Average piston velocity (%)

120 100 80 60 40 20 0 0

20

40 60 80 Valve open time, percent of stroke (%)

100

FIG. 5.26 Plot of piston velocity vs. compressor valve open time. (Courtesy of Ariel Corporation.)

But the velocity used in the above relationships for valve pressure drop and VLP is the average piston speed during the time the compressor valves (suction or discharge) are open as shown in Fig. 5.26. For example, if the suction valve is open for 40% of the stroke, the average piston velocity would be about 87% of the full stroke average velocity. If the above relationships are substituted back into the VLP equation, the following results are obtained: VLP 

PðMWÞðVEÞðRP ÞðABORE Þ3 ðS  RPMÞ3 ðVEÞ ZðT ÞðN  AVLV PKT Þ2

VLP and friction represent all of the inefficiency in a reciprocating compressor (not considering pressure drop in getting gas to and from the compressor and possible efficiency losses due to P-V diagram distortion resulting from gas pulsation). The magnitude of typical friction power is 5%, meaning the majority of the inefficiency is associated with VLP. Some comments about VLP: l

l

VLP varies with (S  RPM)3. This is a large number and therefore significantly impacts VLP. VLP varies with stroke and rotating speed and not just rotating speed. Comments are sometimes made that “high-speed compressors are inefficient.” This is not correct. A more accurate statement is “high piston speed compressors are relatively inefficient.” The following table lists several combinations of stroke and rotating speed that result in the same piston speed:

Reciprocating Compressors Chapter

Stroke (in.)

Rotating Speed (rev/min)

21.0 19.5 18.0 16.5 15.0 13.5 12.0 10.5 9.0 7.5 6.0 4.5 3.0

257 277 300 327 360 400 450 514 600 720 900 1200 1800

5

207

Piston Speed (ft./min) 900 900 900 900 900 900 900 900 900 900 900 900 900

Everything else being equal (admittedly very difficult) all of these combinations would have the same relative compression efficiency. l

l

l

VLP varies directly with MW. For example, a hydrogen compressor (MW ¼ 2) would have 89% less VLP as compared to the same compressor compressing natural gas (MW ¼ 18) simply because of the very low MW. The very basic relationship between the cylinder bore diameter (ABORE 3 ) and the number and size of the compressor valves (N  AVLV PKT 2 ) determines the basic efficiency of a given cylinder. In simple terms—the larger the valves for a given cylinder bore diameter the better the efficiency. As the cylinder bore increases in diameter the relative efficiency decreases. This results from simple geometry:

The bore area grows by the bore diameter to the second power, but the circumference, which is the space available to locate compressor valves (represented by the rectangles in Fig. 5.27), only grows by the bore diameter to the first power.

D Bore area =

πD2 4

Circumference = πD FIG. 5.27 Drawing showing the space available for compressor valves in a typical compressor cylinder design. (Courtesy of Ariel Corporation.)

208 SECTION

II Types of Equipment

Fundamentally, a compressor cylinder is made more efficient by using larger compressor valves for a given cylinder bore diameter (everything else being equal). But something else happens with the compressor valves being larger—the fixed clearance becomes larger. Larger fixed clearance results in lower VE which is lower capacity. A given cylinder diameter with more and/or larger compressor valves will compress less gas, but compress that gas with better energy efficiency (lower power per capacity). The cylinder designer must make a trade-off between compression efficiency and VE by optimizing valve flow area and clearance. The cylinder clearance is the ratio of fixed clearance to swept volume. Most of the fixed clearance is in the valves and valve clearance C. The swept volume is πD2S/4, and the resultant ratio is 4C/πD2S, so the percent clearance is proportional to the inverse of stroke. For example, a 1000 diameter cylinder on a 600 stroke machine might have 20% clearance and operate at 900 rpm. However, the same diameter cylinder with the same valves on a 1200 stroke would operate at 450 rpm and would have the double the swept volume per stroke but the same displacement per minute. The clearance would be only half or 10%. In practice, however, the cylinder designer makes the short stroke machine a nonlinear design which cuts clearance considerably. On long stroke cylinders with small bores more flexibility is available to maximize the valves.

Friction A reciprocating compressor is a mechanical device and as such encounters and must overcome friction. Friction is accounted for very simply: BP ¼

IP M:E:

where BP ¼ brake power IP ¼ indicated power M.E. ¼ mechanical efficiency, typically 95%–97% A definition of indicated power is: Adiabatic Power +Suction Valve Loss Power +Discharge Valve Loss Power Indicated Power Indicated power is all the power derived from the P-V diagram. Brake power is then the total power required to be input to the compressor to get the indicated power to the gas (for the P-V diagram). Friction develops due to the crankshaft turning in the bearings, driving oil pumps, windage, crossheads sliding in the crosshead guides, packing rubbing against the piston rod, piston rings, and wearbands rubbing against the cylinder

Reciprocating Compressors Chapter

5

209

bore and other items. Friction manifests itself as heat, for example, causing the crankcase oil to get hot. Compression Efficiency Compression efficiency is usually defined as the adiabatic efficiency also known as isentropic efficiency and so is the ratio of the adiabatic power to the PV card indicated power: EFF ¼ AP=IP  100% where EFF ¼ compression efficiency AP ¼ adiabatic power IP ¼ indicated power Fig. 5.28 is a plot of compression efficiency vs. compression ratio for a given compressor cylinder when compressing two different gases, hydrogen and nitrogen: Note how efficiency increases as compression ratio increases. The efficiency curve will have this shape for any reciprocating compressor cylinder. Also note how much higher the efficiency is for hydrogen (with a very low MW of two) as compared to natural gas (with a medium MW of 18). Everything else being equal, compressing hydrogen will have one-ninth (11%, 89% less) the VLP, hence much better efficiency.

Pulsation Bottle Losses The losses in a pulsation bottle can be considered in two parts, the relatively static portion which is the pressure drop at the line connection and the choke

FIG. 5.28 Plot of compression efficiency vs. compression ratio.

210 SECTION

II Types of Equipment

tube restriction between the two volumes in a volume-choke-volume acoustic filter. As a result, the pressure at the volume side of the cylinder nozzle is near constant pressure. The other part is a dynamic loss which occurs due to the orifice plate on the cylinder flange plus the acceleration and deceleration losses and pulsation of the gas in the bottle nozzle. The steady-state loss has the effect of moving the inlet and discharge line on the PV card Ps and Pd by the amount of the loss (0.5%–1.0% pressure drop is typical). The dynamic nozzle and orifice loss appears on the PV card as a part of the valve loss. Taken together these losses will typically add between 2% and 5% to the area of the PV card (power).

Cylinder Cooling Effects Cylinder cooling continues to be embraced by the downstream (refining) industry, where the upstream and midstream segments have fully embraced noncooled cylinder designs. The origins of using of water jackets in reciprocating compressors are somewhat uncertain. An article published in an 1891 edition of The Scientific American states that water jackets were used in an effort to achieve near isothermal compression in early air compressor designs. According to the article, the subject very long stroke, very slow-speed compressors showed increased efficiencies when cooling water jackets were utilized. Cylinder cooling is required by API-618 such that the coolant temperature is 6°C above the inlet gas temperature to avoid the possibility of condensation. This helps to minimize potential issues caused by corrosives in the process gas and of possibly washing the lubricant away. This function is most useful during compressor start-up, as the heat of compression will usually provide protection in operation. Cylinder cooling is most effective for partially and fully unloaded cylinders. Consider a cylinder partially unloaded by delayed suction valve unloaders. As the volume efficiency is reduced the discharge temperature will rise due to the throttling losses through the unloaded inlet valve and the reduced gas flow available to remove the heat. In fact such an unloader system will normally have a minimum flow below which the temperature in the cylinder becomes excessive. However if water cooling is used the excess heat can be rejected to the water jacket allowing lower permissible capacity before overheating occurs. In a fully unloaded cylinder the throttling losses may cause the cylinder to overheat unless the heat can be removed. Careful consideration of the unloaded cylinder power consumption vs. the heat rejection capability is required to determine if the cylinder will overheat. Cylinders unloaded by suction valve unloaders typically will have an unloaded kilowatt of 5%–10% of the full-load power. So consider a 750 kW cylinder, if the unloaders result in 75 unloaded kW then this is probably too much for a water jacket and the unloaded time must be limited. However, if efficient unloaders are used and there is only 37 unloaded kW then the water jacket may well be adequate to remove the heat within the temperature limitations and long-term unloaded operating is acceptable.

Reciprocating Compressors Chapter

5

211

Noncooled Cylinders In September 1955, a technical paper was presented at an ASME conference that offered a contrasting view to the need for cylinder cooling systems and ultimately had a major impact on some future compressor designs. The paper was titled “Operation of Compressor Cylinders Without Cooling Water” and was written by H.W. Evans and J.L. Gallagher, two Sinclair Oil and Gas Company engineers based in Tulsa, Oklahoma [22]. In 1940, while performing maintenance on a cylinder cooling system, the decision was made to keep the compressor running for 3 days. During that period, there was no noticeable decrease in compressor performance or increased discharge temperatures. This led to the study eventually detailed in the paper. For 2 years, the cooling water systems for all of the compressors at the Sinclair facility were turned off and drained. The performance of all the compressors was monitored and the following conclusion was reached and published in the paper: “After more than two years of experience in operating plants without the use of jacket water in the compressor cylinders, we believe there is no measurable loss in gas handling capacity and no increase in cylinder wear. On all new installations, we have discontinued the use of cooling water on all gas compressor cylinders and thereby have saved the installation cost of the pumps, headers, compressor cylinder water piping and have reduced water usage, treating cost and operating and maintenance cost.” Capacity Control Capacity control can be achieved in several ways with a reciprocating gas compressor. In the following discussion a distinction will be made between methods that are efficient (reduced power with concurrent reduced mass flow so that the power per unit gas is held constant) and those that are not. The following methods briefly discussed here include: – – – – –

recycle; varying rotating speed; adding/subtracting fixed clearance; end deactivation; and suction valve controlled closing

Recycle Recycle, as summarized in Fig. 5.29, is the simplest and the therefore the most reliable form of capacity control. It consists of piping that connects the discharge piping with the suction piping through a control valve—a recycle valve. When the valve is opened discharge gas is allowed to flow back to suction and not proceed on down the discharge pipe thus reducing the amount of gas the proceeds down the pipe. While the capacity is changed (lowered when the recycle valve is opened) the power is not, making recycle an inefficient form of capacity control.

212 SECTION

II Types of Equipment

FIG. 5.29 Example of capacity control using recycle methods. (Courtesy of Ariel Corporation.)

Variations in Rotating Speed Varying the rotating speed is a simple and efficient method of capacity control. When the speed is reduced the capacity and power are reduced. If the driver is an electric motor then there is substantial cost in the VFD and the VFD draws approximately 2% of the motor power. However with an engine driver the variable speed comes basically for free. A drawback of varying speed is that the compressor valves now have to function reliably over a range of speed which can be difficult. Also the varying speed might cause the compressor to operate closer to frequencies that can excite pulsation or torsional-related issues. Adding/Subtracting Fixed Clearance Adding and subtracting fixed clearance is one of the most common methods used to control the capacity of a reciprocating compressor. When fixed clearance is added to the compression chamber of a compressor cylinder the VE is reduced thus reducing capacity. This is an efficient method of capacity control. There are many mechanical devices used to add clearance. Some of the most common as follows: – – – –

head-end fixed volume clearance pocket; head-end variable volume clearance pocket; valve cap fixed volume clearance pocket; and valve clearance spacer.

Head-End Fixed Volume Clearance Pocket This device replaces the standard head-end head. When the plug valve opens, fixed clearance is added to the compression chamber and the capacity decreases. This device also replaces the standard head-end head. As opposed to opening and closing a fixed volume, the volume is variable, as shown in Fig. 5.30. Most common is a manually operated device, as shown in Fig. 5.31).

Reciprocating Compressors Chapter

5

213

FIG. 5.30 Head-end variable volume clearance pocket. (Courtesy of Ariel Corporation.)

FIG. 5.31 Manually operated variable volume clearance pocket. (Courtesy of Ariel Corporation.)

A version of this device is automatically actuated, using electric or pneumatic motors, or by using hydraulic actuators, and is becoming more common. The advantage is this device can be tied directly to a control system allowing the clearance, and therefore the capacity, to be adjusted completely

214 SECTION

II Types of Equipment

FIG. 5.32 Valve cap fixed volume clearance pocket. (Courtesy of Ariel Corporation.)

automatically thus allowing the compressor to match demand at all times (Fig. 5.32). A valve cap fixed-volume clearance pocket is a device installed over a compressor valve. The added fixed volume is typically less than that provided by a head-end pocket simply due to less space being available over a valve. A disadvantage of this device is that it requires a hole through the compressor valve to provide a passage to the added clearance volume. This hole removes active valve flow area reducing the efficiency of that valve.

Valve Clearance Spacer A valve clearance spacer is a metallic ring that is installed between a compressor valve and the cylinder body effectively moving the compressor valve farther from the cylinder bore adding fixed clearance. The amount of added clearance is small so this device is most often used to provide fine tune of the interstage pressures in a multistage compressor. Of course, this device cannot be added or removed while the compressor is running. End deactivation, as shown in Fig. 5.33, refers to the situation where an end of a double-acting cylinder, usually the head end, is deactivated effectively reducing the capacity of that cylinder 50%. There are three devices commonly used to deactivate an end: – finger-type suction valve unloader; – plug-type suction valve unloader; and – head-end bypass.

Reciprocating Compressors Chapter

5

215

FIG. 5.33 End deactivation. (Courtesy of Ariel Corporation.)

All three devices provide a means for gas in the compression chamber to never be compressed and return that gas to suction. In other words, create an internal bypass. As mentioned earlier end-deactivation reduces the capacity of a cylinder 50% but the power does not reduce 50% because the deactivated end consumes power even if the capacity is zero. The pressure-volume diagram for a deactivated end looks like this (red) (Fig. 5.34): The amount of power absorbed by a deactivated end will vary mostly with piston speed, gas MW, and pressure level. Unloaders are relatively efficient. For example, if one end is unloaded so the capacity of the cylinder drops to 50%, the power might drop to 60% due to the throttling losses in the unloader.

Finger-Type Suction Valve Unloader A finger-type suction valve unloader is a device that holds the suction valve seal elements open for 100% of the crankshaft rotation. Suction gas is drawn into the compression chamber during the suction stroke and returned to suction during the discharge stroke by flowing backward through the suction valve. A plug-type suction valve unloader (Fig. 5.35) creates a bypass to suction utilizing a hole through a suction valve. The hole is opened and closed using a plug valve. When the plug valve opens (moves of its seat), the hole opens allowing a direct connection between the compression chamber and the suction gas passage—creating the bypass. A disadvantage of this device is the hole that

216 SECTION

II Types of Equipment

Pressure

Pd

Ps

0

Vmin

Volume

Vmax

FIG. 5.34 Pressure-volume diagram for a deactivated end (red). (Courtesy of Ariel Corporation.)

FIG. 5.35 Plug-type suction valve unloader. (Courtesy of Ariel Corporation.)

Reciprocating Compressors Chapter

5

217

FIG. 5.36 Suction valve closing regulation. (Courtesy of Ariel Corporation.)

is required through the active suction valve. This removes active valve flow area thus reducing the efficiency of this valve. Suction valve closing regulation (Fig. 5.36) provides a mechanical means to hold suction valves open past the point of the pressure-volume cycle where they would normally close and remain closed during the compression portion of the cycle. This allows suction gas to flow backward through the suction valves effectively reducing the capacity of that pressure-volume cycle. Typical minimum flows are to about 20% of a cylinder end’s capacity, or an 80% reduction. The capacity is then completely variable between 100% and about 20% (Fig. 5.37). Head-End Bypass A head-end bypass is another mechanical device used to deactivate the head end of a double-acting compressor cylinder. The device replaces the standard

218 SECTION

II Types of Equipment

FIG. 5.37 Component layout for suction valve closing regulation. (Courtesy of Ariel Corporation.)

FIG. 5.38 Typical head-end bypass device. (Courtesy of Ariel Corporation.)

head-end head and connects, through a valve, the head-end compression chamber to that cylinder’s suction pressure. A typical head-end bypass device is shown in Fig. 5.38. The plug valve in the bottom half of the sketch is closed and open in the top half. With the valve closed the head-end compression chamber is compressing gas. With the valve open the compression chamber is now directly connected to

Reciprocating Compressors Chapter

5

219

FIG. 5.39 Reciprocating compressor with small diameter high pressure cylinders. (Courtesy of Ariel Corporation.)

suction pressure creating a bypass. When the piston is on the compression and discharge portion of the cycle the gas simply goes back to suction and is not compressed thus bypassing and the head end is deactivated. The head-end bypass is attractive to use when the suction pressure is too high for a finger-type suction valve unloader (about 7 MPa and higher) or when the compressor valves are too small to accommodate a hole through the center for a plug-type suction valve unloader. This means the head-end bypass can be an attractive option for use on small diameter higher pressure cylinders (Fig. 5.39).

Pulsation Control Reciprocating compressors receive and deliver gas in discrete slugs (or pulses) as the pistons moves back and forth. For a double-acting compressor there will be two pulses of gas for each revolution. As these pulses of gas enter or exit the compressor cylinder they will give rise to pressure pulsation that moves in both directions at the speed of sound down the pipe. These pulses can be quite large and so attenuation devices generally called “pulsation bottles” or “dampers” are mounted on the inlet and discharge of the cylinder. Usually there will be one, two or three or even more cylinders per stage and so all the cylinders operating in parallel on each bank will have a common pulsation bottle. So for a twothrow single-stage compressor there are one cylinder on each bank therefore there will be two inlet and two discharge pulsation bottles. For a six-throw two-stage machine the three first-stage cylinders would normally be on one

220 SECTION

II Types of Equipment

bank using one common inlet and one common discharge bottle, similarly for the second stage on the opposing bank. For the case of a single double acting cylinder, the H.E. and C.E. pulses are not identical, because they have different swept volume (C.E. is less due to area of piston rod. Internal clearances are often different and the instantaneous piston velocity is different on the outboard and inbound stroke due to the slider crank mechanism. Note that because there are two pulses of gas the second harmonic (2  running speed) dominates, but there is a small excitation at almost all the harmonics. For a four throw, the two cylinders sharing a bottle will have pulses every 90 degrees of crank rotation so the max pulse harmonic with be 4 . For a six throw, three cylinders will share a bottle and with be phased every 120 degrees (pulses every 60 degrees for double-acting cylinders) and the sixth harmonic will dominate. There are two basic types of pulsation damper that are commonly used. Surge volume, here the cylinder feeds into a simple volume bottle. This uses the principle of a volume and some pressure drop, either an orifice plate or choke at the bottle line connection. Typically, volumes of 10–30  cylinder swept volume and pressure drop at the line connection of 1% are adequate to meet the API-618 pulsations requirements. This bottle may have an internal baffle to control the shaking forces and an orifice at the cylinder nozzle if necessary to control the quarter wave nozzle resonance. Simple surge volume bottles are typically used on low-molecular-weight gases and slow-speed compressors. Volume-choke-volume acoustic filters. These are highly effective pulsation control devices where the two volumes and connecting choke are designed so that the Helmholtz resonant frequency is below either the 1  or in case of double-acting cylinders with very low 1 pulsation, below the 2  frequency. Then the pulsation bottle will effectively attenuate almost all frequencies greater than 1.4 the Helmholtz frequency. Acoustic filters are used on high-speed compressors and at higher molecular weights (to the right of the curve). Surge volumes are typically used for applications on the left of the curve and acoustic filters to the right. For guidance in sizing acoustic filters refer to Ref. [20].

Bottle Design The sizing and selection of pulsation bottles are done as a part of a pulsation study which is an acoustic simulation of the cylinders and piping and is done in accordance with API-618 (Reciprocating compressors and API-688 (pulsation and vibration control for positive displacement compressors—due to be published in 2019). API 618/688 describes the theory of pulsation control and provides limits for pulsation levels at the cylinder and also in the piping

Reciprocating Compressors Chapter

5

221

FIG. 5.40 Pressure pulsation complex wave and resulting Fourier analysis of pulsation amplitude vs frequency. (Courtesy of SwRI.)

away from the compressor. It provides acceptable limits for pressure drop in the pulsation bottle as a function of the compression ratio. It should be noted that the packager can do a lot to reduce pulsation levels by selecting pipe routing and lengths. For example, in Fig. 5.40 note that a doubleacting cylinder has 2 pulses per revolution. So it will have a strong 2  running speed pulsation. Now if this is a 6-throw single-stage compressor with three cylinders on each bank feeding into one common pulsation bottle on each bank then Fig. 5.40 would have 6 pulses for each pulsation bottle. The pulses would better approximate a continuous flow and so now the dominant pulse would be 6  and would be considerably less strong than the single throw. Also if the bottle on each bank was piped together to the common manifold out of phase rather than in phase then the resulting Fig. 5.40 at the manifold would now have 12 pulses resulting in a 12  pulse that is quite low in strength. Considering a 1200 rpm natural gas compressor the pulsation frequency at each three cylinder bank would be 20 revolutions/s  6 pulses per rev giving 120 pulses per sec. The sonic velocity is approximately 519 M/s so the wavelength for a 120 Hz pulse is 519/120 ¼ 4.3 m. Ideally the pulses from the opposing banks will come together at the header out of phase (phased at 180 degrees) so the piping from the two banks to the header should ideally be 4.3/2 ¼ 2.15 m different in length. This should be quite simple and minimal cost to do during the package design phase and is highly effective in reducing pulsations and this can be reflected in reduced size and cost of the pulsation dampeners. There is a trade off in sizing pulsation bottles. For a given pulsation limit the designer can use smaller bottles with higher pressure drop (higher friction loss at the orifice plate and choke) vs. large bottles with lower pressure drop. But the

222 SECTION

II Types of Equipment

user is cautioned about the use of excessively large bottles. These have two technical problems, potentially high shaking forces due to the gas pulsation acting on the baffles and because they are large and heavy the difficulty in supporting safely above the compressor cylinders. Piping and bottles are usually designed to have a mechanically resonant frequency above 2.4  the running speed. Especially on high-speed machines the 2.4 limit can be difficult to achieve and it is often necessary to consider secondary bottles as a part of the pulsation control such as the separator or a secondary pulsation bottle.

