Boiler Dynamics and Controls

Copyright © 2013 by F. Paul de Mello.

All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without permission in writing from the copyright owner.

Rev. date: 11/18/2013

To order additional copies of this book, contact:

Xlibris LLC

1-888-795-4274

www.Xlibris.com

Orders@Xlibris.com

141415

Contents

F. P. De Mello   Consulting Engineer

Acknowledgement

Chapter I   Boiler Process Dynamics And Control-Overview

Chapter II   General Principles And Structures In Boiler Controls

Chapter III   Drum Boiler Pressure Effects

Chapter IV   Drum Boiler Feedwater Controls

Chapter V   Fuel And Air Controls For Drum Type Boilers

Chapter VI   Furnace Draft Controls

Chapter VII   Steam Temperature Controls

Chapter VIII   Miscellaneous Loop Controls

Chapter IX   Controls For Once-Through Boilers

Chapter X   Direct Digital Boiler Control

Chapter XI   Modeling From First Principles

Appendix A   Dynamic Systems, Differential Equations—Transient And Steady State Solutions—Operational Impedance

Appendix B   Laplace Transforms26,27

Appendix C   Transfer Functions, Block Diagrams25,30

Appendix D   Analog Computers—State Space—Numerical Methods Of29 Differential Equation Solution

Appendix E   Feedback Control System Concepts25, 30

Appendix F   Notes On Process Control

Addendum   Digital Control Algorithms And Control Tuning

F. P. DE MELLO

Consulting Engineer

Mr. de Mello graduated with BS and MS degrees in Electrical Engineering from MIT where he was elected to Tau Beta Pi and Sigma XI. His academic experience included several test engineering and laboratory assignments with the General Electric Company (GE) between 1945 and 1948.

In 1948, he joined the Rio Light and Power Company in Brazil and over several years held position of increasing technical responsibility in system planning and design studies concerning expansion of the Rio, Sao Paulo, and City of Santos systems.

In 1955, Mr. de Mello joined the Analytical Engineering Section of GE’s Apparatus Sales Division in Schenectady. Here he undertook design and analysis studies of controls of industrial, power apparatus and aircraft power systems, making extensive use of analog computers. In 1959, he was assigned to specialized studies of dynamics of electrical machines, excitation control, prime-mover systems, and overall power systems.

From 1961 to 1969, he conducted and guided extensive research efforts on modeling of dynamics of power systems and power plants for use in advanced boiler and plant control design studies. He made major pioneering contributions in the development of digital computer methods for dynamic analysis and process control design. Of particular note were the development of computer techniques for the simulation of complex boiler dynamics and for the synthesis of multivariable boiler-turbine controls for which he was awarded GE’s Managerial and Ralph Cordiner Awards. He also made significant contributions in the study of electrical machine dynamics, their voltage and governing controls and to the analysis and implementation of system load-frequency controls.

Mr. de Mello joined Power Technologies, Inc., at the time of its formation in August of 1969 as Principal Engineer, Dynamics and Control, and Secretary-Treasurer. He was appointed Vice President-Secretary in 1973. He was a Director of PTEL, PTI’s affiliate in Brazil. From 1974 to 1976, he was project manager for PTI and PTEL in system and design studies for transmission from Itaipu, the world’s largest 800 kV system, and served on the advisory Board of the Study Group for Itaipu transmission. Prior to his appointment as Principal Consultant in 1987 he was Manager of PTI’s Consulting Services Department.

Mr. de Mello has three patents and authored more than 100 technical papers in IEEE, ISA, American Power Conference, World Power Conference, and other utility industry publications, and also lectured to professional society groups. He has served on the IEEE Systems Controls Subcommittee and the joint IEEE Working Group on Plant Response and also served as US representative on CIGRE Study Committees 38 and 39. He has taught dynamic and operational subjects in PTI’s Power Technology Course and conducted one week courses given to over 2000 engineers worldwide on Power System Dynamics.