Cylinder Orifice Plates Gas enters (and leaves) a cylinder in two discrete slugs per revolution for a double-acting cylinder. Depending on the velocity of the gas flowing through passageway the velocity head results in a pressure pulsation. Applying an FFT analysis to this gas flow gives the harmonic content of the flow. Considering the gas passageway from the compressor valve to the pulsation bottle volume this acts as a closed-open pipe and the column of gas has a natural frequency of a quarter wave. So for 1200 rpm natural gas compressor with a sonic velocity of 519 M/s and a distance from the valve to the bottle volume of 1.2 m, the quarterwave resonant frequency would be 519/(1.2  4) ¼ 108 Hz. The operating speed of 1200 rpm is 20 Hz so the quarter-wave resonant frequency is 108/20 ¼ 5.4  running speed. Looking at Fig. 5.40 a double-acting cylinder will have significant pulsation energy at 4 and 6 running speed. Considering that the sonic velocity for a given gas composition will vary primarily with gas temperature (higher temperatures have higher sonic velocity), then the pulsation analyst will often consider that an orifice plate in the cylinder nozzle is necessary to dampen the quarter-wave pulsation resonance. The orifice plate creates a pressure drop that is quite effective at damping the quarter-wave resonance (usually at 4  or 6 machine speed), but this does nothing to reduce the 2 pulsation, in fact it is made worse. In addition, the pressure drop will cause an increase in overtone and undertone losses in the PV card causing an increase in compression HP and rod load. API-618 gives an allowable pulsation bottle pressure drop of ΔP ¼ 1.67((R  1)/R)%. This gives a value of 1% when the pressure ratio is 2.5. Many packagers use a value of 1% pressure drop for all applications even though this is higher than the allowance for compression ratios below 2.5. In addition, the limit is calculated on the basis of Steady flow. As can be implied from Fig. 5.40 the steady flow is only a small fraction of the peak flow rate for the nozzle orifice plate, usually around 2  on the inlet and 3  on the discharge is typical. As pressure drop through an orifice is a function of flow velocity squared the effective instantaneous pressure drop through the orifice plate is typically 4–10 the value calculated based on steady flow through the orifice plate. This can have a significant effect on compression horsepower and rod load possibly causing overloading. As previously noted resonant pulsations do need to be controlled and API-618 gives allowable pulsation values at the

Reciprocating Compressors Chapter

5

223

cylinder nozzle, but the point of interest is actually the compressor valve which can be damaged by excessive high-frequency gas pulsations. But too often orifice plates are used “just to be safe” and sized basis average flow. It is recommended that orifice plates be sized based on instantaneous flow so that the effective pressure drop is calculated. A low (nonconservative) estimate of effective flow rate can be estimated based on the total flow divided by the percentage of time the valve is actually open (70% on inlet and 50% on discharge is typical). The true pressure drop will be even higher that this calculation, but this is a useful and simple to calculate method of estimating cylinder nozzle orifice plate pressure drop. For further discussion of this topic refer to Ref. [21].

Nozzle Pulsation As noted previously the column of gas on the outside of the valve all the way back to the junction of the pulsation bottle nozzle-pulsation bottle goes from being stationary while the valve is closed and has to accelerate to maximum velocity when the valve opens and then slows back to zero and the piston reaches the end of the stroke. This acceleration and deceleration of the gas causes pressure pulsations. The severity of the pressure pulsation is related to the velocity head of the column of gas, so it is related to the gas velocity, the gas acceleration and the mass of the gas column. This effect is the dominant cause of pressure pulsations in the inlet and discharge nozzle at 1  and 2 machine speed. The 1  and 2 frequency are nonresonant. In the paper by Hinchliff and Bratek (Ref. [7], a discussion of the loads used to rate reciprocating compressors, GMRC 2014) a severity index was proposed to indicate the severity of pressure pulsation “severity index ¼ (specific power  gas velocity  S.G.  rpm  column length)/60,000” Where specific power is the cylinder IHP/cylinder power rating, this is just a measure of how heavily loaded the cylinder, high-loaded cylinders are problematic than lightly loaded. Gas velocity is the average piston speed  (piston area/cylinder nozzle passageway area) fpm, S.G. is the gas specific gravity so is MW/28, rpm is the compressor rotation speed, column length is the linear distance along the gas passageway from the compressor valve to the nozzle-pulsation bottle volume junction in feet. The paper proposed that values less than 30 resulted in insignificant pulsation, whereas greater 60 resulted in significant levels of pulsation. Significant pulsation levels may exceed the API 618 limit at 1  and 2  and may cause increased rod load and horsepower over that which is calculated in the OEM’s compressor performance program. Note that cylinder orifice plates will do little or nothing to reduce 1  and 2  pulsations, indeed as they simply apply a resistance they will normally make the situation worse. Orifice plates are useful to reduce resonant pulsations which are typically at 4 or 6  or greater harmonics. 1  and 2  pulsations can only be reduced by reducing the severity index, which means reducing the velocity head of the gas. (Use a weld neck flange nozzle to

224 SECTION

II Types of Equipment

maximize nozzle diameter, reduce the compressor speed, select a cylinder with larger inlet and discharge flange size, use a higher power rating compressor, etc.)

Cylinder Lateral Gas Pulsation Effects A basic assumption in compressor performance is that pressure everywhere inside a cylinder is equal. In fact, this is often not the case. The worst point is when the discharge valve opens, then gas next to the valve starts to flow out, but gas remote from the valve can only begin to move toward the valve when a pressure wave traveling at the speed of sound reaches the gas. This results in an over pressure in the cylinder causing a pressure wave pulsation which travels laterally across the face of the piston from discharge to inlet valve. This pressure wave goes back and forth across the piston face and gradually decays as the piston reaches the end of the stroke. The strength of the pressure wave is function of three things: l

l

l

The relative cylinder diameter as indicated by the degrees of crankshaft rotation it takes for a sound wave to traverse the cylinder. (This factor includes cylinder diameter, rotative speed, gas sonic velocity.) Volume in the cylinder when the discharge valve opens, that is, the lower the fixed volume and closer the piston is to the end of stroke the worse the pulsation. Approximated as 1/R1/n where R is the compression ratio and n is the volume exponent. Instantaneous piston velocity at the time the discharge valve opens. (Piston velocity is a maximum at midstroke but starts to drop significantly as it reaches 75% stroke.)

The degrees of crank rotation for a sound wave to traverse the cylinder can be calculated from the formula A ¼ 6ND/u where A is the angle in degrees, N is the rotational speed in rpm, D is the cylinder diameter in meters, and u is the sonic velocity in m/s. The low medium and high level correspond to approximately 5%, 10%, and 15% p-p pulsation as a percent of the discharge pressure. Effect of high lateral pulsation; high pulsation will have an adverse effect on discharge valve operation and will impose a large bending moment on the piston and cylinder. This can result in piston rocking and can result in piston reliability problems in sensitive pistons (large diameter compared to piston length and rod diameter), cylinder vibration (rocking mode about horizontal plane at 90 degrees to piston rod axis).

Rotordynamics Torsional Rotordynamics Torsional failures are sometimes referred to as “silent killers,” because the events are rarely preceded by an increase in frequently monitored vibration

Reciprocating Compressors Chapter

5

225

signals and are often catastrophic. Typical package instrumentation is focused on the monitoring of lateral vibration probes, and most often does not include provisions for monitoring torsional oscillation. In addition, reciprocating machinery tends to generate significantly large forcing torques at discrete orders of operating speed, which can result in problematic excitation of critical speeds. A proper torsional analysis is key to avoiding these types of problems. The intent of this section is not to provide a listing of requirements for a complete torsional analysis, as several notable resources are available that outline such information (e.g., see the GMRC publication Guideline and Recommended Practice for Control of Torsional Vibrations in Direct-Driven Separable Reciprocating Compressors [5], and API 684 [6], among others). Rather, the following discussions provide a brief overview of the various components involved in a typical torsional analysis, and some practical guidance for dealing with common issues encountered during this type of work when reciprocating machinery is involved. Torsional evaluations generally involve a steady-state analysis, preparation of interference diagram(s), forced response analyses, and transient analyses (if necessary and applicable). These analyses are most often conducted for trains involving compressors, pumps, motors, turbines, or engines, and any associated rotating components such as gearbox shafts, couplings, and viscous dampers. One common question that arises when considering this type of effort is when such an analysis is required, and what type of analysis is necessary. Generally speaking, a complete torsional analysis is recommended when a nonduplicate system is installed, or significant operating condition changes are planned. Section 2.2 of the GMRC document [5] referenced above also provides a methodology for determining what types of analyses are needed in specific instances. The steady-state torsional analysis involves preparation of a torsional masselastic model derived from manufacturer provided drawings and mass elastic information data for the equipment involved. The steady-state torsional natural frequencies (critical speeds) and mode shapes are then calculated. The mode shapes are plotted to graphically show the deflection associated with the torsional natural frequencies. This allows for investigation of controlling stiffness, and relative model participation of the major inertias in the system. An interference diagram (Campbell diagram) is prepared to graphically display prevalent excitation energy orders vs. speed and frequency. The specified operating speed range(s) are superimposed on the interference diagram, and any coincidences between calculated critical speeds and prevalent excitation energy in the system are identified. The likelihood of exciting the critical speeds involved in any such coincidences is assessed by studying the respective mode shapes and potential excitation mechanisms. A forced response analysis is usually conducted to determine anticipated stress levels in the shafting during normal operation. For reciprocating equipment, these calculations involve the characterization of dynamic torque content for nominal operations, in addition to single cylinder engine misfire and locked

226 SECTION

II Types of Equipment

viscous damper conditions for engines (if applicable). The resultant compressor or engine cylinder excitation is applied to the train during the forced response stress analysis. With trains involving motors, slip and/or VFD frequencies are normally evaluated in addition to mechanical orders of running speed. During the forced response calculations, stress concentration factors (SCFs) are developed to account for keyways, major diameter changes, fillet radii, etc. based on shaft geometry provided by the equipment manufacturers and/or experience with similar machines. The resultant intensified stress and/or dynamic torques developed at each station in the model are then compared against allowable values (provided by the OEM or independently developed) to determine acceptability. A torsional analyst will typically evaluate the provided train configuration and make recommendations for modifications, if necessary, to shift the calculated torsional critical speeds or otherwise improve the damage tolerance of the shafting.

Torsional Mass Elastic Models Torsional models are normally prepared by lumping major system masses (e.g., motor cores, piston/rod inertias, flywheels, etc.) and connecting these with torsional stiffness values to represent the shafting and couplings. It is important to maintain sufficient fidelity in the model to accurately reflect the dynamically important modes. It is also prudent to lump the model in such a way as to maintain constant diameters within each interconnecting stiffness value when practical, thereby limiting the effects of stress concentration to the parts of the shafting that will experience it. This latter practice avoids overly conservative forced response stress and cumulative fatigue results. For typical solid shafting, the stiffness of the shaft may be estimated by using the following formula: K ¼ JG=L ¼ π∗ G∗ D4 =ð32∗ LÞ Likewise, the following formula may be used to estimate the polar mass moment of inertia for solid cylinders or discs: Ip ¼ π=32∗ ρ∗ L∗ D4 In many cases, torsional stiffness and inertia information is available from the OEM involved (EG: coupling stiffness, hub inertias, motor core inertia, reciprocating throw inertias, etc.). In most cases, these values are not considered controversial. Although calculating torsional stiffness values for the webs of reciprocating machines can be challenging for some geometries, several resources exist to provide guidance in mass elastic model generation, including the book by W. Ker Wilson entitled “Practical Solution of Torsional Vibration Problems” [8]. If necessary, solid models and/or FEA calculations may be utilized to confirm or estimate these values.

Reciprocating Compressors Chapter

5

227

It should be noted that the torsional stiffness and damping properties of elastomeric couplings are nonlinear, and require special consideration since they typically become the controlling stiffness for (at least) the first mode when installed in a train. For systems with gearboxes, the effects of the speed ratio(s) must be taken into account in order to accurately calculate the critical speeds. In practice, one end of the model becomes a reference speed, and the inertia and stiffness values of the remaining portions of the model are referenced with respect to (gear ratio) [8]. Input torques and interpretation of calculated torque/stress must also take speed ratio into account. Another issue that arises with gearboxes is how to represent the gear mesh stiffness. For most industrial gearboxes, with a fixed ratio defined by a mechanical connection between pinon and bull gears, the gear mesh stiffness is sufficiently high that the first several modes are not usually affected by this parameter. Although the gear mesh stiffness can be estimated, the controlling springs in typical systems are usually located in the couplings and driver/driven shaft ends. For variable speed devices (e.g., torque converters, planetary gear arrangements, etc.) a more in depth analysis is needed to accurately determine the torsional effects on the attached system. The following Fig. 5.41 provides a table and representative graphics describing a typical torsional mass elastic model. A common issue that arises when preparing torsional models for systems with interference fit couplings is how to appropriately represent the torsional stiffness contribution of the hub to shaft interference. The most common industry recognized approach to dealing with this issue is referred to as the “one-third shaft penetration rule.” Fig. 5.42 provides an illustration of how this approach is applied. Each interference fit length is divided into a section 1/3 of the total length, and another 2/3 of the total length. For the section 1/3 of the overall length, the shaft is assumed to be free to twist (unattached to the hub). In the remaining section (2/3 of the overall length), the shafting is assumed to be fully bonded to the shaft. In some cases, preparing an FEA of a torsional model may be advantageous. One example of this would be when significant localized stress concentrations exist, such as those located within a reciprocating compressor shaft with unusual web construction. Fig. 5.43 provides a comparison between FEA and lumped parameter frequency prediction results for a typical reciprocating compressor shaft. In this instance, reasonable calculated frequency agreement was found (within about 4%) for the subject mode. For most typical configurations, a lumped parameter model is more cost effective and sufficiently accurate. Peterson’s Stress Concentration Factors [9] provides an excellent resource for estimating the SCF for common shafting geometries. Fig. 5.44 illustrates methods for estimating the stress concentration effects for shafts with shoulder fillets at diameter changes, and for shafts containing keyways. The charts indicate that shoulder fillets can roughly double the dynamic stress developed in shafting for some geometries, and that keyways have the potential to intensify stress by a factor of 4. These findings demonstrate the importance, from

Element stiffness in-lb* 106

Station inertia lb-in-s2

Outboard end Compressor Compressor Compressor Compressor Compressor Compressor Compressor Compressor Compressor Compressor Coupling Motor DE Motor Motor Motor Motor Motor Motor Motor core Motor core Motor core Motor core Motor Motor Motor Motor Motor Motor NDE

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

5.000 5.550 5.000 5.550 5.000 5.550 5.000 5.550 5.000 5.550 4.500 – 6.500 6.500 7.870 7.076 7.870 9.444 10.000 10.000 10.000 10.000 10.000 10.000 9.444 7.870 7.076 7.870 –

93.1 124.1 112.2 124.3 112.7 112.7 124.1 112.2 124.1 85.7 203.4 61.5 421.8 271.1 453.1 276.1 924.4 1568.9 1580.6 1035.1 1035.1 1035.1 1035.1 1580.6 1568.9 924.4 276.1 513.3 –

0.9 1.9 5.0 5.0 1.3 1.3 1.3 5.0 5.0 1.9 1.1 23.3 23.7 0.4 0.8 1.0 0.7 1.0 1.9 71.7 141.2 141.2 141.2 71.7 1.9 1.0 0.7 0.9 0.5

Coupling

Reciprocating compressor

Motor

20 Shaft radius (in)

Element stress in

II Types of Equipment

Station #

228 SECTION

Description

10 1

2

3

4

5

6

7

8

9

10

1112

18 19 20 1314 15 16 17

21

22

23

0 –10 –20

0

40

FIG. 5.41 Representative mass elastic model table and graphics. (Courtesy of SwRI.)

80

120 Axial location (in)

160

24 25 26 27 28 29

Reciprocating Compressors Chapter L1

Drive shaft

229

L2

(1/3)L1 (2/3)L1

5

(2/3)L2

(1/3)L2

Driven shaft

FIG. 5.42 1/3 Penetration rule for torsional modeling of interference fit hubs. (Courtesy of SwRI.)

torsional damage tolerance standpoint, of avoiding keyways and utilizing generous fillet radii when possible. Torsional Stiffening Effect of Induction Motor Webs Longitudinal webs (colloquially known as spider bars) are often placed at the mid-span of induction motor shafts and are primarily used to support the rotor core laminations while allowing sufficient space for cooling airflow. When subjected to a torque, the webs experience a loading configuration that includes bending and torsion while the base shaft experiences pure torsion. These effects complicate the calculation of the torsional stiffness of the base shafting, which tends to increase for such configurations. This discussion seeks to provide a simplified practical approach for dealing with this issue, as outlined in Ref. [10] (Fig. 5.45). Historically, various approaches have been used to account for the base shaft stiffening effect of motor core webs in torsional rotordynamics, including geometrically based approximation methods and FEA techniques. In practical experience, FEA approaches have been found to produce meaningful results, but tend to be time consuming compared to other methods. Nestorides [11] presents various methods to account for the increase in torsional stiffness due to webs rigidly attached to shafts, including a technique described as the Griffith and Taylor method [12, 13]. This method requires dividing up the webbed cross-section and performing various geometric calculations to arrive at an effective second polar moment of area. Nestorides [11] provides a detailed description of this method, including the sectioning techniques and the tabular data necessary for the calculation. This method, although found to produce results satisfactorily comparable to FEA results [10], can still be considerably time consuming to execute. Eq. (5.12) provides a simplified method for estimating the torsional stiffness of a webbed shaft section, based on the Griffith and Taylor method. While the Griffith and Taylor method presented by Nestorides [11] requires several steps, the following approach provides a single equation to estimate the torsional stiffness of a webbed shaft. Several assumptions are necessary to simplify the process. This interpretation of the Griffith and Taylor method assumes that the

230 SECTION II Types of Equipment

FIG. 5.43 FEA vs. lumped parameter model frequency comparison. (Based on SwRI analysis, graphics produced using ANSYS and MATLAB codes.)

Reciprocating Compressors Chapter 5

231

FIG. 5.44 Stress concentration factors for torsion of shafts with shoulder fillet and keyway. (Data from Matthews and Hooke 1971, chart 3.12 from Peterson’s Stress Concentration Factors.)

232 SECTION

II Types of Equipment

FIG. 5.45 Typical geometry of a webbed motor rotor and cross-section under the windings. (Courtesy of SwRI.)

rounding off of the outer webs is negligible and that the λ parameter described by Nestorides has a value of unity. Note that this equation also assumes uniform, identical webs that are fully attached to the base shaft along the entire axial length. In addition, this equation assumes that the material properties, specifically the shear modulus, G, are identical for both the shaft and web materials. Furthermore, this simplified approach does not account for welding effects or any potential stiffening from the laminations themselves. "   # 2 3 G π 4 1 N b d + Nb2 d2 + 2 bh  (5.12) Ka ¼ 2 l 32 16 2h

Quantity

Typical SI Units

Typical US Customary Units

d ¼ diameter of base shaft h ¼ height of radial web above base shaft b ¼ web thickness l ¼ axial length of webbed shaft N ¼ number of radial webs G ¼ shear modulus of shaft and webs Ka ¼ approximate webbed shaft torsional stiffness

m m m m – N/m2 N-m/rad

in in in in – psi lbf-in/rad

Note that Eq. (5.12) presented above is different from the torsional stiffening equation provided in API 684, Second Edition [6], which has in some instances provided unreliable results. Fig. 5.46 presents a comparison of FEA and Griffith-Taylor method results for a shaft with six webs of varying thickness [10] and illustrates generally good agreement in the calculated range. It should also be noted that for designs utilizing web geometries with a web height to thickness ratio of greater than 4, an appropriate torsional spring should be added to the model (between the motor core and shaft) in order to properly account for the effects of significant web flexibility on the calculated torsional critical speeds.

Reciprocating Compressors Chapter

5

233

3 FEA

N = 6, D/d = 2.5

Griffith-Taylor

d = 8 in (203.2 mm)

Torsional stiffness ratio, Ka/Kb (–)

2.5

2

1.5

1

0.5

0

0

0.05

0.1

0.15

0.2 b/d (–)

0.25

0.3

0.35

0.4

FIG. 5.46 Torsional stiffness data for six-webbed shaft with varying web thickness. (From Chris D. Kulhanek, Stephen M. James, Justin R. Hollingsworth, Stiffening Effect of Motor Core Webs for Torsional Rotordynamics, presented at ASME Turbo Expo 2012, Copenhagen, Denmark, June 1115, 2012, paper GT2012-69967.)

Steady-State Analysis During a steady-state torsional analysis the critical speeds and mode shapes are produced. Fig. 5.47 provides a typical torsional mode shape superimposed on a graphic of the shafting involved. In this figure, the stations used to represent the motor inertia values do not participate significantly in the mode. However, the compressor inertias do participate significantly. This would be an important consideration when determining the likelihood of the mode to be excited by energy prevalent in the system. In this case, compressor generated excitation would be expected to couple readily, while motor excitation would have less of an impact on this mode. The figure also indicates that the interconnecting stiffness values associated with the coupling and shaft ends play a large part in determining the predicted frequency. Once the predicted critical speeds are calculated, they are superimposed on an interference (or Campbell) diagram, such as the one depicted in Fig. 5.48. This diagram is useful for determining the intersection points between likely excitation orders and the predicted critical speeds, and documenting the

234 SECTION

II Types of Equipment

1.00

Normalized amplitude

0.75 0.50 0.25 0.00 Compressor

–0.25

Coupling

Motor

–0.50 –0.75 –1.00

Mode 1: 5000 cpm (83.33 Hz) 0

50

100 150 Length (in)

200

250

FIG. 5.47 Typical torsional mode shape. (Courtesy of SwRI.)

FIG. 5.48 Representative interference diagram (whole-order excitation). (Courtesy of SwRI.)

proximity of these intersection points to the operating speed range. It is customary to include a range around the operating speeds to account for uncertainty in the provided input data. A range of 10%–20% is typical in the industry. In the example diagram of Fig. 5.48, intersections between the first critical speed and the fifth through seventh orders of excitation is noted. The likelihood

Reciprocating Compressors Chapter

5

235

Interference Diagram for Inter-Harmonics from VFD for LPMR (MCL1605) Train 250 +20%

Speed Range

–20%

f4 = 212.3 200

Frequency [Hz]

f3 = 178.77 f2 = 169.11 150

Fn – 12Fm 12Fn – 12Fm 12Fn – 24Fm 24Fn – 24Fm 24Fn – 36Fm 36Fn – 48Fm 48Fn – 60Fm

100

50

f1 = 16.05 0

0

500

1000

1500

2000

2500

3000

3500

4000

Speed [rpm]

FIG. 5.49 Interference diagram with VFD generated harmonics. (Courtesy of SwRI.)

of exciting this particular mode would need to be evaluated by studying the mode shape, and considering the excitation energy generated at key locations in the train. Whole order excitation, as illustrated in Fig. 5.48 can be generated by various mechanical sources (e.g., misalignment at 2 , or two-stroke reciprocating machinery). It should be noted that four-stroke reciprocating machines can also generate half orders (1.5 , 2.5 , etc.) in addition to whole order excitation (1 , 2, etc.). For machines involving a gear ratio, either separate interference diagrams should be prepared for each shaft speed, or in some cases the multiple shaft speeds can be effectively plotted on the same diagram. At any rate, both speed ranges should always be considered. Fig. 5.49 illustrates another type of interference diagram, which is used to document potential excitation of system critical speeds by VFD harmonics, which can correspond to either whole orders or inter-harmonic frequencies. From a practical point of view, it is difficult to avoid all potential excitation mechanisms by utilizing frequency avoidance for most modern industrial machines. This is particularly true for trains with a wide speed range, gear ratios, VFDs, or reciprocating machinery (which tend to produce multiple strong harmonics). In these cases, a forced response analysis is usually necessary to assess the potential impacts of running on or near an intersection between a critical speed and excitation energy.

Allowable Stress Methodology for Torsional Systems Common industry recommended approaches for assigning an allowable torsional fatigue stress level for reliable long-term operation are generally based

236 SECTION

II Types of Equipment

on a standardized ultimate tensile strength (UTS) reduction methods. In the absence of validated fatigue data for the shaft material in question, an effective endurance limit is generated. On approach is based on Military Standard 167 [14], which involves dividing the UTS by 25 to arrive at an endurance limit. This approach, although simple to apply, can be quite conservative. Another common method starts with UTS and applies a series of reduction factors. The suggested values for these factors represent a conservative interpretation of the method described by ASME B106.1M [15].

Equation (5.13). Endurance Limit Based on an Ultimate Strength Reduction Method Tensile to shear, Fshear : 0:577 Endurance ratio, Fendurance : 0:5 Size,Fsize : 0:7 Surface finish, Fsurface : Calculated ðgenerally 0:7  0:8Þ Reliability,Freliability : 0:8 High Cycle Fatigue stress design, FHCF Design : 0:667 Mean stress, Fmean ¼ 1  mean shear stress=ð0:577∗ UTSÞ Stress concentration factor, SCF varies, typically 1:2  3 for most shafts Shear End:Limit ¼ Fshear ∗ Fendurance ∗ Fsize ∗ Fsurface ∗ FHCF design ∗ Fmean ∗ Freliability ∗ UTS∗ ð1=SCFÞ

To put this into perspective, and neglecting the mean stress factor and SCF, a shaft with a UTS value of 690 MPa could have a target allowable stress level of only about 5%–6% of this value, with typical reduction factors. The shear factor, Fshear, accounts for the relationship of shear endurance limit to bending endurance; 0.577 is the factor given by the shear energy criterion. The maximum shear failure theory yields 0.5 for this factor, and is, thus, somewhat more conservative than the shear energy factor, but the values from these approaches are similar. Endurance ratio factor, Fendurance, reflects the observed factor between ultimate tensile stress and endurance stress for a large body of materials. The factor actually varies between 0.4 and 0.6 for most steels, with 0.5 being a reasonable average factor. If specific fatigue data for the material in question is available, then this can be used to replace the endurance ratio. The size factor, Fsize, represents the fact that tensile fatigue tests are normally performed on relatively small specimens. For a variety of postulated reasons (stress gradient and critical depth), data indicate that the endurance limit of practically sized components is lower, based on Heywood [16]. The value of 0.7 is a conservative factor to account for this.

Reciprocating Compressors Chapter

5

237

FIG. 5.50 Surface finish factor treatment. (From Shigley, Mischke, Mechanical Engineering Design, Table 7.3, 1989, p. 274.)

The reliability factor, Freliability, accounts for the likely variance in endurance stress data obtained by the test. The factor of 0.8, based on ASME B106.1M [15], will cover approximately 99% of the data, based on an 8% standard deviation and a normal distribution. The surface finish factor reflects the fact that most test specimens are polished and that the shaft will not have as fine a surface as the test specimen. Typical size factors range from about 0.7 to 0.8 for reciprocating machines, based upon the Shigley equation for machined surfaces, which is a function of material UTS as depicted in Fig. 5.50, based on “Mechanical and Engineering Design,” Ref. [17]. The HCF (High Cycle Fatigue) stress design factor, Fdesign, reflects the desire to design for survival of high cycle fatigue by a margin, rather than to design for failure. This factor also reflects the limited data for very high cycle fatigue, out to 106 and beyond. ASME has published fatigue data out to several orders of magnitude beyond the traditional “infinite life” value of 106 cycles. This data suggest that the traditional published endurance limits must be reduced by an additional 30%–40% to get to life limits suitable for machinery service. It should be noted that the HCF stress design knockdown factor of 0.667 corresponds to a safety factor of 1.5. The HCF design knockdown factor, FHCF design, is the inverse of the safety factor. The methodology presented here considers an HCF stress knockdown factor of 0.5 (safety factor of 2.0) to represent infinite life.

238 SECTION

II Types of Equipment

Mean stress factor, Fmean, accounts for the observation that the endurance stress decreases as mean stress increases. The modified Goodman formulation, summarized above, provides a working method to account for mean stress. The mean shear stress factor was calculated for each element from the mean torque and the stress diameter. Note that the endurance limit calculations for the forced response analysis include Fmean, since the analysis produces purely alternating stresses. For a transient analysis, Fmean, is not included for the endurance limit calculation because the stress predictions account for both mean and alternating stresses. To account for stress raisers in the shafting, SCFs are applied to each element. For example, SCF values for motor cores, keyways, and reciprocating machinery shaft webs are typically assumed to be in the range of 2.0–3.0. These values are considered reasonably conservative, based on experience with similar machines, and “Peterson’s Stress Concentration Factors” [9]. Note that the shear endurance limits for the forced response and transient analyses are different, due to the mean shear stress factor, Fmean, as discussed above. Torsional Damping The damping applied during the torsional analysis can have a significant effect on the calculated forced response stress results, or cumulative fatigue predictions for the transient events. The intent of this section is to give some brief practical guidance for typical industrial systems, based upon the common experience. When typical disc pack or metallic flexible element couplings is utilized, trains involving gearboxes, centrifugal compressors, and/or turbines are typically modeled with model amplification factors (Q values) in the range of 30–35. Trains with engine-driven reciprocating compressors normally involve a Q value in the range of 35–50. For electric motor-driven reciprocating compressors, Q values in the range of 50–70 are normally assumed. It is very important to note that the Q values quoted above assume typical disc pack or metallic flexible element couplings. If elastomeric couplings are utilized, the resulting Q values will be much lower, and highly dependent on the type of elastomers involved. Reciprocating engines also represent a special case in that they may include viscous dampers, which are usually represented as discrete damping components. Most torsional dampers in use on reciprocating machines contain a highly viscous silicone fluid, which tends to solidify when overheated, and must be maintained and/or replaced on a regular basis to ensure that adequate damping is available to limit stress levels when operating on a torsional critical speed. Similarly, many modern high power engines use dampers with tuned steel springs that incorporate forced oil, supplied by the crankshaft, as the damping fluid. These devices produce high damping levels and allow high heat capacity, but planned maintenance intervals must also be followed to ensure proper operation.

Reciprocating Compressors Chapter

5

239

Forced Response Analysis Torsional forced response analyses are used to document the stress and/or torque levels generated in the interconnecting stiffness elements in the model, as the unit runs at various operating conditions. For reciprocating engines or compressors, the torque generated in the swept volume of the cylinders must be determined and applied to the model as a forcing function. The resultant stress is evaluated on a station-by-station basis to determine acceptability using methods such as those outlined in the Allowable Stress Methodology section. Fig. 5.51 provides representative cylinder torque effort and frequency content information. The individual cylinder torque effort histories are utilized to generate the total applied torque plot, and determine the magnitude of the various orders involved based upon the phasing of the throws. In this particular case, significant 1 and 3  are generated. Fig. 5.52 presents the resultant dynamic torque and stress distribution within the train. The dynamic torque levels are typically compared to the OEM provided coupling and/or shaft torque ratings. The dynamic stress levels for the shafting are typically plotted against the allowable stress limits described in the Allowable Stress Methodology section for design factors corresponding to 1.5 and 2.0. For the methodology presented here, design factors equal to 2.0 or above are considered to correspond to infinite life. Changes to operating conditions, speeds, or mass elastic characteristics (to shift critical speeds) are generally recommended in cases where a design factor below 1.5 has been calculated, in order to improve long-term reliability. For reciprocating machines, it is frequently useful to conduct the stress calculations over a range of speeds in order to document the effects of operating on the flank of a critical speed. Fig. 5.53 provides a typical graphic for displaying this type of information in a waterfall format. Transient Torsional Analysis Transient torsional analyses are usually conducted for trains involving electric motors. The transient events most often studied include start-up, two-phase short circuit, and three-phase short circuit. It should also be noted that generator/motor synchronization events may also be studied, along with the varying effects of soft start strategies (as opposed to across-the-line or line-to-line starting events). A significant difference exists between synchronous and induction motors during the start-up event. Generally speaking, a synchronous motor produces more significant dynamic torque during start-up, and involves a forcing function which varies with respect to speed (generally twice the line frequency at zero speed, dropping to 0 Hz at synchronization speed). However, in most cases the resultant shaft stress levels are tolerable if the total driven inertia (compressor, flywheel, coupling hubs, etc.) is less than the motor inertia (see Ref. [5], section 3.4.2.1). Induction motors tend to generate forcing torques at various frequencies and relatively lower amplitudes during the start-up event.

240 SECTION

3

Total applied torque

× 105

1 0

Compressor cylinder pressure

800

Cylinder

600

1231.1916

500

92.0532

2

6 in-lbs

Psig

700

400 300 50

100

2.5

150

200

250

2

350

1 0.5 0

–0.5 0

150

50

100

150 200 250 Degrees of rotation

300

350

400

200

250

300

350

400

Total applied torque

x 104

4 2 0 0

400

1224.065

1.5 In-lbs

300

100

5

10

15 Orders

20

25

30

5

10

15 Orders

20

25

30

Cylinder torque

3

x 10

Percent total torque

0

50

30 20 10 0 0

FIG. 5.51 Representative applied torque magnitude and order content for trains with reciprocating machines. (From SwRI results developed using MATLAB.)

II Types of Equipment

in-lbs

2873.0445 2

Reciprocating Compressors Chapter

15 × 10

4

Dynamic torque

12,000

Coupling

psi (pk)

in-lbs (pk)

10

5

241

Maximum allowable stress and computed combined stress

10,000

Design factor = 1.5

8000

Design factor = 2.0

Design factor = 1.5 Design factor = 2.0

6000 4000

5

2000 0 0

5

10

15

20

0 0

25

5

10

15

20

25

Field

Field

Compressor Motor Dynamic torque distribution

Compressor Motor Dynamic stress distribution

FIG. 5.52 Typical dynamic torque and stress distribution results for trains with reciprocating machines. (From SwRI results developed using MATLAB.) Stress order content

× 104 3

Fifth order excites first critical speed

psi (pk)

2.5 2 1.5 1

1100

Separation margin

1000

0.5 900

0 0

Running speed 800

5

10

15

20

700 25

30

600

loadstep/rpm

Shaft stress order content FIG. 5.53 Speed sweep analysis stress results for reciprocating compressor train. (From SwRI results developed using MATLAB.)

The short circuit events usually involve relatively strong torque excursions over a brief period of time, with most cumulative fatigue damage occurring in the first few cycles. Various critical speeds can be excited during the transient events, which can amplify the forced response stress levels occurring as the critical speeds are traversed. The transient events are typically simulated based on equations provided by the motor OEM. The torque is represented as a function of time and is defined with time steps small enough to adequately represent the frequencies involved, but large enough not to exceed the variable arrays used by the analysis codes.

242 SECTION

II Types of Equipment

Transient torsional stress calculations are accomplished by first translating each of the torque values for the section in question to stress by normal strength and material relationships, and taking the outside and inside diameters into account. A stress diameter is determined that corresponds to the smallest diameter in that particular element section. The stress variation is clearly complex in the resulting time histories. Its application to fatigue analysis is accomplished using Miner’s linear damage rule, as implemented in the “rain flow” cycle counting algorithm. This algorithm extracts from the complex stress variation, a count of the number of cycles, as a function of stress range. For each of these stress ranges, the S-N diagram of the material in question is queried to establish the number of cycles to failure at this stress range. The fractional damage is calculated as the ratio of actual number of cycles at this stress range divided by the number of cycles to fail at the stress range. Cumulative damage at each stress range for the entire cycle is totaled to give cumulative damage per event. The allowed number of transients is the inverse of this damage per event. During the transient torsional analysis, system integrity is investigated using an ultimate strength reduction approach, as described previously in the Allowable Stress Methodology section. This approach relates stress limits for high cycle fatigue to the UTS. In addition, the maximum shear strength and low cycle fatigue limit are necessary and calculated for the transient analysis. The maximum shear strength for the transient analysis is provided in Eq. (5.14). A factor of 0.9 is applied to the maximum shear strength to arrive at the low cycle fatigue limit (103 cycles). Note that the safety factor (assumption) is roughly equivalent to the term design factor (result). A safety factor of 1.1 is assumed for the max shear stress and low cycle fatigue limit, while a 1.5 safety factor is used in the high cycle fatigue limit (106 cycles). It is noted that the larger safety factor of 1.5 is applied to the high cycle fatigue limit because of the many uncertainties associated with the high cycle fatigue calculations. The assumed safety factor is reduced to 1.1 for the max shear stress and low cycle fatigue limit because less uncertainty exists regarding the mean loading and the allowable stress values at zero to low cycles. The low and high cycle stress limits are calculated for each element in the torsional model and printed on each relevant stress plot. These limits allow for the construction of an S-N diagram, constant at the low cycle limit for 1 to 1  103 cycles, constant from 1  106 cycles to an infinite number of cycles,

Equation (5.14) Max Shear Strength Equation Derived From the Ultimate Tensile Strength (Source: SwRI) Max Shear Stress ¼

UTS  Tensile to Shear Factor Safety Factor  SCF

Reciprocating Compressors Chapter

5

243

Torsional Vibration Analysis Demonstration Case - Synchronous Motor 2 Phase Short Shaft section 8 Amplification factor = 30 Effectiv radius = 7.35 in Inner radius = 0.00 Length = 80.00 in Sped ratio = 1.0 Shear modulus = 1.18 e+7 Psi

16,000

Maximum shear stress Low cycle failure stress shear endurance limit

Stress (psi)

12,000 Predicted number of tolerable events: 491

8000 4000 0 –4000 –8000

–12,000 –16,000

0

0.5

1

1.5 2 Time (s)

2.5

3

3.5

FIG. 5.54 Representative transient torsional results—two-phase short circuit. (From SwRI results with University of Virginia ROMAC code TORTRAN.)

and sloping linearly on a log-log plot between the 1  103 and 1  106 cycle points. While stress limits are used to evaluate the shafting elements, the predicted torque levels are typically used to evaluate the couplings. Typical output from this type of analysis includes calculated dynamic torque and stress plots for key portions of the trains, a summary of shaft stress results for the transient analysis, and an evaluation of dynamic torque developed at the couplings during the transients. Fig. 5.54 provides a representative dynamic stress history for a synchronous motor two-phase short circuit event, along with the associated cumulative fatigue results.

Resolution of Typical Torsional Problems Many torsional issues can be resolved through the practice of frequency avoidance, which generally involves tuning the predicted critical speeds such that an adequate separation margin is maintained from excitation energy present in a system. This approach is most useful for fixed-speed applications, but may not be fully effective for some machines operating over a wide speed range. In existing reciprocating systems, the frequency shifts are most often accomplished by modifying system inertias or interconnecting stiffnesses. Examples of inertia values that can be most readily adjusted in a reciprocating system would include external flywheel inertia, or internal flywheels placed between compressor throws. A common torsional stiffness change would involve a

244 SECTION

II Types of Equipment

different coupling. Two types of couplings widely used in reciprocating systems are of the flexible disc and elastomeric varieties. These couplings both have advantages and disadvantages and can be exchanged to produce a significant shift in torsional rotordynamics. Similarly, couplings within a series of a particular type may be exchanged for a less dramatic shift in the critical speeds. Flexible disc couplings are very commonly specified in reciprocating applications. These couplings can provide a low cost alternative in many situations, generally require little maintenance, and are available for a wide variety of torque and power ranges. Elastomeric couplings can be useful when a low torsional stiffness is required, for adding damping to a system, or allowing operation over a wide operating speed range with no lock-out (exclusion) speeds. However, elastomeric couplings generally require more maintenance than comparable flexible disc designs and may not be available for some high torque applications (see Ref. [5], section 3.1.2 for a more detailed discussion of this subject). In cases where the machinery must transition through or run on a critical speed for some operating conditions, it may be useful to increase the damage tolerance of the shafting. Some common systems that might require this approach would be a synchronous motor driving a reciprocating compressor during a start-up condition, or reciprocating compressors utilizing disc-pack couplings over a wide speed range. One common method for increasing the damage tolerance in new machinery designs is to increase the shaft diameter. This change has the additional effect of increasing the torsional stiffness of the shafting, which may be useful as well. Another common approach in new machines is to specify a shaft material with an increased UTS. Many equipment manufacturers offer an alternate shaft material for this purpose, which can greatly increase the ability of the machines involved to tolerate dynamic torque. Yet another approach is to decrease stress concentration in the shafting an associated rotating elements by removing keyways, increasing fillet radii, etc. In some cases involving reciprocating compressors, loading strategies may also be modified to reduce shaft stress to acceptable levels. In general, asymmetric loading strategies (some cylinders loaded, others unloaded) tend to produce the highest dynamic torques. A common misconception is that the most loaded (highest power) case would result in the most cumulative damage. In actuality, the total stress level experienced by a shaft is a combination of mean stress (which generally does increase with power setting) and dynamic stress (which can be significantly influenced by excitation of the torsional modes). It is not unusual for the highest composite stress level to occur at a load step that produces less power than the full-load case. As such, it may be possible to utilize a more symmetric loading strategy to achieve a particular operating condition, while limiting the shaft stress to a tolerable level. VFDs may also introduce significant alternating torques in a system that can result in excitation of a critical speed. The design and operation of these drives is considered beyond the scope of this chapter. However, in some cases it may be

Reciprocating Compressors Chapter

5

245

possible to tune the VFD in the field to avoid the generation of certain alternating torques (see Ref. [18], appendix c6 for further information). As a last resort, it may be necessary to avoid operation at some conditions (often referred to as “lock-out” speeds). Field data can be very helpful for identifying the exact speed ranges and conditions to avoid. In some lightly damped systems, it may only be necessary to change the operating speed slightly to avoid the resonance causing the high shaft stress levels. This approach can potentially allow for larger allowable speed ranges (smaller “lock-out” ranges), as less uncertainty would be involved than if the recommended speeds were based on a torsional analysis alone.

Testing Methods for Torsional Systems Field testing of torsional systems can involve significant challenges. Typical package instrumentation is focused on the monitoring of lateral vibration probes, and most often does not include provisions for monitoring torsional oscillation. An exception to this general statement is when a torsional mode involves significant participation in gearbox shafting. Due to the physical nature of the gear mesh, a mechanism is provided for torsional vibration to couple with lateral vibration. As a result, torsional modes can sometimes be monitored in the lateral probe signals from a gearbox. This section provides a brief summary of typical testing methods and instrumentation used for documenting torsional behavior. Fig. 5.55 illustrates typical field data that indicates the excitation of a torsional critical speed. The plot clearly shows resonant behavior as fifth- and sixth-order excitation energy from a reciprocating compressor transitions through the appropriate frequency. A waterfall plot can be very useful for estimating separation margins between excitation of critical speeds by various orders and running speeds, and also for estimating the damping of the modes in question. In this particular case, a large third-order component is also present, which is primarily a function of phasing within the reciprocating compressor involved and not related to a resonance. Historically, one reason that torsional vibration measurements were not prevalent involved the relative lack of durability or difficult installation requirements of typically available instrumentation. The subsequent paragraphs provide some further information about the most common types of torsional instrumentation and effective torsional data capture techniques. One predominant type of instrumentation for gathering torsional data, used widely in the industry, is a strain gage telemetry system, as shown in Fig. 5.56. The system consists of the strain gages on the shafting, a rotating collar attached to the coupling hub or spacer, and a stationary antenna/receiver. This system is intended for short-term use only, and each collar has an operating speed limitation due to centrifugal force. The data records strain levels, which can be directly related to shaft stress, and evaluated for acceptability.

246 SECTION

II Types of Equipment

Torsional field testing: Representative field data N1 = first natural frquency = 4620 rpm = 77 Hz During startup, 5× and 6× energy excite the first critical speed, even though 3× is the largest energy order 3× 1× 2× 980 rpm

4×

5× 6× 7× 8×

5× Excites NC1 (924 rpm)

6× Excites NC1 (770 rpm)

0 rpm 0

PWR SP A

200 Hz

FIG. 5.55 Representative torsional field data. (Courtesy of SwRI.)

FIG. 5.56 Strain gage telemetry system components. (From SwRI photograph of Wireless Data Corporation system.)

A torsiograph may be used to measure oscillations in rotational velocity. This type of test is normally performed on the outboard ends of the equipment due to shaft mounting requirements (a threaded pilot hole and/or stub shaft must be used). The data are most useful for determining torsional critical speeds. The instrument is best suited to short-term measurements. Similar data can be obtained with an encoder, although the signal must be appropriately processed,

Reciprocating Compressors Chapter

5

247

FIG. 5.57 Torsional testing with a torsiograph. (Courtesy of SwRI.)

and care must be taken to ensure that the bearings in the encoder are capable of prolonged operation at the intended speeds. Fig. 5.57 provides a photograph of a torsiograph being used to gather torsional data on an industrial electric motor. Instrumented couplings provide many of the benefits of a strain gage telemetry system, in a more robust package. These couplings are relatively expensive and are normally installed to allow for measurement of static torque, which can be used to estimate unit load in a real-time manner. However, a side benefit of many systems is to allow access to a dynamic strain signal, which can be used to monitor torsional critical speeds, and infer shaft stress levels in the remainder of the system provided a torsional forced response analysis is available. Typical instrumented coupling types are illustrated in Fig. 5.58. Laser rotational vibrometers are becoming more popular in the industry, as more options become available, and the instrument cost decreases. The most common types in use for torsional testing are of the twin beam variety, one of which is illustrated in Fig. 5.59. Over time these devices have become more user friendly, allowing use at further distances from the shaft. The units are also very portable, which is useful when multiple test locations are involved. However, these devices are still fairly sensitive to orthogonal placement of the beams to the shafting, and the shaft surface condition. Regardless of the test instrumentation selected, several procedures must be carefully considered to ensure that useful torsional data are obtained. One of the most important items to consider before a test is the mode shape of the critical speeds that are likely to be excited during the test. A preexisting torsional analysis can be very useful in this regard. A review of the mode shapes involved may

248 SECTION

II Types of Equipment

Instrumented coupling types Industrial coupling

Strain gages Bonded to spool

Rotary transformer

Strain gage type

Strain gages

Electronics

Driving machine

Stator support

Intermeshed pickup teeth

Driven machine

Phase shift type Monopole sensors

Flexible coupling

Cond. unit

Display unit

FIG. 5.58 Instrumented couplings—typical components. (From Bill Meier, Dave Edeson, Developments in Continuous Torque Monitoring Couplings, Presented at Ethylene Producers’ Conference, Orlando, FL.)

Laser rotational vibrometers fo

Mirror fo

Reference beam

fo + fb + fd

Beam splitter

Test beam

Beam splitter

Bragg cell fo + fb

Basic Operating Principle

fo + f b + f d

Laser fo

Photo detector

Beam dplitter

fo + fb

Target

Representative twin beam portable unit

FIG. 5.59 Laser rotational vibrometer operating principle. (Sources: Basic operating principle (http://en.wikipedia.org/wiki/Laser_Doppler_vibrometer) and Representative twin beam portable unit (http://www.polytec.com/us/products/vibration-sensors/special-application-vibrometers/rlv5500-rotational-laser-vibrometer/.)

Reciprocating Compressors Chapter

5

249

indicate that some test locations are very sensitive to the modes of interest, while other potential test locations are not likely to be sufficiently sensitive. Another common issue involves recording start-up or shutdown data with sufficient fidelity to adequately capture the transient events. Historically, it has been useful to record the signal with digital tape recorders (usually sensitive to 20 kHz or so), or some other means, which allows for quickly playing the signal back to the computerized DAS (Data Acquisition System) at several sample rates. This allows for optimized data reduction, since it can be difficult at times to estimate the optimal sample rate in advance of observing the events. Such a procedure avoids a potential need to start and or stop the machine several times while the DAS is optimized. Another consideration for potentially increasing the shutdown time of a train (and thus increasing the amount of captured data for a given sample rate) is to reduce the load as much as possible on the driven equipment during the testing.