Mr. de Mello is a Life Fellow of IEEE, a Life Fellow of ISA, a member of the National Academy of Engineering, a registered Professional Engineer in New York State. He was awarded the IEEE Charles Concordia Award in 2003.

ACKNOWLEDGEMENT

This contribution to the technology of modeling and control of large steam generators, both drum boilers, and once through, critical and subcritical, was largely based on my experience gained while working at the Analytical Engineering Department of the Apparatus Sales Division of the General Electric Co., Schenectady, NY, in the mid-60s, before GE’s entry into the analog operational amplifier-based process control business with the GEMAC line. While most of this information has been published before in various ISA and IEEE papers, it is presented here in a cohesive manner to serve as a text to those involved in the important field of plant modeling and control, implemented these days with distributed digital systems. The principles of control are the same, whether analog or digital, although the digital approach makes it much easier to use adaptive features, nonlinear logic, multiplication, division, function generation, square roots, etc., which were expensive to implement with analog controls. The design process with modern digital simulation tools and modeling capability can be used to greatly improve the performance of controls through the phases of start-up and wide ranges of operating conditions some of which, in the past, had to be handled by manual control subject to operator error. Other very important applications of the technology are in the field of operator training simulators.

A most important phase of the modeling effort involved field testing through extensive night work at power stations. Especial recognition for encouraging the work of model validation by tests goes to Carolina Power and Light with the test on Plant Robinson in 1961, to Georgia Power which insisted on having the design of the Hagan analog controls for Plant McDonough be done by GE, based on simulation, and for the Plant McDonough tests in 1964; to South Carolina Electric & Gas which awarded GE the supply of a GEMAC boiler turbine control system for its sub-critical once-through unit at the Canadys plant; to Florida Power Corp for similar effort in the design of boiler controls for its Crystal River Plant; to Baltimore Gas & Electric for the tests at the Crane plant; to Consolidated Edison with their tests at Astoria related to furnace draft and implosion studies; to NY Power Authority at the Charles Poletti Plant for similar studies.

In all this work I was blessed by the collaboration of colleagues who were of immense help in the conduct of the tests (John C. Westcott, recently deceased) and in carrying out the simulation studies, and testing ( D. N. Ewart, D. J. Ahner, and R. J. Mills). I am also grateful for the encouragement and support of my manager at GE, Dr. L. Kirchmayer (deceased), W. M. Stephens (deceased) of Georgia Power, and V. C. Summer (deceased) of South Carolina Electric & Gas. In the computer simulation effort I had a team of expert Green Berets, John Undrill whose expertise in modeling dynamic systems excelled not only in electrical large scale network dynamic and steady state performance modeling, but also in the plant area, both hydro and fossil-fired. Witness his work in furnace draft problems in Chapter VI.

Lou Hannett and Jim Feltes were always there to tackle computer simulation work for any type of electrical, electro-mechanical and thermo-mechanical system. Dick Mills made enormous contributions in the large scale plant operator training simulator area, starting at GE, developing the software for the first nuclear plant simulator at Dresden II in the late 60s, and then at PTI in similar applications in Scandinavia, and extending this to the fossil plant simulator business at PTI headed by John Westcott. I mention this because a key element in this business is the efficient and accurate modeling of components in the path of water and steam flow through heat exchangers, steam generators with their economizers, waterwalls, superheaters, reheaters, as well as the dynamics of the combustion process and heat transfer through the gas path, contained in these notes.

I would also remember my deceased wife, Barbara, who bore with my frequent absence for business including prolonged trips with plant testing, while she had the responsibility of handling four children who turned out to be responsible and successful citizens, and to my wife Margaret who puts up with me in my senior years.

Finally, this publication would not have been possible without the help from my good friend Cyrus Taft who first suggested that the work merited publication, and who worked with ISA to make this possible.