Reciprocating Compressor Lateral Rotordynamics Lateral rotordynamic studies describe the vibration behavior of a rotating shaft due to the various radial forces acting on the rotor. The lateral rotordynamics of reciprocating compressors are often not considered a priority, as compared to other machinery, but a few special considerations are noteworthy and discussed further. This text is not intended to be a tutorial on subject of lateral rotordynamics or vibration, but rather give insight to the unique considerations for reciprocating compressors. Furthermore, it is noted that lateral rotordynamics must be carefully considered for many other types of rotating machinery, including driving equipment, connecting equipment such as couplings and gearboxes, and other gas machinery, such as centrifugal compressors. A more indepth discussion of lateral rotordynamics for centrifugal compressors is provided in Chapter 3. This section will cover four main topics, including a discussion on lateral natural frequencies of a crankshaft internal to a compressor, lateral modes involving overhung mass, lateral implications of rigid couplings between drivers and crankshafts, and lateral-torsional coupling. In general, the lateral natural frequencies (or critical speeds) of a crankshaft within a reciprocating compressor (lateral mode shapes that involve amplitude internal to the compressor) are typically not a concern for two main reasons: (1) the crankshaft natural frequencies are high in frequency and therefore well above potential excitations, and (2) the damping from the crankshaft bearings helps dampen the rotor response. As described earlier in this chapter, the crankshaft includes multiple journal locations for the bearings, which are spaced axially in between or on either side of the crank throws. Fluid-film stiffness and damping forces exist at all journal-bearing locations. The relatively short axial spacing between multiple-bearing constraints results in crankshaft natural frequencies (internal to the compressor) that are relatively high in frequency, and

250 SECTION

II Types of Equipment

well above potential excitations. In summary, lateral natural frequencies (or critical speeds) of a crankshaft that involve mode shapes with amplitudes within a reciprocating compressor are not a typical concern. Lateral natural frequencies for reciprocating compressor crankshafts that involve an overhung mass, such a coupling hub or flywheel, can become important. Specifically, this is referring to a lateral natural frequency that involves a mode shape with large amplitude at the overhung mass. These particular modes are controlled by the shaft stiffness of the overhung portion, and the mass of the overhung component. These modes are potentially excited by imbalance at 1 , misalignment at 1  and 2, and also by the rod loads acting from the connecting rod at the two throws adjacent to the drive end, primarily at 1 and 2 . With relatively high amplification and little damping (mostly from the crankshaft-bearing oil films), operating on or very near an overhung lateral natural frequency (primarily at 1  and 2) can cause significant lateral vibration and potentially failed parts. In most cases, a simple hand calculation considering the crankshaft held at the two drive-end bearings, shaft, and overhung mass may be adequate to determine that the lateral natural frequency is separate from 1  and 2  running speed. Such a check (or a complete lateral analysis) should be performed for these types of designs during the design phase to determine the potential for any overhung modes to be excited in the running speed range. The lateral rotordynamics of a reciprocating compressor also become important when the crankshaft is rigidly connected to the driving equipment. When a rigid coupling is used, the lateral rotordynamic response of the crankshaft is coupled to the driver shaft. Therefore, the lateral rotordynamic model and analysis should (at a minimum) include both the driver and reciprocating compressor crankshaft through the first two throws and crank bearings at the drive end. Traditional calculation of natural frequencies, mode shapes, and imbalance response should be considered for the coupled driver and compressor crankshaft model. It is noted that most oil and gas reciprocating compressors today are driven through flexible couplings, where the lateral response of the driver and compressor is de-coupled, or independent of one another. Lateral and torsional coupling, from a vibration standpoint, only occurs when a kinematic constraint relates lateral motion to torsional motion. This can be seen in a geared system or gearbox where lateral motion of the bull or pinion can influence or become affected by the torsional motion through the connection at the gear mesh. Typically, lateral-torsional coupling is seen when a torsional mode of a machinery train is excited, resulting in measureable lateral vibration of the pinion. There is the possibility, although considered very rare, of a coupled lateral-torsional mode in a geared system. Further information on lateral-torsional coupling in rotordynamics is discussed in Childs [18]. In addition, lateral and torsional vibration coupling has been noted in reciprocating compressors. The literature shows that lateral and torsional vibration coupling can be problematic in reciprocating compressors. Stephens et al.

Reciprocating Compressors Chapter

5

251

Ref [19] discuss how moderate levels of torsional vibration can correspond to high levels of lateral vibration at the torsional resonant frequencies (and  1 harmonic) as a result of lateral-torsional coupling within a compressor. Compressors operating with torsional vibration are not rotating at a constant angular velocity, which is typically assumed for inertial rod load calculations. The nonconstant angular velocity creates additional inertia rod loads that primary respond at 1 harmonic from the torsional vibration harmonic. These rod loads couple into the main bearings and react on the frame. The nonconstant angular velocity also creates vertical forces at the torsional vibration harmonic and 2 harmonics that react in the guides. These shaking forces can be calculated with simple rigid body kinematics at nonconstant angular velocity. The frequency shift is due to cosine/sine multiplication at different frequencies (i.e., 1  and n ) in the acceleration terms of the force calculations. The forces caused by the torsional oscillation are proportional to the rotating inertia, and importantly, to the square of the torsional frequency. For high-speed reciprocating compressors with heavy pistons and moderate torsional vibration, these forces can be on the same order as the primary unbalanced inertial forces, which can be several thousand pounds of dynamic force. While lateral-torsional coupling can and does exist in many reciprocating compressors, elevated lateral vibration levels caused by this phenomenon often pass undiagnosed. In addition, lateral-torsional coupling in reciprocating compressors is typically not analyzed on a regular basis in the oil and gas industry, as opposed to other topics, such as torsional vibration or pulsation studies, which are very common. As these force calculations require the torsional response to be calculated and can only be used in a design approach three mechanical study, the absence of API methodology to address them is a current design gap in the industry.

References [1] These descriptions of diaphragm compressors were provided by PDC Machines, Inc. [2] NEMA MG1 21.36.1. [2a] ANSI/NEMA MG 1-2016 (NEMA MG1 21.36.2), Motors and Generators. https://www. nema.org/Standards/ComplimentaryDocuments/ANSI_NEMA%20MG%201-2016% 20CONTENTS%20and%20FOREWORD.pdf. [2b] API 618, Reciprocating Compressors for Petroleum, Chemical, and Gas IndustryServices, Fifth Edition, December 2007, API Publishing Services, 1220 L Street N.W., Washington, DC, 20005, United States. [3] GMRC, Guideline and Recommended Practice for Control of Torsional Vibrations in Direct Driven Separable Compressors, Section 3, GMRC, 2015. [4] G. Phillippi, Basic Thermodynamics of Reciprocating Compression, in: Turbo Symposium Tutorial, 45th Turbomachinery and 32nd Pump Symposia, Houston, Texas, September 12–15, 2016, 2016. [5] Gas Machinery Research Council (GMRC), Guideline and Recommended Practice for Control of Torsional Vibrations in Direct-Driven Separable Reciprocating Compressors, 2015.

252 SECTION

II Types of Equipment

[6] American Petroleum Institute (API), API 684 Ed. 2, API Standard Paragraphs Rotordynamic Tutorial: Lateral Critical Speeds, Unbalance Response, Stability, Train Torsionals, and Rotor Balancing, 2010. [7] Hinchliff, Bratek, A Discussion of the Loads Used to Rate Reciprocating Compressors, GMRC, 2014. [8] W. Ker Wilson, Practical Solution of Torsional Vibration Problems (With Examples from Marine, Electrical, Aeronautical, and Automobile Engineering Practice), third ed. Revised, vol. I, John Wiley & Sons Inc, New York, 1956. [9] W.D. Pilkey, Peterson’s Stress Concentration Factors, second ed., John Wiley and Sons, New York, 1997. [10] C.D. Kulhanek, S.M. James, J.R. Hollingsworth, Stiffening Effect of Motor Core Webs for Torsional Rotordynamics, 2012. Presented at ASME Turbo Expo 2012, Copenhagen, Denmark, June 11–15, 2012, paper GT2012-69967. [11] E.J. Nestorides, A Handbook on Torsional Vibration, Cambridge University Press, Cambridge, 1958. [12] Griffith, A.A., Taylor, G.I., Rep. Memor. Adv. Comm. Aero., Lond., nos. 333, 334 and 392. [13] G.W. Trayer, H.W. March, N.A.C.A. Rep. No. 334, 1944. [14] Department of the Navy, MIL-STD-167, Military Standard, Mechanical Vibrations of Shipboard Equipment, 1954. [15] ANSI/ASME B106.1M (Ed.), Design of Transmission Shafting, 1985. [16] R.B. Heywood, Designing Against Fatigue of Metals, Reinhold Publishing Corporation, New York, 1962. [17] Shigley, Mischke, Mechanical Engineering Design, Table 7.3, 1989, p. 274. [18] D. Childs, Turbomachinery Rotordynamics With Case Studies, 2013. [19] W. Wang, J. Braun, R. Chundi, R. Khan, IDC Technical Services; Tom Stephens and Fred Newman, Ariel Corporation, Linear and Torsional Vibration Study for An Engine Driven Compressor System, GMRC, 2013. [20] GMRC Guideline for High Speed Reciprocating Compressor Packages, 2013. appendix 3.1. [21] Hinchliff, Greenfield, Bratek, A Discussion of the Various Loads Used to Rate Reciprocating Compressors, GMRC, 2014. [22] H.W. Evans, J.L. Gallagher, Operation of Compressor Cylinders Without Cooling Water, Sinclair Oil and Gas Company, Tulsa, OK, 1955. technical paper, ASME conference.

Chapter 6

Screw Compressors Eugene “Buddy” Broerman*, Tim Manthey†, J€ urgen Wennemar‡ and Justin Hollingsworth* *

Southwest Research Institute, San Antonio, TX, United States, †Aerzen USA, Coatesville, PA, United States, ‡MAN Energy Solutions SE, Oberhausen, Germany

Two Types of Screw Compressors Screw compressors generally come in two basic designs: dry screw compressors or wet/oil-flooded screw compressors. Both compressor types consist of rotors that are closely mated. Dry screw compressors do not have any oil in-between the screws. A dry screw compressor consists of two rotors, a closely mating pair that are installed in a tight clearance cylindrical bore. Wet or oil-flooded screw compressors have oil in-between the screws. An oil-flooded screw compressor also has a pair of closely mating rotors that are installed in a tight clearance cylinder bore; however, oil-flooded screw compressors do not require clearances to be as tight as those that are required for the dry screw compressors. The subsequent sections of this chapter describe the various aspects (similarities and differences) of these two types of screw compressors.

Working Principle of Screw Compressors Screw compressors consist of two rotors in a common casing. Both rotors carry intermeshing helical lobes and rotate against each other with tight clearances between the rotors, and between the rotors and casing. During rotation, the lobes and casing form compression chambers that steadily decrease in volume as the rotors turn, changing cyclically from maximum volume to zero and back to maximum again. Thus, the working principle is similar to other positive displacement machines like reciprocating compressors. In an oil-flooded screw compressor, a slide valve is available for capacity control. This slide valve moves axially beneath the rotors and changes the effective rotor length, and also opens an internal recycle volume on the suction side of the compressor. This is explained in more detail in the sections further. Fig. 6.1 shows the rotors in different positions and with a varying working chamber volume. Compression Machinery for Oil and Gas. https://doi.org/10.1016/B978-0-12-814683-5.00006-7 © 2019 Elsevier Inc. All rights reserved.

253

254 SECTION

II Types of Equipment

FIG. 6.1 Compression progression.

Positive displacement machines do not transfer a certain speed-related amount of kinetic energy to the gas which is converted into pressure like centrifugal compressors but compress the gas “quasi-statically.” The enthalpy difference per mass unit of gas transferred into the gas can vary heavily for different gases. Therefore, the concept of head does not make sense for positive displacement machines and the use of an h-s diagram is not helpful. The working process consists of three phases and is best described in a pressure-volume diagram (Fig. 6.2). (Note that all pressures are absolute pressure unless specifically noted.) Suction phase: The working chamber is connected to the suction line via the inlet port in the casing. The size of the working chamber increases from zero to its maximum value and the chamber is filled with gas at the suction pressure, p1. Compression phase: The working chamber is closed off from the suction and discharge line. The working chamber decreases its size from the maximum value V1 to a defined value V2 that is defined by the following equation: V2 ¼ V1 =vi

(6.1)

with V1 and V2 as defined in the nomenclature section. vi is the built-in volume ratio.

FIG. 6.2 Best fit.

Screw Compressors Chapter

6

255

The built-in volume ratio vi is an important value for the design of screw compressors and is determined by the shape and size of the discharge opening. In oil-free screw compressors, the vi is a fixed value. In oil-flooded screw compressors, some models are available with an adjustable vi slide that is independent of the capacity slide valve; however, even for those without this adjustable vi feature, the vi is only truly realized when the capacity slide valve is in the maximum capacity position. When the slide valve moves toward the minimum capacity position, the effective vi is reduced due to the shorter effective rotor length. Due to the size reduction of the working chamber, the gas is compressed as it is moved by the screws, and the gas pressure and temperature increase. At the end of the compression phase the internal pressure p2i is reached. As described in the following equation, p2i is a function of p1, vi, and k. p2i ¼ p1  vi k

(6.2)

with k being the isentropic exponent of the gas. For ideal gases: k ¼ cp =cv

(6.3)

The ratio p2i/p1, also expressed as vik, is often called the “compression ratio.” It should be noted that the “built-in volume ratio” is a design property of a certain screw compressor, the “compression ratio” depends on the volume ratio and the kappa value of the gas, and the “pressure ratio” (p2/p1) also depends on the discharge line pressure. Therefore, when discussing the pressure ratio for screw compressors, it is important to distinguish whether the discussion is about the “pressure ratio” or the “compression ratio.” Discharge phase: When the compression chamber has reduced its size to the volume V2, the rotor lobes pass the outlet port in the casing and the chamber is connected to the discharge line. This is the beginning of the discharge phase. By further rotation of the rotors the working chamber reduces its size to zero and the gas is ejected into the discharge line at the discharge pressure p2. The screw compressor has no dead volume like a reciprocating compressor therefore no reexpansion of trapped gas happens. During all phases, a small amount of gas leaks across the rotor clearances from the discharge line into the compression chamber. Another leakage flow occurs during compression from the closed compression chamber to the trailing chambers with lower pressure and to the suction line. The leakage to the suction line acts like an internal bypass and reduces the inlet volume flow. This is known as slippage, which also has thermal effects on the compressor: the gas that “slips” backwards from higher pressure to lower pressure has already been heated by the compression process. As it is recompressed, the heat increases further. Thus, higher slippage leads to higher discharge temperature. Slippage increases as internal clearances increase, as pressure ratio increases,

256 SECTION

II Types of Equipment

and as gas molecular weight decreases. All compressor sizing and performance calculations take slippage into account; therefore, it is critical to have accurate gas data and clearance information in order to get accurate data with respect to volumetric efficiency and discharge temperature. Please note that the working process description given above is idealized because the real compression phase is not isentropic but polytropic. The work W needed for the compression is indicated by the area limited by the suction, compression, and discharge line and is mathematically given by. ð W ¼ Vdp (4) It must be noted that the beginning of the discharge is only determined by the position of the outlet ports in the casing and is independent from the actual pressure in the discharge line. Therefore, the internal pressure p2i may deviate from the discharge line pressure p2. This may occur if the suction or discharge line pressures change or if the isentropic exponent k of the gas changes.

Undercompression A case where the internal pressure p2i is lower than the discharge line pressure p2 is called undercompression (Fig. 6.3). In this case the compression chamber opens before the inner pressure p2i has reached the line pressure p2 and gas is flowing back rapidly from the discharge line into the compression chamber until the pressures have equalized. By further rotation of the rotors the volume is finally reduced to zero and the gas is expelled into the discharge line at the pressure p2. Fig. 6.3 shows that the area between the lines is larger than the ideal process in Fig. 6.2 thus indicating that for undercompression a larger compression work W is needed. This is equivalent to a lower efficiency and higher

FIG. 6.3 Undercompression.

Screw Compressors Chapter

6

257

discharge temperature. In addition, the backflow of gas into the compression chamber causes gas pulsations and noise in the chamber and the connected discharge line.

Overcompression The opposite case where the internal pressure p2i is higher than the discharge line pressure p2 is called overcompression (Fig. 6.4). Here the gas is compressed inside the closed compression chamber to a higher pressure than in the discharge line. When the rotor lobes have reached the outlet port edges the chamber opens and the gas expands rapidly into the discharge line until the pressures equalize at p2. By further rotation the chamber volume is reduced to zero and the gas is expelled into the discharge line at the pressure p2. The triangular pressure peak in Fig. 6.4 shows that overcompression also leads to a larger work than in Fig. 6.2 which indicates a higher power consumption and worse efficiency. Again the rapid equalization of gas pressures may cause gas pulsations and noise. The compression to a high internal pressure p2i also leads to high internal gas temperatures which are higher than the temperature in the discharge line. In extreme cases, this may cause damages due to internal overheating that may not be detected by the temperature sensors in the discharge line. In some cases, overcompression can lead to internal pressures that exceed the pressure rating of the machine, and can also cause very high loads on the radial and thrust bearings, leading to reduced bearing life and possibly even shaft damage. It should be noted that Figs. 6.2–6.4 are idealized cases. In reality, the compression phase is not isentropic and also the pressure equalization in case of undercompression or overcompression is not instantaneous but takes some time. Therefore, Fig. 6.5 shows an example of a more realistic simulation.

FIG. 6.4 Overcompression.

258 SECTION

II Types of Equipment

FIG. 6.5 p–v-Diagram simulated.

A certain mismatch of “built-in pressure ratio” and external pressure ratio is often unavoidable and does not cause a real disadvantage. Large deviations between both values, however, may have severe disadvantages such as a drop of efficiency, gas pulsations, and internal overheating in case of overcompression. Undercompression is sometimes unavoidable, especially in oil-flooded screw compressors where high-pressure ratios are possible; cases with an internal compression ratio of 10 and a pressure ratio of 25 have been known to operate without any issues. In any case a moderate undercompression is better than overcompression. The working phases are repeated for each lobe of the male rotor. Thus a compressor with four lobes on the male rotor performs four compression cycles during each rotation of the male rotor. The number of compression cycles per second is called pocket passing frequency (PPF). Gas is discharged into the discharge line discontinuously at the PPF. Table 6.1 shows examples of different compressor types and speed with their respective pocket passing frequencies. The PPF is much higher than the speed of reciprocating compressors which means that the discontinuous flow pulsations occur at higher frequencies. The discontinuous flow requires a careful discharge silencer design.

Comparison of Positive Displacement Machines (Screw Compressor, Reciprocating Compressor) Versus Centrifugal Compressors Screw compressors are positive displacement machines with purely rotary motion.

Screw Compressors Chapter

6

259

TABLE 6.1 Pocket Passing Frequency of Different Screw Compressor Types Small Dry Screw Compressor

Large Dry Screw Compressor

Oil-Flooded Screw Compressor

Male-rotor speed (rpm)

18,000

3000

3600

Number of lobes on male rotor

4

4

5

Number of working cycles per minute

72,000

12,000

18,000

Pocket passing frequency (Hz)

1200

200

300

Compressor Type

The thermodynamic behavior of screw compressors is similar to reciprocating compressors. The power consumption is nearly independent of the gas molecular weight. The inlet gas volume flow does not change much when the pressure ratio or the molecular weight of the gas changes. Therefore, screw compressors do not have a surge line. Due to the insensitivity against mole weight changes a screw compressor can operate with a variety of gases. A screw compressor can operate at very high-pressure ratios as long as the allowable discharge temperature or mechanical limits like bearing loads or shaft stress limits are not exceeded. With oil-flooded screw compressors pressure ratios of as high as 25 in one stage can be achieved. With dry screw compressors a pressure ratio up to 10 is possible with liquid injection. The compact rotor design with a small number of lobes gives a robust rotor. The first lateral critical speed is always higher than the maximum allowable speed (“stiff rotor design”). Due to the purely rotating motion the vibrations of screw compressors are similar to centrifugal compressors and much lower than for reciprocating compressors. While vibration in the machine itself is possible (due to imbalance, bearing failure, torsional resonance, etc.), vibration in a screw compressor package (piping, vessels, baseframe, etc.) is much more likely to come from the effects of gas pulsations than from the motion of the compressor itself. The clearances between the rotors and the casing are only fractions of a millimeter which means that any buildup of process gas residues is scraped away and the balance status of the rotors is not affected. In contrast to reciprocating compressors screw compressors do not have valves for controlling the gas inlet and outlet of the working chamber. The gas inlet and outlet are controlled by ports in the casing. Therefore, failure of valves is not an issue with screw compressors.

260 SECTION

II Types of Equipment

Screw compressors do not have any contacting parts other than shaft seals. Oil-injected screw compressors are typically direct driven by electric drivers or gas engines. Therefore, the typical speeds are between 1500 and 3600 rpm. The speeds of dry screw compressors can range between approximately 1000 rpm and up to 25,000 rpm depending on the machine size. The tip speeds for both screw compressor types are lower than those of centrifugal or axial compressors. Therefore, screw compressors are less susceptible to erosion by droplets or contaminants. With oil-flooded screw compressors, erosion is not an issue due to the relatively low rotor tip speeds. Oil-free screw compressors with continuous liquid injection at the suction side can, due to higher rotor tip speeds, be subject to erosion of the rotors and—to a lesser extent—the casing, if carbon steel is used. Stainless steel rotors are often used in this case, and a stainless steel casing is sometimes used as well.

Differences Between Dry Screws and Oil-Flooded Screws Dry Screw Compressors Fig. 6.6 shows sectional views of a dry screw compressor. A synchronizing gear (often called timing gear) is used to avoid contact between the rotors. In oil and gas service the bearings are typically hydrodynamic journal and thrust bearings. At each shaft end a shaft seal is placed between the compression chamber and the journal bearing (sometimes called conveying chamber seals). Depending on the type of coupling and coupling guard a labyrinth seal for sealing the driveshaft may be necessary. An important value for the characterization of screw compressors is the male-rotor tip speed uM. For dry screw compressors the typical tip speed range for all compressor sizes is between 50 and 150 m/s. In some cases, even lower or higher tip speeds are realized. The rotor diameters are in the range of approximately 100 mm up to >800 mm. Therefore, the largest types have speeds of 1000 to 3600 rpm while the smallest sizes may have speeds up to 25,000 rpm with medium-size compressors in between. Due to the high speeds a gearbox between driver and compressor is needed for small- to medium-sized machines. For large machines a direct drive with a steam turbine is often used but an electric drive with or without gearbox is also common. The built-in volume ratio vi is defined by the casing geometry and is a fixed value for a dry screw compressor. Dry screw compressors perform well with many gases (e.g., corrosive, toxic, flammable), and with changing molecular weights. If required a liquid injection for cooling or washing purposes is possible.

Screw Compressors Chapter

6

261

A

4

1

3

B

5

7

6

5

2

4

8

Key A B

Inlet Outlet

1 2 3 4

Casing Male rotor Female rotor Shaft seal

5 6 7 8

Radial/thrust bearing Timing gear End cover Drive shaft

FIG. 6.6 Axial split dry screw. (From MAN Energy Solutions.)

Oil-Flooded Screw Compressors Figs. 6.7 and 6.8 show sectional views of an oil-flooded screw compressor. In oilflooded screw compressors, sometimes referred to as oil-injected or wet screw compressors, oil is injected directly into the rotor chamber continuously during operation. The oil is discharged with the gas into an oil separator vessel and then must be separated from the gas on the discharge side before being injected back into the compressor again. The injection oil serves several purposes:

262 SECTION

II Types of Equipment 1

A

8

11

9 12

10 4

B 4

2

5

5

7

14

6 3

13

5

5

Key A B

Inlet Discharge

1 2 3 4 5 6 7

Casing Male (main) rotor Female (secondary) rotor Radial bearings Axial (thrust) bearings Oil pump (shaft-driven, optional) Shaft seal

8 9 10 11 12 13 14

Capacity slide valve piston Capacity slide valve vi slide (for variable vi, optional) Capacity slide valve position sensor vi adjustment screw (for variable vi, optional) Oil injection port Thrust balancing piston

FIG. 6.7 Oil-flooded screw main components.

1. It provides a lubricating film between the male and female rotors. Oilflooded screw compressors do not have timing gears. Instead, one rotor drives the other through direct contact (with an oil film between the two rotors). The drive rotor refers to the rotor that is coupled to the motor, while the driven rotor refers to the rotor that is moved by the drive rotor.

Screw Compressors Chapter A

1

263

6

6

5 11

4 9

10

7

4

3

2

8

B

Key A B

Inlet Discharge (not visible)

1 2 3 4 5 6

Casing Male (main) rotor Female (secondary) rotor Radial bearings Axial (thrust) bearings Shaft seal

7 8 9 10 11

Capacity slide valve hydraulic cylinder Capacity slide valve (not visible – under rotors) Capacity slide valve position sensor Oil injection port Thrust balancing piston

FIG. 6.8 Oil-flooded screw main components—3D.

Most oil-flooded screw compressors are male-rotor drive, but many femalerotor drive machines are available as well. 2. It carries away much of the heat of compression. This enables higher pressure ratios than are thermally possible in oil-free screw compressors (in which the compression power increases the gas temperature, and a small part of the heat of compression is absorbed by the casing and rotors). 3. It fills the internal clearances, increasing volumetric efficiency. The volumetric efficiency of an oil-flooded screw compressor can be 10%–20% higher than a similar oil-free screw compressor. 4. It continuously flushes away contamination that might enter the machine from the suction header. There are no internal seals at the conveying chamber of an oil-flooded screw compressor—all of the internal components are in contact with the oil and the gas. Therefore, all of the compressor internals must be compatible with the oil and the gas, and also must be rated for the full range of operating pressures and temperatures. The only seal is at the driveshaft, typically an oilpurged mechanical seal, either single or double. When a single mechanical

264 SECTION

II Types of Equipment

seal is used, the seal is usually fed by the common oil system. When higher safety requirements call for a double mechanical seal, it is typical for a separate oil system to be used for the driveshaft seal. This allows better control of the temperature, pressure, and flow of the oil to the seal, and also ensures that any oil leakage on the atmospheric side of the seal does not contain any process gas. Capacity control is possible via an integral slide valve (refer to Fig. 6.9). The slide valve is located below the rotors and moves axially, actuated hydraulically by the oil system and a system of control valves. Since the slide valve is not truly a valve, it is more accurate and clear to use the terms “load” and “unload” rather than “open” and “close” when referring to its movement. Moving the slide valve toward the unloaded position opens a bypass area on the suction side of the machine, which reduces the volume flow of the machine. Moving the slide valve toward the loaded position closes this internal bypass area and increases the volume flow of the machine. With a single-acting slide valve configuration, the slide valve is moved toward the unloaded position via oil pressure, and is moved toward the loaded position via discharge pressure. Thus, a small pressure difference must exist between suction and discharge in order to move the slide valve toward the loaded position. With a double-acting slide valve, the movement toward loaded and unloaded positions is done by the oil system and is not dependent on discharge pressure.

3

2

1 5

4

Key 1 2 3 4 5

Capacity slide valve Capacity slide valve piston vi cut in capacity slide (shape/length of cut determines maximum vi) Internal recycle (when slide valve is away from maximum capacity position) vi slide (changes location of vi cut in capacity slide relative to discharge port)

FIG. 6.9 Slide valve function.

Screw Compressors Chapter

6

265

Design Features In large screw compressors the casing is normally split horizontally (axially) for ease of maintenance (see Fig. 6.6). For oil-injected screw compressors or for small- and medium-sized dry compressors or for higher discharge pressures, vertically (radially) split casings are common (Fig. 6.10). The suction nozzle is normally on top of the casing while the discharge nozzle is in most cases directed downwards, in order to ensure a free draining of any liquids into the discharge line. Thus the gas flow in the casing is directed from top to bottom and also axially from suction to discharge end. Some screw compressors have top or side discharge arrangements, which offers some advantages in terms of piping layout, but this is not common in process gas machines. The pressure difference between discharge and suction causes not only axial gas forces but also large radial gas forces on the rotors. The radial gas forces cause bending stresses in the shafts and loads on the journal bearings. Journal and thrust bearing forces increase linearly with the pressure difference. The direction of the radial forces changes for different operating conditions. Normally the radial forces are much higher than the rotor weight and therefore robust journal bearings are required. The allowable bearing loads and the shaft stress give a limit for the allowable pressure difference. For example, the thermodynamic process gives the same discharge temperature and efficiency when operating with a pressure ratio 3.0 like compressing from 1 bar abs to 3 bar abs, or from 2 bar abs to 6 bar abs or from 10 bar abs to 30 bar abs. Due to the large pressure difference this machine will be mechanically overloaded when compressing from 10 abs to 30 bar abs.

FIG. 6.10 Vertical split dry screw.

266 SECTION

II Types of Equipment

Rotor Design The rotor design is defined by the shape and number of lobes on the rotors and the rotor length indicated by the L/D ratio. For a given rotor diameter the geometrical volume flow increases proportional with the rotor length. It should be noted that the achievable pressure ratio is independent of the L/D; however, the allowable pressure difference is inversely proportional to L/D. The rotor design is carefully chosen with respect to the intended applications. Any design is a compromise between the conflicting requirements of a large volume flow for a given machine size and high rotor bending stiffness. Compressors for high flow and low-pressure difference have long rotors with few lobes. These are normally dry screw compressors. Lobe combinations may be three lobes on the male rotor and four or five lobes on the female rotor (3/4 or 3/5). L/D is >1.8 up to approximately 2.2. Due to the long and slender rotor these machines are only suitable for low-pressure differences. The most common profile is the asymmetric 4/6 profile (Fig. 6.11). This profile is a good compromise for the conflicting requirements of pressure difference and volume flow for a given size and is used for oil-flooded and for dry screw compressors. Depending on the intended pressure difference the L/D ratio may range from 2.5 the torque increases with increasing suction pressure. A massive increase of suction pressure may lead to overloading the compressor and coupling. Therefore it is important that the SOP is as close to the suction pressure as practical. The SOP depends on suction pressure, discharge pressure, and the gas volumes at suction side and discharge side up to the check valve. For start-up and shutdown the discharge side gas volume should be as small as practical and the suction end volume should be as large as practical. Check valves in the suction line must be avoided. Block valves in the suction line should be closed only after the pressures have equalized at SOP. If all these measures are not sufficient to reduce the SOP to an acceptable level a quick opening and sufficiently sized blow down valve to a flare should be placed in the discharge line.

Stopping of Flooded Screw Compressors and SOP Oil-flooded screw compressors must also be unloaded during coast down to stop, to protect the bearings. However, it is also important to avoid a rapid depressurization of the discharge side. The oil in the bulk oil separator, any secondary separator vessels, and everywhere in the oil system is at discharge pressure or higher, with some amount of process gas dissolved in the oil. Rapidly reducing the pressure in the system will cause the gas to rapidly bubble out of the oil, which can lead to foaming of the oil (champagne bottle effect). It can take many hours for the foam to return back to liquid form. If undetected, foaming in the oil can damage the compressor if the system is running. The rapid drop in the pressure can also cause oil migration toward the lower pressure area, resulting in a total loss of oil from the bulk oil separator vessel. It is important to have a check valve on the suction side as well as the discharge side, and for the suction check valve to be as close as possible to the compressor. This ensures that the SOP is only slightly lower than the discharge pressure. During coast down, the slide valve should be moving quickly to the unloaded position to protect the bearings from high load at low speeds. If the recycle valve connects downstream of the suction check valve, it can be opened during coast down. Otherwise, it should remain closed. After the compressor has stopped, if it is necessary to reduce the system pressure, it should be vented slowly to prevent oil foaming.

288 SECTION

II Types of Equipment

Liquid Injection for Dry Screw Compressors Liquid injection is essential for oil-injected screw compressors and is often used for dry screw compressors as well. Dry screw compressors use liquid injection for cooling and washing purposes. The most common injection liquid is demineralized water but other liquids like methanol or heavy hydrocarbons may also be used if they are compatible with the process. Fig. 6.36 shows the effect of water injection on the discharge temperature of a large dry screw compressor at two different discharge pressures. Two modes can be identified which are separated by the saturation temperature. The saturation temperature depends on the humidity of the suction gas and the discharge pressure.

Injection Above Saturation Temperature For low injection flows there is a sharp linear decrease of discharge temperature with increasing injection flow. The cooling effect is caused by evaporation of the injected liquid during compression. This characteristic is valid until the saturation temperature is reached. In this mode the injection flow is controlled by a control valve to adjust a defined discharge temperature. Due to temperature measurement precision this mode is practically limited to a discharge temperature of at least 10 K above saturation temperature. This injection mode is used if the discharge temperature must be reduced due to process or compressor design limits.

250

Discharge temperatrure with water injection [°C]

Discharge pressure 5 bar abs

200 Discharge pressure 3 bar abs

150 100 50 0 0

2000

4000

6000

Injection water flow [kg/h] FIG. 6.36 Temperature versus injection flow for a dry screw compressor.

8000

10,000

Screw Compressors Chapter

6

289

Injection Below Saturation Temperature By increasing the water injection flow the discharge temperature eventually reaches saturation temperature and no more water will evaporate. In this mode even a massive increase of injection water flow leads only to minor reduction of discharge temperature. In this case the injection flow cannot be controlled by discharge temperature but is controlled by a flowmeter and a control valve at a predefined flow. Injection below saturation temperature is often used for processes with dirty gases like acetylene crack gas, lime kiln gas, or coke oven gas. A part of the water does not evaporate but leaves the compressor in liquid form and washes the rotors, casing, and piping. For these processes it should be noted that a practical discharge temperature is only approximately 1 K below saturation temperature. In case of polymerizing gases solvents may be injected in addition to water to prevent sticking of the rotors or clogging of piping components.

Liquid Injection Flows for Dry and Oil-Injected Screw Compressors/Liquid Hammer The quantity of injected liquids differs very much between oil injected and dry screw compressors. For dry screw compressors the volume flow of liquid is determined by the evaporation of the liquid and is normally 5 MPa have been built. Horizontal split casings generally have design pressures up to 1.6 MPa. However, it is important to note that in many cases the compressor may not be able to operate at the maximum discharge pressure of the casing due to other limitations. The maximum discharge pressure must be clearly distinguished from the maximum allowable differential pressure. The maximum allowable differential pressure depends on the rotor profile, lobe numbers, L/D, shaft diameters, and bearing design and size. The maximum allowable differential pressure is often much less than the maximum discharge pressure. For example, even if the

292 SECTION

II Types of Equipment

TABLE 6.3 Optional Instrumentation for Dry Screw Compressors

Discharge pressure Pressure difference between suction and discharge p2–p1

Pressure ratio p2/p1 Compressor speed for variable speed drivers or turbines Seal liquid temperature (for oilcooled seals or watercooled seals) Seal gas pressure (for all kinds of seals with external gas supply) Seal liquid pressure (for oil-cooled seals or water-cooled seals) Shaft axial thrust

Alarm low

Trip low

In order to prevent excessive overcompression

In order to prevent excessive overcompression

Alarm high

Trip high

If bearings or shafts may be overloaded

If bearings or shafts may be overloaded

Alarm high

Trip high

For compressors with varying operating pressure levels and if discharge temperature may become too high

For compressors with varying operating pressure levels and if discharge temperature may become too high

Alarm high

Trip high

Alarm low

Trip low

Alarm high

Trip high

Alarm low if liquid may freeze or clog

Alarm low if liquid may freeze or clog

Alarm low

Trip low

Alarm high, if seals can be damaged by high pressure

Trip high, if seals can be damaged by high pressure

Alarm low

Trip low

Alarm high, if seals can be damaged by high pressure

Trip high, if seals can be damaged by high pressure

Alarm high

Trip high

design pressure of the casing is 40 bar the maximum allowable differential pressure may be limited to 1.5 MPa by the rotors and bearings. Thus the compressor may be able to compress from 1.5 to 3 MPa but not from 1.5 to 4 MPa. In this case a differential pressure monitoring is required to ensure that the compressor rotors and bearings are not overloaded. Due to the variety of rotor designs and L/D only very rough values for the maximum allowable differential pressure of dry screw compressors can be given. This may range from 0.4 MPa for long rotors with 3/5 profile up to approximately 1.5 MPa for short rotors with 4/6 profile. The allowable differential pressure must be determined on a case-bycase basis. Examples for high discharge pressure or high-pressure difference are no. 3 in Table 6.6 and no. 12 in Table 6.8.

Screw Compressors Chapter

6

293

TABLE 6.4 Standard Instrumentation for Oil-Flooded Screw Compressors Suction Pressure

Alarm Low

Trip Low

Suction pressure

Alarm high

Trip high

To prevent overcompression or high torque Discharge pressure

Alarm high

Trip high

Suction temperature

Alarm high

Trip high

Discharge temperature

Alarm high

Trip high

Thrust bearing temperature (if hydrodynamic)

Alarm high

Trip high

Journal bearing temperature

Alarm high

Trip high

Casing vibration or shaft vibration monitoringa

Alarm high

Trip high

Shaft axial thrust

Alarm high

Trip high

Pressure difference between lube oil and discharge gas

Alarm low

Trip low

Lube oil temperature

Alarm high

Alarm high Trip high

Start-up interlock low Shaft seal oil flowrate

Driveshaft seal oil pressure

Alarm low

Trip low

Alarm high

Trip high

Alarm low

Trip low

Alarm high Temperature difference driveshaft seal oil outlet over inlet

Alarm high

Trip high

a Due to the stiff casing design, 1 accelerometer per compressor or 1 accelerometer at each compressor end (max 2 in total) is sufficient for automatic monitoring with shutdown.

The pressure ratio is limited by the discharge temperature and the undercompression. Extreme undercompression may cause severe gas pulsations in the discharge line. If the pressure ratio is too high the discharge temperature may be higher than allowable for the rotor clearances. The discharge temperature depends also on the gas molecular weight and the isentropic exponent. For dry screw compressors without liquid injection the maximum pressure ratio may range from approximately 2.5 for light gases up to approximately 6 for hydrocarbons with molecular weights of 50 or higher. Again a case-by-case decision is required.

294 SECTION

II Types of Equipment

TABLE 6.5 Optional Instrumentation for Oil-Flooded Screw Compressors Alarm low To prevent excessive overcompression and oil carryover

Trip low

Alarm high

Trip high

If bearings or shafts may be overloaded

If bearings or shafts may be overloaded

Pressure ratio

Alarm high

Trip high

p2/p1

For compressors with varying operating pressure levels and if discharge temperature may become too high

For compressors with varying operating pressure levels and if discharge temperature may become too high

Compressor speed for variable speed drivers or turbines

Alarm high

Trip high

Alarm low

Trip low

Discharge pressure Pressure difference between discharge and suction

With delay to allow normal start-up

p2–p1

Dry screw compressors with liquid injection for discharge temperature limitation can operate with pressure ratios up to approximately 10 in low-pressure applications. A typical example is the styrene monomer process with suction pressure 20–25 kPa abs and discharge pressures of 160–200 kPa abs or even 250 kPa abs (see examples nos. 6 and 7 in Table 6.7). The actual suction flow depends on the compressor size (given by rotor diameter and L/D) and speed. The lower end of the flow is approximately 300 m3/h for the smallest dry screw compressors. The largest dry screw compressors in operation today have an actual volume flow of 77,000 m3/h (see Fig. 6.37, and Table 6.2 nos. 6 and 8, and Table 6.8 no. 14). Dry screw compressors with even larger flows of 120,000 m3/h have been announced but are not yet on the market. The driver power of dry screws can range from 1%–3%, even on hot days. The impact of humidity tends to increase at higher ambient conditions. In two-shaft engines, the power turbine speed impacts the available power and efficiency. For any load and ambient temperature, there is an optimum power turbine speed. Usually, lowering the load (or increasing the ambient temperature) will lower the optimum power turbine speed. Small deviations from the optimum, such as 10% have very little impact on power and efficiency (Fig. 7.6).

Maintenance and Reliability One key advantage of gas turbine drivers is the high availability and reliability, which can be higher than 98%. Typically, the recommended time between major inspections is 30,000 h of operation or more. Maintenance intervention is performed on site or in dedicated facilities. While heavy industrial machines are usually maintained on site, modern light-industrial gas and aero-derivative gas turbines allow the choice between on-site maintenance, component exchanges, or complete engine exchanges. Component exchanges, or engine exchanges, allow for minimal downtime for maintenance interventions.

Reciprocating Engines As a compressor driver, reciprocating gas engines occupy a level of complexity between gas turbines and electric motors. They also bring characteristics that allow them to serve in applications where the use of either the turbine or motor may be challenging or more costly to execute. The reciprocating piston design is

Drivers Chapter 7

FIG. 7.6 Performance characteristics.

315

316 SECTION

II Types of Equipment

older than most people realize, yet its use in the automobile is so routine that most take it for granted as part of their daily lives. It is a proven technology that serves society effectively every day. The reciprocating engine most commonly used in gas compression is a sparkignited internal combustion design fueled by the natural gas itself. The thermodynamic theory involved is, at the basic level, no different than that for the turbine: fuel and air are combined, compressed, and ignited, and the expansion of the burning gas is tapped to extract usable shaft work to power other processes, such as the gas compressor at the back of the compressor package. Reciprocating internal combustion engines can be sorted at the highest level into two basic designs: the two-stroke cycle and the four-stroke cycle (Fig. 7.7). Both use a piston moving inside an enclosed cylinder to draw fuel and air in, compress the mixture, ignite with a spark, then use the expansion of the burning gases to force the piston to turn a crank. But while the four-stroke cycle uses separate movements of the piston to do each step (intake-compressionpower-exhaust), the two-stroke design accomplishes those activities with just two piston movements. The two-stroke design claims the advantages of relative simplicity and power density. The two-stroke designs commonly use open ports for intake and exhaust, with flow being regulated by the position of the piston instead of actuated valves as are found in a four-stroke engine. Also, in the two-stroke each cylinder has a power stroke on every rotation of the crankshaft, giving twice as many firing impulses as a four-stroke engine. The four-stroke design enables more precise control over the air/fuel mixture and the intake and exhaust valve timing, factors which can be manipulated to achieve highefficiency and low exhaust emissions. When speaking of drivers for high-speed separable compressors, most engines today are four-stroke designs. Four-stroke cycle Intake valve open Air-fuel mixture

Spark plug

Valves closed

Valves closed

Intake valves closed

Exhaust valve open

Exhaust valve closed Spark plug firing

Exhaust gases

Combustion chamber Piston Connecting rod

Crankshaft

Intake Air-fuel mixture is drawn in

Compression Air-fuel mixture is compressed

Power Explosion forces piston down

Exhaust Piston pushes out burned gases

© 2007 Encyclopædia Britannica, Inc.

FIG. 7.7 Schematic representation comparing two- and four-stroke cycles.

Drivers Chapter

7

317

Engine Starting Starting an internal combustion engine is simple in concept: a small motor spins the engine up to a speed where combustion is consistent enough to sustain the rotation by itself, at which point the starter motor is disengaged. The size of the starter depends on the total torque it must generate to turn the engine’s crankshaft, which makes it dependent not just on the size of the engine, but also on the equipment that is connected to the engine during cranking. In gas compression applications, the compressor is directly coupled to the engine and thus places additional load on the starter during cranking. Site design considerations like an adequately sized compressor recycle loop are important to achieving a successful starting solution for the engine. Temperature is also a factor in engine starting—a cold cylinder block saps heat energy from combustion during cranking, and cold oil increases in viscosity resulting in higher frictional loads at start-up. Auxiliary heaters for the engine’s coolant and lubricating oil may be required as starting aids in a cold weather environment.

Power, Torque, and Speed The rated power and operating speed of an engine tend to follow from the engine’s combustion chamber size. Because the piston’s movement in the cylinder determines the volume of fuel and air taken into the engine, the combined volume of all the cylinders becomes an important basic design detail—the engine’s displacement. A larger-displacement engine has the ability to pull in greater amounts of fuel, thus giving it more potential to produce power. Larger pistons also have greater mass that must change direction at the ends of each stroke, and the inertia forces involved in that turnaround tend to keep operating speeds lower on larger engines. While the exact shape of the torque-speed curve varies from engine-toengine, the heavy-duty gas engines used in gas compression typically advertise constant torque across the range of speeds over which the engine may operate continuously. This fits well with the operating behavior of reciprocating compressors, which, for a given set of operating conditions, demand a fixed operating torque regardless of speed. Because the engine is optimized for operation at the rated speed, the amount of “speed turndown” (operation at reduced speeds) available tends to be limited, with the actual minimum operating speed being a characteristic of each individual engine.

Durability As with all machines, durability and ease of maintenance are prime factors in the design of the engine. Long life adds value, both in extending the period of time over which a given engine continues to perform well, and in limiting the

318 SECTION

II Types of Equipment

accumulated time during which the installation must be stopped to conduct routine maintenance. Gas engines require scheduled stoppages to inspect and/or replace the spark plugs and to change the oil and filters. Other required adjustments or component servicing can be performed less frequently, but still do factor into the overall owning and operating costs of the engine. Effective operating strategies seek to extend these servicing intervals as much as possible without risking component failures. One area of focus relatively unique to the gas engine driver is in addressing torsional vibrations. Torsional vibrations are small irregularities in the turning speed of the package that can have consequences because they repeatedly flex the rotating parts (crankshafts, etc.) during operation (Fig. 7.8). If these oscillations tune-up with the operating characteristics of the rotating driveline (a condition known as resonance), the flexing can become significant and could potentially lead to a fatigue-related failure. The discrete linear motions of reciprocating machines (engines and compressors) are sources of torsional vibration, so the engine’s design includes two main devices to help to limit such vibrations: the flywheel and the torsional damper. The flywheel serves as a high-inertia barrier, smoothing the rotation of the engine’s crankshaft and limiting the amount of oscillation that can reach the crankshaft from the connected compressor. The damper absorbs much of the torsional oscillations in the engine’s crankshaft, dissipating them to the surrounding air as heat. Even with the these in place, a torsional vibration analysis (TVA) is considered essential to verify that the engine, compressor, and coupling work together at all of the expected speed-load operating points to ensure the risk of fatigue-related issues is minimal.

Simplified drive train

FIG. 7.8 A simplified piston-and-crank drive train, showing how torsional inputs result in rotational “wind-up” of the crank.

Drivers Chapter

7

319

Electric Motors The two common electric motor types are the induction (asynchronous) and the synchronous motors. Both types are made up of stationary (stator) and rotating (rotor) components, where an alternating current is applied to the stator to create a rotating magnetic field. The rotor, a permanent magnet or electromagnet, reacts with the magnetic field to rotate within the stator.

Induction Motor Basic Principles The induction motor is the motor in which alternating current is supplied directly to the stator and current in the rotor is induced from the stator. When the stator is excited from a balanced three-phase source, it will produce a magnetic field in the air gap rotating at synchronous speed ns as determined by the number of stator poles and the applied stator frequency fe:
 120  fe (7.1) ns ¼ poles Three-phase current in the stator produces a magnetic flux of constant magnitude rotating at synchronous speed. As per Faraday’s law, rotating flux cuts the rotor conductors causing EMF (elector-motive force) induced in the rotor bars. EMF in shorted rotor bars induces rotor current flow. By the Lenz’s law, the current flow in the rotor conductors creates mechanical force. The result is the rotor turns in the direction of the magnetic field trying to catch up. The rotor speed nm is less than the synchronous speed of the stator-induced magnetic field ns and it is defined by the slip s: n s  nm s¼ (7.2) ns An induction motor can be described as a rotating transformer. Its input is a three-phase system of voltages and currents. For an ordinary transformer, the output is electric power from the secondary windings. The secondary windings in an induction motor (the rotor) are shorted out, so no electrical output exists from normal induction motors. Instead, the output is mechanical. The relationship between the input electric power and the output mechanical power of this motor is illustrated in Fig. 7.9. The input power to an induction motor Pin is the three-phase electric power. The first losses encountered in the machine are I2R losses in the stator windings, that is, the stator copper loss PSCL. Some amount of power is lost as hysteresis and eddy currents in the stator Pcore. The power remaining at this point is transferred to the rotor of the machine across the air gap between the stator and rotor. This power is called the air gap power PAG of the machine. After the power is transferred to the rotor, some of it is lost as I2R losses, that is, the rotor copper

320 SECTION

II Types of Equipment Pconv

PAG Air-gap power

τindwm

Pin = √3 VTIL cos q

PSCL (stator copper loss)

Pcore (core losses)

Pout = τloadwm

Pstray Pfriction PRCL P and windage ( misc.) (rotor copper loss)

FIG. 7.9 Induction motor—power balance diagram.

loss PRCL, and the rest is converted from electrical to mechanical form Pconv. Finally, friction and windage losses and stray losses Pmisc are subtracted. The remaining power is the output of the motor Pout. The synchronous speed is shown as an angular synchronous speed ωs and the rotor speed is shown as an angular rotor speed ωm. Some losses vary with the motor speed. The higher the speed of an induction motor, the higher the friction, windage, and stray losses. On the other hand, the higher the speed of the motor (up to ns), the lower its core losses. Therefore, these three categories of losses are sometimes lumped together and called rotational losses. The total rotational losses of a motor are often considered to be constant with changing speed, since the component losses change in opposite directions with a change in speed. The induced torque in a machine is defined as the torque generated by the internal electric-to-mechanical power conversion. This torque differs from the torque actually available at the terminals of the motor by an amount equal to the friction and windage torques in the machine. Hence, the developed torque is τind ¼

Pconv ωm

(7.3)

While another way to express torque is τind ¼

ð1  sÞ  PAG PAG ¼ ð1  s Þ  ω s ωs

(7.4)

The torque-speed relationship as a general equation for torque as a function of slip is derived from the induction motor equivalent circuit. The induced torque is 0 12

τind ¼

VTH B C R2 3@qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi A  2 s 2 R2 RTH + =s + ðXTH + X2 Þ ωsync

(7.5)

Drivers Chapter

7

321

where VTH, RTH, and XTH are the Thevenin equivalents of voltage, resistance, and reactance derived from the induction motor equivalent circuit and R2 and X2 are the rotor resistance and reactance. If a graph of torque and speed is plotted based upon the changes in slip, a graph similar to the one in Fig. 7.10 would be obtained. The locked rotor torque is the torque developed by the motor at standstill. This torque is sometimes also referred to as the starting torque. The pull up torque is the minimum torque developed by the motor as it accelerates from standstill to the speed at which breakdown torque occurs. The breakdown torque is the maximum torque that the motor is capable of developing. The full load torque is the steady-state torque developed during operation. It is the torque necessary to produce the motor rated horse power at full speed. Torque in pounds at a foot radius (lb ft) is equal to the horsepower times 5252 divided by the full speed in rounds per minute (NEMA Standards Publication No. MG 1-2016, Motors and Generators, 2016). At no load, the rotor of an induction motor would rotate at near synchronous speed. The synchronous speed can never be reached with an induction motor. If rotor of an induction motor was rotating at synchronous speed, the rotor bars would be stationary relative to the stator magnetic field. There would be no induced voltage in the rotor and no rotor current. Therefore, the induced torque would be equal to zero and the rotor would slow down below the synchronous speed due to friction. To use an induction motor to drive a load, for example, a fan, a pump, a compressor, or a conveyer, it is important to know the load torque-speed curve and to analyze its interaction with the motor torque-speed curve. This is to ensure that the motor will start successfully and operate the load in the stable region of the torque-speed curve. In Fig. 7.11, the stable region is on the right side of the breakdown torque from points C to E.

A – Locked rotor/starting torque

C – Breakdown torque D – Full load torque E – Synchronous speed

Percent of full-load torque

B – Pull-up torque

Locked rotor torque (starting torque)

Breakdown torque (pullout) C

200 Pull-up torque (pull-in) A B

100

D Full-load torque

No-load speed 0

0

50 Percent of synchronous speed

FIG. 7.10 Typical induction motor torque-speed curve.

E 100

322 SECTION

II Types of Equipment

Breakdown torque (pullout)

Percent of full-load torque

Locked rotor torque (starting torque)

C

200 Pull-up torque (pull-in) A B

100

D Full-load torque

Load torque-speed curve

Stable operating point

No-load speed 0

0

50

E 100

Percent of synchronous speed

FIG. 7.11 Stable operating point.

From Fig. 7.11, the motor would accelerate the load successfully since the motor torque is greater than the load torque up to the operating point. If load torque at the operating point increases, the motor will slow down. As it slows down, at lower speed, the motor torque increases and catches up with the increased load torque. The motor operation is stable. From Fig. 7.12, the motor would also accelerate the load successfully since the motor torque is greater than the load torque up to the operating point. If load torque at the operating point increases, the motor will slow down. As it slows down, at lower speed, the motor torque decreases so the motor slows down further. Eventually, the motor would stall. The motor operation is unstable. For a given speed, the motor torque varies at the square of voltage. This is important during motor start. When motor starts direct on line, the starting current is much larger than the full load current (e.g., 600% of the full load current) and causes the voltage drop at the motor terminal. At the moment when the motor starts from the locked rotor condition, the voltage at the motor terminal is typically between 80% and 90% of the rated voltage. As the motor accelerates, the motor current is reducing from the locked rotor current to the full load current and the motor terminal voltage recovers from the initial drop to a value approximately equal or close to the full-rated motor voltage. So if for example, initially, the voltage drops to 80% of the rated voltage, the motor starting torque would drop to 64% of the theoretical starting torque if starting voltage was equal to 100% of the full voltage. As the voltage recovers from the initial drop, the torque recovers too. Between the initial voltage drop at the motor terminals and the full motor voltage, there is a family of torque-speed characteristics, and the actual motor torque will be a point on a corresponding torque-speed

Drivers Chapter

323

Breakdown torque (pullout)

Locked rotor torque (starting torque) Percent of full-load torque

7

C

200 Pull-up torque (pull-in) A

Unstable operating point

B

100

D Full-load torque Load torque-speed curve No-load speed

0

0

E 100

50 Percent of synchronous speed

FIG. 7.12 Unstable operating point.

characteristic for a given speed and given voltage, as it transitions from the starting condition to the normal running condition. It is important to have the load torque less than the motor torque for every speed and the corresponding voltage during the motor start, so the motor can accelerate the load successfully. The difference between the motor torque and the load torque is called the acceleration torque. Fig. 7.13 shows the range of motor torque-speed characteristics from starting voltage of 80% to the motor full-rated voltage.

% Full load torque

200

Torque at 100% voltage Torque at 80% voltage

100

0

0

20

40

60

80

100

% Synch speed FIG. 7.13 Motor torque-speed curve at 100% voltage and 80% voltage.

324 SECTION

II Types of Equipment

% Full load torque

200

Torque at R1 Torque at R2

100

Torque at R3 R3 > R2 > R1

0 0

20

40

60

80

100

% Synch speed FIG. 7.14 Motor torque-speed curve for different rotor circuit resistances.

Rotor Circuit Resistance and Motor Torque The rotor circuit resistance has significant impact on starting torque, the speed at which the breakdown torque occurs, and the slip during the normal running operation. Fig. 7.14 shows how the motor torque-speed curve would change if rotor resistance is increased and all other parameters in the motor equivalent circuit stay the same. From Fig. 7.14, a high rotor resistance will provide a high starting torque, leading to rapid acceleration of the mechanical system. This is desirable because short acceleration times reduce the stress on the power system caused by high starting currents. While high starting torques are desirable, high rotor resistance results in a relatively high slip during normal running operation. The high resistance causes increased losses and reduced efficiency during normal operation. For most applications, it is desirable to have high starting torque and high efficiency at rated speed. However, designs with high starting torque will have low efficiency at rated speed and designs with high efficiency will have low starting torque. The National Electrical Manufacturers Association (NEMA) has established four different designs for electrical induction motors: designs A, B, C, and D. Different motors with the same nominal horsepower may have different start current, torque curves, speeds, and other variables. When a motor is being selected for an intended use, all engineering parameters must be considered. The four NEMA designs have unique torque-speed characteristics making them suited for different types of applications. The typical torque-speed characteristics of the NEMA design motors are shown in Fig. 7.15. As explained previously, the rotor circuit resistance plays the key role in the motor torque-speed relationship. The NEMA designs are based on different designs of the rotor circuit. Fig. 7.16 shows laminations from typical NEMA design cage induction motors and the cross section of the rotor bars.

Drivers Chapter

7

325

300 Design D

Torque (%)

250 200

Design C

150

Design A and B

100 50 0 0

20

40

60

80

100

Speed (%) FIG. 7.15 Motor torque-speed curves for different NEMA design motors.

FIG. 7.16 NEMA designs—laminations from typical cage induction motors—cross section of the rotor bars (find clearer pictures of rotor laminations).

326 SECTION

II Types of Equipment

Motor classes A and D are obtained by designing the rotor resistance to be either low (class A) or high (class D). Classes B and C are obtained by exploiting the skin effect to obtain a variable resistance rotor circuit, as the frequency of the current induced in the rotor changes during motor acceleration. The rotor frequency may be expressed as f2 ¼ s  fe

(7.6)

where f2 is the rotor frequency, fe is the stator frequency, and s is the slip that changes during motor acceleration from 1 in the locked rotor condition to some final value at the full speed. Typically, slip s at the full speed is 100% of the full load

332 SECTION

II Types of Equipment

100

% Torque

80 Centrifugal compressor 1 Centrifugal compressor 2

60 40 20 0

0

20

80 40 60 % Synch speed

100

FIG. 7.20 Examples of centrifugal compressors torque-speed curves.

torque, and breakdown torque equal to or >150% of the full load torque. With induction motors, the torque requirements for centrifugal compressors are typically achieved with NEMA Class A and B motors for starting with inlet or discharge valves closed, and NEMA Class C and D motors for starting with inlet or discharge valves open. Synchronous motors or variable speed drives can be used as well.

Synchronous Motors Synchronous motors operate at synchronism with the line frequency and maintain a constant speed regardless of load. There are two major types of synchronous motors: non-excited and direct current excited. Non-excited motors are manufactured in reluctance and hysteresis or permanent magnet designs. These motors employ a self-starting circuit and require no external excitation supply. Non-excited synchronous motors are typically smaller motors rated 50 kW or less. The synchronous motors used in compressor drive applications are typically direct current-excited motors. Excitation circuit is on the rotor and requires external supply. Those motors are designed for much higher power requirements than the nonexcited synchronous motors. The largest electric motors today are the direct-current excited synchronous motors, they are built in the range of up to 80 MW. There are two types of direct online starting methods of the direct current excited synchronous motors: (a) Motor starting with an external prime mover or starter. 4. Synchronous motors are mechanically coupled with another motor. The starter motor can be either a three-phase induction motor or a DC motor. The DC excitation of the synchronous motor is turned off initially. The starter motor helps the synchronous motor accelerate and rotate at speed

Drivers Chapter

7

333

close to its synchronous speed, at which the DC excitation is turned on. When magnetic locking between the stator and rotor takes place, the power is no longer supplied to the external starter motor. (b) Motor starting with damper winding. 5. The synchronous motor is of salient pole type, additional winding is placed in rotor pole face. Initially, when rotor is at standstill, relative speed between damper winding and rotating air gap flux is large and an electromagnetic force is induced in it to produce the required starting torque, according to the same principles as in the induction motor. As the motor approaches synchronous speed, excitation on the rotor is turned on by applying DC power. The rotor is locked in and synchronized with the stator field. Hence, in this case, the synchronous motor is first run as an induction motor using additional winding and finally it is synchronized with the frequency. Synchronous motors with direct current excitation are typically supplied with excitation control equipment to apply the excitation in the correct moment and control the power factor. Once the motor is in operation, the speed of the motor is dependent only on the supply frequency. When the motor load is increased beyond the breakdown load, the motor falls out of synchronization and stalls, such as when the applied load is large enough to pullout the field winding from following the rotating magnetic field.

Why Use Synchronous Motors? Synchronous motors need excitation and they are generally more complex and more expensive than the induction motors. Why are they used? The main advantages of synchronous motors are: l

l

Rotational speed is independent of the load. The motor operates at constant RPM (revolutions per minute). This is important in process industries where constant speed is important for the quality of the product, for example, paper industry. Efficiency is higher than of an induction motor of the same output and voltage rating because there are neither losses related to slip nor the additional losses due to magnetizing current. With synchronous motors, there is no difference of speed between air gap rotating magnetic field and rotor. With induction motors, rotating magnetic field and rotor are not at the same speed, so eddy losses are present and those losses introduced by the slip are mainly responsible for reduced efficiency. In addition, with synchronous motor, the excitation is applied directly on the rotor field winding, while with induction motor, the power required for excitation is coming from the stator and induced on the rotor, so additional losses due to magnetization are present with the induction motor.

334 SECTION l

l

II Types of Equipment

Synchronous motors have larger air gap. In induction motors, the electromagnetic force induced in the rotor winding is mutually induced electromagnetic force. If the air gap is large, then the leakage flux will increase and the mutual flux would reduce. As a consequence, the rotor electromagnetic force and torque would be reduced. In a synchronous motor, the magnetic flux is derived separately from the field winding at the rotor. The electromagnetic force induced in the stator armature winding is a dynamically induced electromagnetic force due to relative motion between the field and the conductors. The air gap can be larger and noise and vibration are generally less than with the induction motors. When the synchronous motor is overexcited, it generates reactive power, which improves overall consumption power factor of the plant. Power factor will have a significant impact on the electric utility bill costs. Electric utility companies have a minimum power factor threshold, typically 0.9, that industrial customers must maintain in order to prevent additional power factor charges. Synchronous motors help improve overall power factor and may eliminate the need of power factor correction equipment, for example, capacitor banks.

Current Pulsation In a synchronous motor, AC power is supplied to the stator to generate a rotating magnetic field. DC power is supplied to the rotor which results in discrete North (N) and South (S) poles. The poles in the rotor then lock onto (synchronize) and follow the opposing rotating magnetic pole (N follows S). At zero load, they follow exactly, but at load they follow slightly behind by a load angle which varies between 0 electrical degrees at zero load to typically between 20 and 30 degrees at 100% load and approximately 70 degrees at stall. Fig. 7.21 shows example where load angle at full load is 32 degrees. There are 180 electrical degrees between each adjacent N and S pole. So take the previous example where the torque variation was 40%, the torque would vary between 60% and 140% and the magnetic lag would vary between 0.6  32 ¼ 19.2 degrees and 1.4  32 ¼ 44.8 degrees. In a synchronous motor, however, the exciting amps are varied to keep the power factor constant with load and so the amps would also vary between 60% and 140% nameplate, the average amps would be 100%, average power 100%. The current pulsation would be (140–60)/ 100 ¼ 80%. So in this case, the NEMA limit of 66% current pulsation is adequate to protect the motor because a synchronous motor is less affected by torque pulsations. API 618 also recommends 66% as a current pulsation limit. Natural Frequency Note that the rotor lags the stator magnetic by an amount proportional to the torque. The magnetic field acts as a spring and the rotor inertia and drive inertia will have a natural frequency that is equal to:

Drivers Chapter

7

335

gnetic field rotation Stator ma h

m Magnetic flux in air-gap

Dotted line unstable operation

C B

N

S

A

d

d 0°

Electrical degrees

90°

180°

32°

S

S'

N

Rotor FIG. 7.21 Synchronous motor operation in synchronism: at full load (example 32 degrees) and at no load (0 degree).

35; 200  fn ¼ n

rffiffiffiffiffiffiffiffiffiffiffiffi Pr  f WK 2

(7.7)

where fn is the undamped natural frequency in cycles/min, n is the synchronous speed in rpm, Pr is the synchronizing torque coefficient, this is the stiffness of the electrical field applying torque to the rotor to restore it to the neutral position in units of KW/rad (see NEMA MG1 21.36.2), and f is the electrical line frequency in Hz, WK2 is the inertia in units lbs-ft2. The electrical natural frequency obviously must be separate from the running speed in order for the current pulsation to be below 66%. Usually, it is between 0.5 and 0.75 times the running speed. Also referring to the NEMA MG1 21.38, there is a term called the compressor factor C. C¼

0:746  WK 2  n4 Pr  f  108

(7.8)

Therefore, the term compressor factor (C) gives a dimensionless measure of the inertia in the motor and driven system. At C ¼ 9.24, the system is in electrical resonance and the natural frequency is proportional to 1/C1/2. Generally for the multiple throw machine such as a six-throw, satisfactory current pulsation will normally be achieved at C ¼ 20 (fn ¼ 0.68  running speed) and for a two-throw at C ¼ 30 or greater (fn ¼ 0.55  running speed). It is typically acceptable to have a motor with minimum C ¼ 20 for a two-throw and C ¼ 15 for four-throw or more. Any additional inertia required will be in the compressor or can be added to the compressor flywheel. The current pulsation for both induction and synchronous motors should always be checked during the engineering phase by the motor manufacturer using the full load and part load

336 SECTION

II Types of Equipment

crank effort and the driven inertia (supplied by the packager). Most motor manufacturers will calculate the current pulsation for both induct and synchronous motors using a simple single degree of freedom model, electrical bus-magnetic field stiffness inertia of motor and driven equipment lumped as one mass. This is accurate for most systems where the torsional natural frequency is above four times running speed. For applications where the first torsional natural frequency is below four times running frequency, a more accurate result can be obtained using a 2 degree of freedom model; electrical bus—magnetic field stiffness— motor inertia—equivalent shaft stiffness—flywheel and compressor stiffness, lumped as one mass. Not all motor manufacturers can do this calculation, but it will show more accurate and lower current pulsation for torsionally soft systems. The motor manufacturer will determine the minimum amount of system inertia required to achieve 66% for synchronous and 40% for induction motors. The required inertia can be added using a compressor mounted flywheel if the motor inertia is not adequate. Note for a synchronous motor it is recommended that the driven inertia always be less than the motor inertia. This is because of the strong (up to 40% of nameplate torque) 2  slip frequency pulsating torque that is induced during acceleration. During acceleration this 2  slip frequency torque will inevitably coincide and excite torsional resonance at some brief point during start-up assuming that torsional resonances occur between 0 and 120 Hz (for a 60 Hz electrical frequency). Reciprocating compressor drives are torsionally robust so if the driven inertia is less than the motor inertia then the resulting torsional stresses are unlikely to exceed permissible limits. See Chapter 5 for a further discussion of torsional analysis.

Summary Comparison between Induction and Synchronous Motors Synchronous motor Typically, 1%–2% better efficiency than induction and runs at a power factor of 1.0 or 0.8 leading. Therefore, synchronous motors can improve the overall power factor of the plant resulting in reduced electrical demand charges and are less affected by the pulsating compressor torques. Has low starting current than an induction motor. Starting torque is lower than induction motor, 40/30/ 150 is typical but 60/60/175 is available. Care is required for proper compressor start up unloading especially at the 95%–100% speed point where the motor pulls into synchronization. A synchronous motor is especially useful in slow speed applications (under 400 rpm) when it is comparable to the cost of an induction motor and is normally used in a single-bearing configuration directly and rigidly bolted to the crankshaft. Induction motor Lowest-cost driver provides good starting torque, usually 60% at breakaway which continuously rises to a peak at about 95% speed. Higher torques are

Drivers Chapter

7

337

available but it is usually better to ensure adequate start-up unloading of the compressor. Starting currents are high for full voltage across the line starting and so a strong power supply is required. Power factor is quite low usually in the range of 0.5–0.8 lagging. Lower speed motors and high-efficiency motors tend to have lower power factors. This can result in higher power costs due to demand charges unless separate power factor correction equipment is installed. Induction motors are quite sensitive to pulsating torques especially electrically stiff motors (low slip). The user is cautioned about the use of high-efficiency motors with low slip (1% or lower), as compared to a standard motor with 1.5% or greater slip. Due to the negative effect of pulsating torque on current pulsation and power factor the hope for improvement in efficiency may not be achieved and much greater inertia is required in the compressor flywheel to achieve acceptable current pulsation (ref API 618).

Mechanical Effects Torque Pulsation The compressor imposes a strong pulsating torque on the motor shaft at 1 and all higher harmonics of the compressor speed, the strongest harmonic is a function of the number of throws and whether the compressor is at full load (all cylinders double acting) or part load (some or all cylinders single acting or unequal load head end to crank end). A two-throw compressor will have a very strong 2  harmonic. The strongest harmonic of the four-throw is the fourth harmonic. The six-throw strongest is usually the sixth. The GMRC (gas machinery research council) torsional guideline [2] notes that for a two-throw compressor the motor should be designed for a torque of 100% of nameplate mean torque plus a pulsation torque of 250%, for a 100% mean plus 200% pulsating, sixthrow compressor 100% mean and 150% pulsating. The pulsating torques can be reduced by a compressor flywheel and a flywheel normally will be required on a two- and four-throw compressor in order to meet the current pulsation requirement. The motor should be a heavy duty (also called severe duty). For heavy duty applications, the motor manufacturer will typically provide a large shaft, heavy duty fan, more internal bracing of windings, etc. Keyways are best avoided, but if necessary they should include a radius fillet to minimized torsional stresses [2]. It is best if the motor shaft at the drive end from the core through shaft end (including bearing) be no less than the crankshaft stub shaft [2]. Lateral Pulsation The compressor imposes a strong lateral pulsation on the motor. This comes from the rod loads of the compressor. They are transmitted to the motor in two ways. Through the baseframe and foundation, and through the crankshaft to the motor shaft. At each revolution, the two throws adjacent to the drive apply

338 SECTION

II Types of Equipment

the rod load to the crankshaft at the instantaneous angle of the connecting rod. As the crankshaft rotates, it roughly orbits and because the two drive end throws are in phase but opposed, the crankshaft also angles. The motion of the crankshaft is limited by the bearing clearance which will typically not exceed 0.1% of the shaft diameter as a diametral clearance. The motion at the crankshaft is primarily at 1 but there is some excitation at all higher harmonics. For a two bearing motor connected via a flexible coupling the crank lateral motion will cause vibration of the flywheel (0.2 mm p.p. for a 200 mm crankshaft assuming nonresonant motion), but only a limited amount is transmitted through the coupling to the motor. However, for rigidly connect motors, for example, a singlebearing synchronous motor, the entire lateral motion of the crankshaft is transmitted directly. The outboard bearing and motor shaft needs to be designed for the crankshaft orbital motion while limiting the vibration at the bearing to acceptable limits. Because of the compressor pulsating loads the vibration at the drive motor will exceed what is normally considered acceptable. For example, ISO 10816-3 list an acceptable vibration at the motor of 4.5 mm/s rms, however, this standard specifically excludes motors driving reciprocating compressors. A more realistic limit would be the acceptable value listed for the compressor frame which is 8 mm/s rms as a typical vibration limit for the motor frame and the bearings. Reciprocating compressors impose a cyclic torque and lateral vibration transmitted through the crankshaft to the motor shaft and lateral vibration transmitted to the motor through the foundation. This means that driving a reciprocating compressor needs to be considered severe duty and appropriate considerations are required.

Variable Frequency Drives Induction and synchronous motors are designed for a specific voltage per frequency ratio (V/Hz). Voltage is the supply voltage to the motor, and frequency is the supply frequency. The V/Hz ratio is directly proportional to the amount of magnetic flux in the motor magnetic material (stator and rotor core laminations). The torque developed on motor shaft is proportional to the strength of the rotating flux. The type and the amount of magnetic material used in motor construction are factors to define motor power rating. With constant supply power frequency, higher voltage causes higher V/Hz ratio and higher flux. With constant supply voltage, lower supply frequency would cause higher V/Hz ratio and higher flux. Higher flux increases the motor torque capability. When motor operates at higher V/Hz than rated, the overfluxing occurs, which may cause saturation of the stator and rotor magnetic core. Saturation causes overheating and can lead to motor failure. When motor operates at lower V/Hz than rated, the magnetic flux is reduced. Reduced flux reduces the torque capability and affects the motor ability to handle the load.

Drivers Chapter

7

339

When motors are supplied directly from the power network, the supply power frequency is constant, while voltage and current change during motor starting. During motor acceleration to synchronous speed (synchronous motors) or close to synchronous speed (induction motors), the current would initially rise to multiple times the rated current and cause voltage drop. Lower voltage while supply frequency is constant means lower V/Hz ratio and lower flux which affects the torque. Once the motor accelerates, the voltage recovers to close to rated value and the torque available at the motor shaft is at the rated value. The speed of the motor is then constant and synchronous (synchronous motors) or close to synchronous (induction motors). With motors connected directly to power network, the speed is dictated by the fixed network frequency and cannot be controlled. To manage the speed when necessary, additional mechanical systems are used: dampers, valves, gear boxes, brakes, etc. Mechanical systems reduce the overall system efficiency. In addition, as explained previously, induction motors consume reactive power, so maintaining the power factor may be a challenge with induction motors. Synchronous motors do not cause issues with the power factor, they can actually help. There are four categories of challenges with motors connected directly to the power supply network: high starting current, torque control, speed control, and power factor (only with the induction motors). One of the effective ways to address the challenges is to use VFDs. When VFDs are used, the drive is supplied from the power network, and the motor is supplied from the drive. VFDs control the motor speed and motor torque by controlling the frequency and magnitude of voltages and currents supplied to the motor. Each VFD has three sections: rectifier, filter with energy storage, and inverter. Typical conceptual configuration is shown in Fig. 7.22. Rectifier takes the fixed frequency and magnitude voltage sinusoid from the grid and rectifies it into DC waveform.

Rectifier

Filter

Inverter

A B C

M

Input

Controller

FIG. 7.22 Typical VFD configuration.

340 SECTION

II Types of Equipment

Filter takes the DC waveform from rectifier and provides almost pure linear DC. Energy storage is used to support instantaneous energy balance. If with balanced three-phase load, the total power remains constant from instant to instant and with the ideal converter, the energy storage would not be required. In practice, converters require energy storage to store sufficient energy to supply the motor during the brief intervals when load power is greater than the input power. Capacitors and inductors are used for energy storage. Inverter inverts the DC power back to AC through a set of electronic switches (MOSFET (metal-oxide semiconductor field-effect transistor), IGBT (insulated-gate bipolar transistor), IGCT (integrated gate-commutated thyristor), GTO (gate turn-off thyristor), etc.). These switches, by opening and closing at certain speeds and durations, can invert DC and recreate output currents and voltage waveforms that mimic sinusoidal AC waveforms. The motor is then supplied from the output of the inverter. The output waveforms are pulse width modulated (PWM) waveforms. They are called PWM waveforms because they are created by multiple pulses of the switches at short intervals. The magnitude and frequency of PWM voltage waveforms are adjustable. By varying the time, the pulses are on and which switches are firing, the frequency can be increased or decreased. By changing the width and duration of the pulses, the average voltage to the motor can be increased and decreased. Typical PWM waveform with sinusoid being approximated is shown in Fig. 7.23. With an induction motor used as an example, the induction motor can run efficiently only at close to synchronous speed of the rotating field. The speed control requires continuous variation of the rotating field speed, which requires variation of frequency. When the inverter output voltage at each inverter output frequency is controlled so that the V/Hz ratio is kept constant up to the rated speed, a family of torque-speed curves can be derived similar to Fig. 7.24. Point “a” in Fig. 7.24 corresponds to no load torque and no load speed at inverter supplied frequency of 25 Hz. From no load in point “a” to full load

Yd Waveform being approximated 0

–Yd FIG. 7.23 Typical PWM waveform with sinusoid being approximated.

t

Drivers Chapter

7

341

Torque

b

10 Hz 5 Hz

c

25 Hz

37.5 Hz

50 Hz

a Speed

FIG. 7.24 VFD Supplied induction motor torque-speed curve with voltage and frequency variation and constant V/Hz ratio.

in point “b,” the speed will drop slightly. If it is required to maintain the constant speed from point “a,” the VFD control would raise the frequency so that the full load operating point moves to point “c.” The VFD control would also raise the voltage proportionally to the frequency increase, to maintain the constant V/Hz ratio at the full load, and thus maintain the full load torque. From Fig. 7.24, it can be observed that the pullout torque is constant at all points below the rated speed, except at low frequencies. At low frequencies, the pullout torque is reduced because of the effect of stator resistance. As the frequency approaches zero, the voltage drop due to stator resistance becomes important, and flux reduction that causes the torque reduction becomes prominent. This effect is known and easily mitigated by low-speed voltage boosting: increasing the V/f ratio at low frequencies to restore the flux. Fig. 7.25 shows a typical set of torque-speed curves for a drive with low-speed voltage boosting. Beyond rated speed, V/Hz ratio cannot be kept constant anymore because voltage cannot increase beyond rated voltage of the motor to avoid motor Torque

12.5 Hz

25 Hz

37.5 Hz

50 Hz 62.5 Hz 75 Hz

87.5 Hz Speed

FIG. 7.25 VFD supplied induction motor torque-speed curve with voltage and frequency variation, constant V/Hz ratio up to rated speed and low-speed voltage boosting.

342 SECTION

II Types of Equipment

insulation breakdown. The increase in frequency beyond rated frequency is possible and will produce higher speed but with voltage kept at rated voltage, and consequently reducing V/Hz ratio, the flux density will reduce and the torque will reduce. The advantage of VFD supplied motors is that the motor can supply the same maximum torque from zero speed to rated speed. This area of the motor torquespeed characteristic is called “constant torque” area. Continuous operation at peak torque is not done in practice because of the heat limitations. The upper torque limit equal to motor rated torque is usually set in the controller. With VFD supplied motors and with their availability of high torque at low speeds, the starting problems common to fixed frequency operations (initial high slip, high starting current, voltage drop, and torque reduction) are avoided. The VFD-driven motor starts with low frequency, which is gradually increased. The slip speed of the rotor is always small and the rotor continuously operates in the optimum torque condition. Rated torque is available at low speeds and starting current does not exceed the rated full load current. The motor can start from a week power supply system without causing voltage disturbances in the supply network. As mentioned previously, the VFD-driven motor can develop any torque up to rated torque at any speed up to rated speed. This area is called “constant torque” area. Above rated speed, V/Hz will reduce because voltage is kept constant at rated motor voltage, stator, and rotor current are also kept constant and speed and frequency are increasing, so the flux density will reduce and the torque will reduce inversely with the frequency. This area in the motor torque-speed characteristic is called “constant power” area. Constant power area is up to approximately twice the rated speed. Beyond constant power area is the high-speed area where current limit coincides with the pullout torque limit, which reduces inversely with the square of the frequency, so the constant power cannot be maintained any further. Constant torque, constant power, and high-speed areas are shown in Fig. 7.26. Torque

Constant torque region

Constant power region

High-speed region

Speed FIG. 7.26 VFD supplied induction motor torque-speed curve in constant torque, constant power and high-speed area.

Drivers Chapter

7

343

In VFD supplied motor applications, it is important to note that torque-speed curves show the torque the motor can produce for each frequency, but not for how long and if motor can operate in each condition continuously. If in a VFD supplied motor application, a standard induction motor is used, heat limitations need to be taken into consideration. Standard industrial motor is usually an enclosed with an external shaft mounted fan which blows air over the finned external case. The standard design and motor cooling is for the continuous operation for the fixed network supplied frequency and rated speed. When standard industrial motor operates connected to a VFD which produces low frequency and runs the motor at low speed, the motor cooling becomes an issue. The motor will be capable to produce rated torque at low speed, but in those conditions, it will operate at higher temperature which may significantly impact the service life of the motor or cause overheating and motor failure. When motor is used in VFD applications, it is important to specify operating scenarios, design the cooling accordingly, and use motors suitable for inverter duty. Other than cooling, there are other considerations that must be considered in the design when VFD-driven motors are used, such as impact of the harmonics from the VFD to the network, cable configuration and sizing from the VFD to the motor, etc.

Steam Turbines A steam turbine extracts thermal energy from pressurized steam and converts this energy to mechanical work in the form of rotary motion of an output shaft. Following Fig. 7.27, the potential energy of falling water at a hydroelectric plant is analogous to the potential energy in pressurized steam as used fuel a steam turbine. Using a Mollier Chart, the thermodynamic characteristics of a steam turbine can be determined by following the theoretical and actual steam

FIG. 7.27 The potential energy of falling water is analogous to potential energy in pressurized steam.

344 SECTION

II Types of Equipment

expansion lines, from the turbine inlet to the turbine exhaust. Theoretically, a perfect turbine would have isentropic expansion. Knowing the enthalpy (or total heat energy) change for the perfect turbine, its theoretical steam rate (TSR) can be determined. Also, knowing the enthalpy change from the actual turbine operating steam conditions, we can determine the actual steam rate (ASR). From these steam rates, we can determine the overall turbine efficiency. Steam turbines (Fig. 7.28) have been used in industry for well over a century for many purposes including driving of gas compressors, electrical generators, pumps, ship propulsion, and general drive line applications. Steam turbines have a wide operating speed range which lend themselves to being ideal drivers for many styles of gas compressors including both dynamic (radial centrifugal and axial centrifugal) and positive displacement (rotary and reciprocating).

Types Steam turbines can be categorized including: By steam supply conditions—Descriptors include very high-pressure (VHP), HP, medium-pressure (MP), and LP steam. For example, VHP steam typically indicates steam pressures over 14 MPa as well at temperatures over 813 K. By extraction—Extraction is term used to indicate that the steam turbine exports steam to a process or steam header. There are uncontrolled and controlled extraction designs. Uncontrolled extraction typically extracts steam from a predetermined connection in casing with the extraction pressure varying with turbine load. Uncontrolled extraction is typically used in reheat applications at power plants. Controlled extraction designs extract steam out of a similar casing connection, however, directly following is extraction connection the back-end steam flow is metered through a set of governor valves or a control valve which modulate based on an extraction pressure set point. For example, typical process may require a VHP inlet steam turbine to extract HP or MP steam to a header to supply lower power steam turbines in the process (Fig. 7.29). Induction—Induction is a term used to indicate that a side stream of steam is inducted in through a controlled or uncontrolled connection on the steam turbine. The manner of induction is similar to that of an extraction design and in certain cases, a steam turbine may start-up by inducting steam through the extraction connection before VHP steam is produced in the process (Fig. 7.30). Exhaust conditions—steam turbines either exhaust to a pressure above or below atmospheric pressure. If above atmospheric pressure, the steam turbine is known as a back pressure design (sometimes topping turbines). In these cases, the header pressure that the turbine exhausts to is fixed. For example, the turbine may supply steam to the MP or LP header (Fig. 7.31). If the turbine exhaust pressure is below atmospheric pressure the design is known as condensing. Typical applications exhaust to a condenser which

Rocker arm bearing

Lubrican connection

Governor linkage assembly Valve stem and packing Valves, seat, and bar assembly Breather cap Bearing Journal housing bearing Bearing housing deflector Shaft end with coupling bolt pattern

Gland packing assembly case

Steam chest

Turbine case

Exhaust end packing gland assembly

Breather cap Interstage shaft seals

Journal bearing Thrust bearings

Steam end packing gland assembly

Bearing housing end seal

Oil drain

Casing drain

Gland seal leak off

345

FIG. 7.28 Cross section of steam turbine.

Stream end flexible support

7

Steam exhaust

Bearing housing end seals

Bearing housing Oil drain

Drivers Chapter

Rotor

346 SECTION

II Types of Equipment

FIG. 7.29 Extraction of steam in a steam turbine.

FIG. 7.30 Induction of steam in a steam turbine.

High pressure steam Reducing control valve

Turbine Noise

Load Load: driven equipment

Low/mid pressure steam

FIG. 7.31 Example of steam turbine exhausting to above-atmospheric pressure.

condenses the steam and reclaims the condensate for the boiler feed water process (Fig. 7.32). Casing or shaft arrangement—Casings designs sometimes indicate the type of turbine. For example, a casing with double shell or double casing design

Drivers Chapter

7

347

High pressure steam Load: driven equipment

Turbine

To boiler

P

Load

Surface condenser

FIG. 7.32 Example of steam turbine exhausting to subatmospheric pressure.

typically indicates a high inlet pressure and temperature. (See picture for VHP inlet/extraction/back pressure/double casing example.) Other designs have multiple casings known as tandem casing arrangements or compound designs where HP, MP, and LP turbine stages are contained in separate casings. In these designs, the rotors may or may not be coupled together and rotating at same speed (Fig. 7.33). Blade design—Steam turbine blades and nozzles are primarily impulse or reaction designs (Fig. 7.34). With impulse designs, the pressure drop across each stage is taken over the fixed row. This corresponds to zero reaction of rotating row and with this the rotor thrust is minimal. These designs are usually a disk/wheel and diaphragm construction. Fig. 7.34 shows the range of operation for each classification of turbine stage. This chart shows wide usage of impulse blades. This type of blading is generally less efficient however the design less susceptible to the effects of leakage flow around the stage. With Reaction designs steam expands in both the fixed and rotating blades. The pressure drop across each stage is split between fixed and rotating rows. This split results in higher reaction imparted on the rotating rows and therefore larger thrust loads which are counter balanced using balance pistons. Single or multistage—This indicates the number of stages of expansion the steam takes from high to LP. Single-stage turbines are mostly used for general

FIG. 7.33 Example of casing arrangement.

II Types of Equipment

Steam in

348 SECTION

Nozzle vones fixed

fixed

Moving buckets

atio Fixed

P

Moving

Steam pressure Steam speed

V A Velocity-compounded (curtis) stage and reaction stages

B Pressure-compounded (rateau) stage

C Reaction stage

FIG. 7.34 Example of (left) impulse and (right) reaction steam turbine blades.

purposes including pump and fan drives. Multistage designs are used when higher HP and better efficiencies are needed (Fig. 7.35). Nature of steam generation—The source of steam often is the descriptor of the steam turbine design. Examples include terms like “geo-thermal” which typically indicates a high flow LP turbine design. In most industrial processes, the steam system is in some way integrated with the process. This is because the heat used to generate the VHP steam comes from the process. In these cases, alternate steam sources are used to get plant process initially running.

Application as a Compressor Driver The variable speed “Mechanical drive” or “industrial drive” designations are often used to describe the type of steam turbine best suited for driving compressors, pumps similar rotating equipment. Fig. 7.36 represents the range that typical mechanical drive turbines must span for an ammonia process. The design of this style steam turbine is generally known for its simplicity, reliability, ruggedness, and flexibility in terms of operation. Mechanical drive steam turbines are typically multistage units and can be either straight through flow or extraction/induction designs. They typically use a range of steam

Drivers Chapter 7

349

FIG. 7.35 Example cross-section of steam turbine.

350 SECTION

II Types of Equipment

FIG. 7.36 Typical range of mechanical drive turbine for an ammonia process.

conditions up to 14 MPa and 813 K and ranging in power to 69 MW with speeds up to 14,000 rpm. This class of steam turbine can be direct coupled to the compressor or coupled through a speed increasing gear to meet compressor speed requirements. Variable speed requirements make this turbine design well suited to compressor drive use as a wide variety of speed ranges can be supplied.

Major components Steam chest—contains the valve rack and governing valves. This portion of the turbine casing sees steam inlet conditions and admits steam to the turbine nozzles. Designs can either be integral to main turbine casing or separated and connected with interconnecting piping with main casing. Steam chests can have a single-valve or multivalve schemes. Single-valve applications typically either have double seated valve designs or use venture valve designs. Multivalve designs typically use venturi valves (Fig. 7.37). Nozzle ring and diaphragms—A nozzle ring contains a series of nozzles arranged on a base diameter. The purpose is to direct steam to the rotating buckets of a rotor. The nozzle ring with a predetermined admission percentage is generally associated with the “control stage of the turbine directing steam to first rotating row of the rotor.” Diaphragms similarly contain a circle of nozzles which direct steam to the downstream rotating blades. Diaphragms contain attachment point for an interstage labyrinth seal which seals against the shaft between disks (Fig. 7.38). Valve rack—The term “valve rack” typically describes the mechanical workings related to admitting steam to the nozzles of a steam turbine.

Drivers Chapter

FIG. 7.37 Illustration of steam chest containing valves.

7

351

352 SECTION II Types of Equipment

FIG. 7.38 Nozzle ring and diaphragms.

Drivers Chapter

7

353

There are different categories of valve rack designs depending on the amount of flow being passed, degree of control precision required. Major types include: Cam—A cam valve rack is known for its individually controlled venturi valves where each valve is lifted based on cut in points sequenced from cams mounted on a cam shaft and cam followers intern connected to the valve stems (Fig. 7.39). Bar lift—A bar lift arrangement consists of multiple venturi valves mounted in a lift bar within the steam chest. The bar is lifted via lift rods which protrude through the steam chest cover thru a set of lift rod seals. The flow sequence is controlled by the distance the valves hang from the bar (Fig. 7.40). Casing—The steam turbine casing is typically cast steel integral to the steam chest. The casing has mounting provisions for nozzle rings and diaphragms. The bearing housings are bolted to the casing. Steam turbine casings come in single double casing configurations depending on inlet temperature and pressure. Double casing designs are typically employed so that the inner casing is subjected to the high steam temperatures and pressures while the outer casing and sealing joints are exposed to lower values. This helps with thermal stresses and sealing capability (Fig. 7.41). Rotor—The rotor is the rotating element within the turbine. The rotor has disks which have attachment provisions called roots in which the rotating buckets or airfoils are mounted. The disks are either integral or machined from a shaft forging or, individually shrunk on a shaft. The rotor has machined journal bearing surfaces, a thrust disk, or thrust runner, and typically houses the rotating trip arrangements. The rotor drives the driven equipment via the shaft end which can have either an integral machined coupling hub, or removable coupling hubs. Shaft end fit designs include tapered and straight fits with and without shear keys as well as hydraulic fits (Fig. 7.42).

FIG. 7.39 Cam valve rack.

354 SECTION

•

II Types of Equipment

Multivalve lift bar arranagement 1.803” ACCUMULATED VALVE TRAVEL 1.064”

0.330” 0.060”

0.697”

1.392”

Position

1.5/8” NO. 1

2.1/16” NO. 2

1.5/8” NO. 3

2.1/16” NO. 4

2.3/8” NO. 5

Order of opening

NO. 4

NO. 2

NO. 1

NO. 3

NO. 5

FIG. 7.40 Bar lift arrangement.

FIG. 7.41 Steam turbine casing.

Drivers Chapter

7

355

FIG. 7.42 Steam turbine rotor.

Thrust disk—The thrust disk or thrust runner provides an axial surface for the thrust bearing pads to ride against. The thrust disk can either can be an integrally machined component, or have interference fit to the rotor (Fig. 7.43). Bearing housings—Bearing housings contain essential components including journal and thrust bearings, overspeed trip hardware, governor drive hardware, and various instrumentation like axial position probes, magnetic pickup probes, grounding brushes, key phase probes, and bearing load cells (Fig. 7.44).

FIG. 7.43 Thrust disk.

356 SECTION

II Types of Equipment

FIG. 7.44 Bearing housing.

Journal bearings—Journal bearings provide a means to support the rotating rotor radially on an oil film. Bearings are typically made of steel backing with a layer of lead-tin babbitt to help with wear and reduce journal scoring and damage during start-up and low load conditions. Major journal bearing designs include liner, tilting pad, and spherical seat tilting pad, flexural pivot. Also, there are different styles of lubrication including flooded design and direct injected designs. Bearing instrumentation includes use of thermocouples and RTD (resistance temperature detectors)’s to measure bearing metal temperature. In addition, bearing housings and drains are outfitted with thermowells to measure specific bearing oil throw-off temperatures (Fig. 7.45). Thrust bearing—Thrust bearings serve to manage a steam turbine rotor axial thrust by providing a lubricated surface for the rotor to thrust against. Two main categories of thrust bearings designs include, self-leveling designs which consist of a base ring, leveling links, and individual pads with pivot on back and fixed ring designs with tapered lands which are often found in with older designs. The self-leveling feature aids in matching pads with the changing thrust disk location and perpendicular relationship to steam turbine axis. Bearings are typically made of steel backing with a layer of lead-tin babbitt to help with wear and reduce journal scoring and damage. Also, there are different styles of lubrication including flooded design, and direct injected designs. Bearing instrumentation includes use of thermocouples and RTD’s to measure bearing metal temperature as well as load cells to measure thrust load (Fig. 7.46).

Selection The selection process for a mechanical drive steam turbine depends on the type of process, the available steam conditions, the required power, and the speed range. In addition, there may be steam rate and possibly casing connection size requirements. Basic steps for ideal back pressure turbine: 1. Establish a TSR (from inlet, exhaust steam conditions, and required power). 2. Determine isentropic heat drop across turbine.

Drivers Chapter

FIG. 7.45 Journal bearings.

7

357

358 SECTION

II Types of Equipment

FIG. 7.46 Thrust bearing.

3. Divide the total isentropic heat drop by the optimum heat drop per stage (typically available from OEM). 4. This value represents the number of stages needed to make power at isentropic conditions. 5. Assume efficiency and multiply by the total isentropic heat drop to obtain the actual heat drop. 6. Subtract actual heat drop from isentropic heat drop to obtain exhaust enthalpy. Other checks needed for a selection include: l

l

l

l l l l

l

Blade mechanical checks—airfoil profile/nozzle selection, blade material selection, Goodman/Campbell review, root design selection, momentary speed limit check. Moisture erosion—erosion protection for rotating and stationary components. Casing design—HP and LP casing designs based on the pressure and temperature limitations. Casing connections—velocity limits and pressure rating. Shaft end size—check to determine torque capability of shaft end. Rotor design—lateral/torsional critical speed analysis. Journal bearing—journal bearing load review, bearing metal temperature prediction. Thrust check—thrust prediction across all operating points maximum thrust bearing limitations.

Performance Once installed, a steam turbine’s performance is subjected to degradation overtime for a number of reasons. Performance monitoring is an important part of operating a plant with steam turbine drivers. The end result of a performance monitoring program is to obtain the lowest sustained heat rate possible. As performance deteriorates, the increase in heat rate is a measure of total degradation of the components in the system. Heat rate itself has a diagnostic value of indicating the magnitude of deterioration over a time period. Heat rate itself does not pinpoint the source of the degradation.

Drivers Chapter

7

359

ASME (American society of mechanical engineers) PTC-6 performance test codes outline the requirements for conducting a code performance test and the associated percentage error to be applied to a given type of instrument. The majority of the time the performance review does not resemble a code test therefore the ASME standards are essentially a set of rules to follow to take account for all nonideal test instrumentation and nonstandard test conditions. For example, the primary flow measurement alone has many contributing factors including: l l l l

the presence and location of straighteners, calibration frequency of instruments, type of instrument and its accuracy, and singular or redundant instruments for each point.

Once summation takes place for each instrument and nonstandard condition the overall test uncertainty is then calculated and applied to final power or flow value. These and other criteria are evaluated and summed to determine overall error to apply to the calculation results in order to close the gap between test data and predicted data. Primary instruments required for a typical steam turbine performance analysis include: l l l l l l l l l

Inlet steam pressure and temperature, inlet steam flow, and valve position Steam chest pressure and ring pressure First-stage pressure Extraction pressure and temperature, extraction flow rate, valve position Second section first-stage pressure Exhaust pressure and temperature and condensate flow rate Speed Torque meter output (if available) Leakage flow rate (if available)

Operation One of the more critical aspects of steam turbine operation is related to the following start-up procedures. HP and temperature units require proper warmup in order to avoid high thermal stress across thick portions of casing and also to prevent uneven thermal growth which can lead to internal alignment issues. Cold and Hot start-up maps are typically provided by OEM for user to follow in order to maintain successful operation between outages. As seen with attached maps, there are hold points at certain speeds, before ramping through critical speed bands, and prior to operating speed in order to identify risks and minimize damage. Cold start is most critical as the chances of damage to turbine are greatest. Note the time difference between cold and hot maps. It can be seen that on a hot start the hold times are much shorter and ramp rates are higher (Figs. 7.47 and 7.48).

360 SECTION

II Types of Equipment Example cold startup curve

6000

Critical speeds Lateral 2SQV9 2900 - 3000 rpm (calc.) 5M6-4 2600 - 2800 rpm (calc.) Torsional 1254, 11852, 14599, 25146 rpm Calculated

5000

MCS Rated speed

Speed values Min. required hold period Low idle#1 500 rpm 10 min 2000 rpm 120 min Low idle#2 Critical speed band 2400 - 3200 rpm Minimum speed 3600 rpm 70 min High idle#1 Min speed 3856 rpm High idle speed MCS 5061 rpm OVS 5567 rpm 4820 rpm Rated

4000 Speed (rpm)

Check overspeed trip

3000

Critical speed band

Low idle speed #2

2000

Governor valve control

T&T valve control

Establish light vacuum by enabling ejectors. Then, open TTV Check manual trip

1000 Low idle speed #1

0

–1.0

Enable GC supply seal steam increase vacuum

0.0

1.0

2.0

3.0

4.0

5.0

Time (h)

FIG. 7.47 Cold start-up curve.

Example hot startup curve (5–30 min after trip) 6000 0 Speed values Low idle#1

5 min 2 min

High idle#1 3600 rpm

3 min

Low idle#2 Critical speed band 2400 - 3200 rpm

5000 Check process

Min speed

Speed (rpm)

4000

Min. required hold period

500 rpm 2000 rpm

3856 rpm

Minimum speed

Stop

High idle speed

3000

Critical speed band

Stop seal steam

2000

Low idle speed #2 T&T valve control

Establish light vacuum by enabling ejectors. Then, open TTV

1000

Governor valve control

Disable GC Low idle speed #1

Break SC vacuum

0 0.0

Enable GC supply seal steam increase vacuum

Restart preparation

0.1

FIG. 7.48 Hot start-up curve.

0.2

0.3

0.4

0.5

0.6

0.7

Drivers Chapter

7

361

Auxiliaries Sealing steam system—These systems are used with condensing turbines in order to seal HP, LP packing from atmospheric pressure during start-up and low load conditions. Gland leakoff system—These systems are used with all turbines in order to minimize outward leakage of steam from HP, LP packing to atmosphere and toward bearing housings. Systems often include a gland condenser and steam ejector in order. Negative pressure is obtained in packing drains in order to encourage flow toward gland condenser, as shown in Fig. 7.49 of a typical system for a condensing turbine application. Lubrication system—System supplies lubrication oil, seal oil, control oil, and/or trip oil to rotating equipment. System typically includes pumps and regulation system and reservoir. Trip valve—Mounts ahead of steam chest connection is an integral part of steam chest. Trip and trip and throttle valves serve to isolate steam from turbine during a trip condition. T&T valves also serve to throttle steam to turbine for start-up (Fig. 7.50). Nonreturn valve—Use with extraction turbines to eliminate reverse flow from extraction header if internal pressure in casing falls below extraction pressure. Surface condenser—Used with condensing turbines. Condenses exhaust flow as well as leakoff flows from steam turbine. Venting equipment—Part of surface condenser system. Hogger, first- and second-stage ejectors used to remove non-condensible’s (air) from exhaust steam flow during start-up and operation. Full flow relief valves—Used in exhaust line of back pressure turbines to protect exhaust casing from overpressure condition. Rupture disks/atmospheric relief valves—Used with condensing turbines to protect exhaust case from overpressure condition. Governor control systems—Range from simple mechanical and hydraulic speed control applications such as fly-weight governors and Woodward PG governors, to sophisticated electronic governor systems which can be programmed to control speed, or a process variable. Since the early 1990s, industry has trended toward supply of new turbines with electronic speed control as well as electronic overspeed trip control.

Steam Turbine Maintenance/Reliability Steam–Steam purity guidelines are followed by customers who are concerned with the quality of steam used for steam turbines, such as shown in Fig. 7.51. End-user reliability departments are interested in instrumentation including radial vibration, phase angle, bearing metal temperatures, bearing oil throw off temperatures, bearing load cells, oil analysis, rotor thrust position, grounding brush current, and steam quality. These and many others are key points of

362 SECTION

II Types of Equipment

FIG. 7.49 Gland leakoff system.

Drivers Chapter

7

363

FIG. 7.50 Trip valve.

Continuous

Startup

Drum Once through

0.3 0.2

1.0 0.5

(ppb, max.) (ppb, max.) (ppb, max.) (ppb, max.) up to 800 psig 801–1450 psig 1451–2400 psig over 2400 psig (ppb, max.)

20 20 3

50 50 10

20 10 5 3 10

20 10 5 3 10–30

Conductivity Micromhs/cm at 25°C

SiO2 Fe Cu Na + K

CL

FIG. 7.51 Steam purity guideline example.

interest that are typically monitored, recorded, and trended in order to track reliability and predict issues with operation, and plan for future outages. A term called mean time between failure (MTBF) is a metric that is analyzed for each piece of rotating equipment to maximize run times and minimize the cost of outages. The main points of start-up center around the warm-up of the casing and rotor. Condensate must be drained from the casing during warmup such that drains are free of liquid. This is an indicator that casing is warm enough to proceed to next step in start-up. On HP and temperature units, casing differential thermocouples are supplied so operations have an indicator that casing is heat soaked (Fig. 7.52). For example, a typical acceptable differential temperature between deep and shallow thermocouples is on the order of 82°C. Higher values indicate outside of casing temperature is too far from inner casing wall temperature and more heat soaking is needed before proceeding to next start-up step.

364 SECTION

II Types of Equipment

FIG. 7.52 Thermocouples to indicate casing temperature.

FIG. 7.53 Illustration of water ingestion into HP packing.

Liquid ingestion through main turbine inlet or intake of liquid into HP or LP packing during start-up is a source of rotor bows, high vibration, and internal seal damage (Fig. 7.53). Care must be taken for turbines with sealing steam, that the source of external steam is dry and hot.

Performance Degradation The challenge with monitoring steam turbine performance is evaluating the power balance between the absorbed compressor power and the produced turbine power. The difference between performance review of turbine-generator and a compressor drive applications is that with TG (turine-generator) applications there is a kW output to compare directly with. A typical back pressure turbine, for example, may have a superheated inlet and exhaust. In this case, the steam temperature before and after the unit is very important and plays a critical role in calculating an accurate enthalpy drop. If the steam rate in such an

Drivers Chapter

7

365

example had deteriorated overtime additional instrumentation must be reviewed to determine what component within the unit is contributing to the inefficiency. Data including steam conditions before the trip valve, steam chest pressure, ring pressure, first-stage pressure, and V1 position are areas to review to further pinpoint internal issues. Also, data such as vibration, thrust, and journal bearing metal temperatures are important. In the case of a typical condensing turbine the complication of knowing the exhaust steam quality is the challenge. Only knowing the exhaust pressure and temperature does not indicate steam quality. Condensate flow serves to validate the inlet steam flow value which closes the uncertainty to a point. The following is a list of possible sources of inefficiency related to steam turbines:

Expanders Turboexpanders provide the most efficient solution when it is required to reduce the pressure of a fluid stream. They have been part of modern gas processing plants since the late 1960s. Turboexpanders extract energy from the fluid flow, thereby producing power, letting down pressure, and dropping down the fluid temperature. The extracted shaft power, which would otherwise be permanently lost, can be used to drive a compressor, pump, or a generator. The work recovered from the expansion is supplemental and increases the plant thermal efficiency. The expander wheel is used to extract energy from the gas flow. The extracted shaft power can be used to drive a compressor, a pump, or a generator.

Expanders in Cryogenic Applications In the second half of the 20th century, pioneers such as Dr. Judson Swearingen introduced expander-compressors for use in cryogenic natural gas plants. Incentives to increase thermal efficiencies of plants, minimizing feed usage in liquefied natural gas (LNG) plants, and reducing CO2 emissions have driven the development of many plant process designs incorporating turboexpanders. In cryogenic applications, most of the economic benefit created by turboexpanders lies with the fluid enthalpy drop in the turboexpander and the associated temperature drop. The cooler fluid leaving the turboexpander helps to increase the thermal efficiency of the plant by saving on the cooling capacity. In addition, the turboexpander shaft power extracted from the fluid can be used to reduce the plant input power. Each of these serves to reduce the refrigeration cycle-specific power consumption and hence provides economic benefit to the refrigeration process. In cryogenic expander systems, a Joule-Thomson (JT) valve is usually provided in parallel to the expander. The parallel JT valve automatically adjusts as necessary to prevent an overall process shutdown, if the expander drops “offline” for any reason.

366 SECTION

Internal fouling of steam path

II Types of Equipment

Foreign object damage to steam path

Increased pressure drop across trip valve, inlet strainer, governor valves, or nozzle ring. As seen in the following figure, as fouling increases the differential across the turbine also increases. This leads to inefficiency, more steam flow required to make power, as well as increased thrust force the thrust bearing is subjected to

Drivers Chapter

Opened up seal clearances in HP/LP packing Opened up internal shaft seal or shroud seal clearances Internal bypass of steam across diaphragms or around nozzle ring Wear

7

Typical wearing parts within a turbine include:

367

Continued

368 SECTION

l l l l l

l l l l l

Nozzle or blade erosion Governor linkage pinned joints Valve rack pinned joints Valve stem and lift rod seals Labyrinth seals in HP, LP packing, and interstage locations (carbon rings) Tip seal (shroud seal) seal strips Oil baffle labyrinth Governor mechanical drive gears Governor drive couplings Journal and thrust bearing babbitt faces

II Types of Equipment

—cont’d

Drivers Chapter

7

369

Hot Gas Expanders The idea of converting the energy in combustion exhaust gas to power has been in practice for a century. Energy costs, environmental regulations, emissions control, and carbon management have driven interests in employing hot gas expanders in modern plants. There are fundamental challenges to design a reliable expander operation: the extreme high temperatures, chemical corrosion, and kinetic erosion caused by the hot, particulate laden flue gas caused the rapid erosion of critical rotating and stationary components such as the rotor disk, blades, and stator vanes. Drastic thermal differentials also imposed substantial stress on the structural elements of the machine. Corrosion is another critical consideration in expander operations. Corrosion is related to the high temperatures at which expanders operate. The effects of creep and gaseous corrosion increase at elevated temperatures. The nature of the corrosive attack is primarily influenced by the feedstock and the additives injected during the cracking process. If not addressed, hot corrosion can lead to rotor blade or disc failure (Fig. 7.54).

Exhaust casing Expansion joint

Inlet casing

Gas inlet

Shaft

Trunnion support

Bearing support pedestal

FIG. 7.54 Basic components of a hot gas expander (Hydrocarbon Engineering, August 2013).

370 SECTION

II Types of Equipment

This harsh operating environment poses special challenges for expander design and performance, reliability, and maintainability. Additional operational considerations are increasingly demanding process parameters, extended maintenance cycles, and more stringent environmental regulations.

Radial Inflow Turboexpanders In radial inflow turboexpanders, energy is transferred from the fluid to the wheel in passing from a larger area at the wheel tip to a smaller area at the wheel eye. Fig. 7.55 shows a schematic layout of a radial inflow turboexpander. The fluid discharging from the wheel may have a considerable kinetic energy (high velocity C4). A diffuser is normally incorporated to recover the kinetic energy, which would otherwise be wasted. In Fig. 7.55, the velocity triangles are shown to clarify that the inlet relative velocity, W3, is radially inward, and the absolute flow at rotor exit, C4, is axial. This configuration of velocity triangles is popular for radial inflow turboexpanders. In a turboexpander, there are three steps that convert energy: 1. The potential energy of the fluid is converted into velocity in the inlet guide vanes. This conversion is approximately 95% efficient. 2. The potential energy remaining is converted to mechanical power in the turboexpander wheel. 3. The exit velocity of the gas is relatively high and is decelerated in an exhaust diffuser.

Thermodynamics of Gas Expanders The load applied to the turboexpander serves as a sink for the power extracted by the turboexpander. Fig. 7.56 shows three reference expansion processes between the higher pressure P0 and the lower pressure P5. It shows the minimum discharge temperature happens for an isentropic process, when energy is extracted from the fluid stream. It also shows the discharge temperature for an isenthalpic expansion process (e.g., across the JT valve) is higher than both the actual and isentropic processes. The temperature difference between 5a and 5h (e.g., ΔT ¼ T5h  T5a) signifies the additional temperature drop, which can be gained if a turboexpander is used in place of a JT valve. The turboexpander efficiency can be defined in terms of static-to-total temperatures when the temperature drop is the main purpose for utilizing a turboexpander. ηexp ¼

Tð0Þ  Tð5aÞ Tt ð0Þ  Tt ð5sÞ

Drivers Chapter

FIG. 7.55 Layout and velocity diagrams of a radial inward-flow expander.

7

371

372 SECTION

II Types of Equipment

0

pe Te m

Pressure

Upstream of the IGV

py tro

rat u

En

re

P0

P5 5s

5a

5h

Downstream of the diffuser

Enthalpy FIG. 7.56 Isentropic expansion process (0–5s), actual expansion process (0–5a), and isenthalpic expansion process (0–5h).

This definition of turboexpander efficiency may be used to determine how efficiently the fluid energy at the inlet nozzle is extracted by the expander to achieve cooler fluid at the expander exit.

References [1] C.B. Meher-Homji, M. Chaker, A.F. Bromley, The Fouling of Axial Flow Compressors: Causes, Effects, Susceptibility, and Sensitivity, ASME, in: Turbo Expo: Power for Land, Sea, and Air, Volume 4: Cycle Innovations; Industrial and Cogeneration; Manufacturing Materials and Metallurgy; Marine, 571–590. https://doi.org/10.1115/GT2009-59239. [2] Guideline and Recommended Practice for Control of Torsional Vibrations in Direct-Driven Separable Reciprocating Compressors, GMRC, 2015.

Chapter 8

Upstream Compression Applications Timothy C. Allison*, Avneet Singh† and Joseph Thorp‡ *

Southwest Research Institute, San Antonio, TX, United States, †Solar Turbines, Inc., San Diego, CA, United States, ‡Saudi Aramco Energy Ventures - North America, Houston, TX, United States

Introduction The upstream sector of the oil and gas industry (also referred to as exploration and production) focuses on the identification and production of natural gas from various oil and gas fields. Compression activities in this sector are focused at the well head and gas processing plant, where the gas is exported to midstream processes for transport and storage. These compression applications are shown on the left-hand side of Fig. 8.1 and include gas gathering, gas lift, reinjection, and boost and export compression at processing plants. Subsea compression is a specialized upstream application and is discussed in a separate chapter. Fig. 8.2 shows the general relationship between various upstream processes for oil and gas fields, although many installations incorporate only some of these elements and others may exclude, combine, or rearrange some elements. In general, upstream processes involve multiple stages of compression, separation, and treatment for preparing the produced fluids for sale. This chapter organizes discussion on the compression processes into the three general categories of gas gathering, gas lift/reinjection, and gas processing. Compression duties vary widely among the different upstream applications, and even for a specific application due to the wide variation in oil and gas field production characteristics of composition, pressure, and flow rate. These characteristics vary from field to field but also over the life of a field [2]. This varying service demands flexible compressor performance to accommodate variations in operating conditions. In addition, the economics of the upstream industry require high availability and reliability of compression resources to begin production from new fields as early as possible and minimize both planned and unplanned downtime for existing facilities.

Compression Machinery for Oil and Gas. https://doi.org/10.1016/B978-0-12-814683-5.00008-0 © 2019 Elsevier Inc. All rights reserved.

375

376 SECTION

III Applications

Gas path Electric power generation

Oil field Gas field

Gas gathering

Pipeline compression Gas lift Gas injection

Export compression

Waterflood

Boost

User Fuel compressor

Distribution/ power generation

Storage/ withdrawal

FIG. 8.1 Natural path processes from gas field to user [1]. (Reproduced with permission from ASME.)

FIG. 8.2 Oil and gas field compression.

Gas Gathering Gas gathering activities from a reservoir are broadly classified by the type of well (oil or gas). Associated gas (flash gas) compression is performed at oil wells, and nonassociated gas is compressed at gas wells.

Associated Gas (Flash Gas) Compression Most oil wells produce oil, gas, and water that must be separated at the surface. The separated water may be used for a variety of purposes, and the remaining hydrocarbon mixture is sent to an oil treatment plant for stabilization of the crude oil (removal of all gas). At this plant, the hydrocarbon mixture is processed in a multistage gas-oil separation system with stepped pressure reductions to flash off gas at each stage. During each decompression stage the associated gas (also called flash gas) is removed from the crude oil in a separator until the pressure is ultimately reduced to slightly above atmospheric pressure [1]. Lighter gases flash at higher pressures, so the resulting associated gas streams are at a variety of compositions, pressures, and flow rates. A final step is to heat the crude oil at low pressure and run it through a stabilizer column where hydrogen sulfide and any remaining light hydrocarbons are boiled off and gathered for disposal or use.

Upstream Compression Applications Chapter

8

377

FIG. 8.3 Oil production with flash gas, gas injection, and vapor recovery compressors [2].

These postseparation gas streams are typically recompressed and combined for use as sales gas, gas injection, or gas lift. The stabilized crude oil is sent to storage/sales. Typical discharge pressures for associated gas compressors are about 7–10 MPa. In many cases, water in the gas is removed via glycol dehydration prior to pipeline transport in order to prevent hydrate formation in the lines [3]. An example flash-gas compression system is shown in Fig. 8.3, where the gas is sent to a gas injection compressor. Flash-gas compressor trains often consist of multiple bodies to accommodate the different gas streams. In general, each independent stream needs a control mechanism in order to allow stable compressor operation in response to process variations in individual streams. For centrifugal compressors, it is usually recommended to incorporate a combination of speed control for the highest flow stream and suction throttling or variable inlet guide vanes for lower-flow streams since speed control is the most energy efficient control method. Reciprocating compressors can operate stages at lower flows via speed control, cylinder deactivation, and variable volume pockets or valve unloaders. Recycling is a potential control method both centrifugal and reciprocating compressors, but is not recommended because it does not reduce power consumption at lower flows.

Nonassociated Gas Compression Natural gas wells do not produce oil but may be classified as dry wells that produce only raw natural gas or condensate wells that produce both gas and natural gas condensate (wet gas). Natural gas condensate is typically composed of hydrocarbons (propane, butane, pentane, hexane) and impurities (may include hydrogen sulfide, thiols, carbon dioxide, alkanes, cyclohexane, and aromatics, e.g., benzene). For many condensate wells, the gas-liquid mixture at the well is

378 SECTION

III Applications

passed through a field separator to remove condensate and water. The natural gas liquids separated at this stage must be transported to a processing plant for recovery. A wide variety of gas field configurations exist that produce at different pressures and flow rates, depending on the size and connectivity of gas pockets in the field and age of the reservoir (pressure decreases over time as the reservoir is depleted). The purpose of gas gathering compression is to raise the gas pressure from wellhead pressure to approximately 7–10 MPa for transport to a gas processing plant or sales pipeline. This includes compression to compensate for pressure reductions in field separators or high pressure losses in the well bore due to flow restrictions from liquid condensate formation. Over the age of a field the suction pressure of gas gathering compressors change drastically, with a very low initial pressure ratio of 1.25–1.5 to final suction pressures near vacuum at well closure [4]. A large number of low-pressure, low-flow wells will require a large number of small (hundreds of kilowatt) compressors connecting to centralized compressor stations that combine streams. Larger fields with better connectivity and volume may have field compressor stations with up to 10–20 MW of compression [1]. In addition, over the life a field it may be necessary to restage compressors to higher-head, lower-flow wheels to match reservoir production or even the purchase of additional compressors as supplemental wells are drilled to maintain production. In addition, there is significant interest in wet gas compressors that are robust to operation with condensates in order to minimize field separation and liquid transport requirements. Typical compression system characteristics for gas gathering applications are summarized in Table 8.1.

TABLE 8.1 Typical Characteristics for Gas Gathering Applications Pressure

Suction pressures range from near vacuum (0.3–0.7 MPa) to 6.5 MPa depending on reservoir. Discharge pressures typically 7–10 MPa

Temperature

30–35°C suction to 176–190°C discharge

Fluids

Natural gas with or without liquids, sometimes corrosive components also if associated gas

Compression power

100 kW to 20 MW; Highly variable depending on number and size of reservoirs

Typical machinery

Small reciprocating compressors to larger turbine-driven centrifugal compressors depending on number of wells, flow rate, and reservoir size

Upstream Compression Applications Chapter

8

379

Gas Lift/Reinjection Gas lift and reinjection are two distinct processes for improving production at oil wells. Gas lift focuses on improvement of well performance, and reinjection focuses on improved yield from the reservoir.

Gas Lift A gas lift compressor injects gas in to an oil well to aerate the crude oil, increasing production [5]. The gas is introduced from the surface into the tubing-casing annulus and introduced into the tubing at the desired depth. This gas injection decreases the fluid density in the tubing, thereby increasing the well pressure at the surface. An example gas lift system is shown in Fig. 8.4. The application may be combined with a gas gathering operation, and some offshore operators use the same compressor train to both feed a gas lift service and export compression to feed gas into a pipeline. The gas lift compressor is supplied with produced gas at the surface, typically at relatively low pressure (0.3–0.7 MPa). This gas is compressed to a higher pressure for injection, typically to pressures of 10–12 MPa or even up to 20 MPa depending on reservoir requirements. Higher injection pressures enable gas lift at deeper depths, maximizing the production increase. Thus, gas lift compressors typically require high throughput and a high compression ratio. The typical operating characteristics for gas lift compression are summarized in Table 8.2.

FIG. 8.4 Gas lift system [6]. (Courtesy of Southwest Research Institute)

380 SECTION

III Applications

TABLE 8.2 Typical Characteristics for Gas Lift Applications Pressure

Suction pressures typically 0.3–0.7 MPa, discharge pressures typically 10–20 MPa

Temperature

303–308 K suction to 449–463 K discharge

Fluids

Natural gas

Compression power