CHAPTER I

BOILER PROCESS DYNAMICS

AND CONTROL-OVERVIEW

INTRODUCTION

Control of the steam generation process in power plants is a vital function to assure reliability, economy, and safety of power production. It is also one of the most complex and difficult tasks representing a challenge to control system designers and plant operators. These comments are in the context of conditions in the mid 1960’s when the transition from pneumatic to operational amplifier technology was under way.

The capabilities of modern control technology implemented through present day analog and computer control hardware have opened vast areas of opportunity in the automation and control of power plants. These capabilities fill a vital need as the trend to larger sizes and new designs of boiler-turbine systems, at higher pressures and temperatures, increases the importance of achieving closer, safer and more sophisticated control of plant variables.

Utility engineering know-how and dedication to the technology is no doubt the most important ingredient in capitalizing on the state of the art and applying it, in logical steps, to yield badly needed improvements in the control of power plants.

Before proceeding with the development of details of the steam generation process and its controls, it is appropriate to place in proper perspective the role of energy supply dynamics in the overall power system structure.

The schematic in Fig. 1.1 shows how various systems and subsystems of automatic controls are involved in the overall control of the power system. At a more detailed level the block labeled Prime Mover System is described schematically, by its main components, typically as in Fig. 1.2. The prime mover system must maintain process conditions, such as, pressures, temperatures, and flow rates at optimum and safe values while responding to the demands of automatic or manual dispatch control. This control, in turn, acts to coordinate, instant by instant, the production from each source to meet system-load requirements at minimum cost. An understanding of the reactions of the various processes and interactions among systems and subsystems is necessary to evaluate response requirements and control system design criteria.

The area addressed in these notes concerns the power plant control system. This system, often referred to as boiler control, must act to maintain process conditions at desired levels and to protect the plant by automatic override or runback action against any upset which may endanger the safety of the equipment. The well-engineered design of a dynamically integrated control system for such a complex process as that of steam power generation cannot be based entirely on intuition and costly trial and error, especially where extrapolation of past experience is not directly applicable. Fortunately, today’s control technology and computation aids permit systhesis and evaluation of the overall boiler-turbine control prior to its installation. The indispensable requirement for engineering of the overall control system is the ability to simulate the dynamic characteristics of the boiler-turbine and auxiliaries system.

In addition to the fundamental requirement of understanding the process, that is, having the ability to quantitatively predict boiler dynamic performance, it is essential that one have a practical knowledge of control concepts and control analysis techniques.

The material in these notes addresses both the understanding of the process physics and control requirements.

FIGURE 1.1 Power System and Power Plant Control Loops

MULTIVARIABLE PROCESS

The term multivariable process characterizes those processes in which two or more mutually coupled variables are to be controlled. A boiler-turbine system is an excellent example of a multivariable process.

Fig. 1.3 describes symbolically the boiler process identifying dependent and independent variables. Independent variables are those whose values may be arbitrarily set, such as fuel, air, tilts, spray, etc. They may be considered as inputs to the process. The dependent variables describe the reaction of the process to the inputs. They are throttle pressure, main steam temperature, reheat steam temperature, drum level, etc. The important point is that these variables are mutually coupled. A change of each input variable affects some or all of the output variables. The output-input relationships are time dependent and can be described by differential equations. Although these relations are, in general, nonlinear in nature, they may often be approximated by linear differential equations where one is concerned with small excursions about a given operating point. Operationally then, a multivariable process can be described by a matrix of equations, as indicated in Fig. 1.4 which, for the sake of illustration, deals with two variables only.

The matrix notation of Fig. 1.4 is shorthand for the following equations:

x1(s) = m1(s)G11(s) + m2(s)G12(s)

x2(s) = m1(s)G21(s) + m2(s)G22(s)

The transfer functions G11(s) and G22(s) represent the self terms:

25787.png with m2(s) = 0

25794.png with m1(s) = 0

G12(s) and G21(s) are the mutual coupling terms:

25803.png with m1(s) = 0

25811.png with m2(s) = 0

The strength of the terms G12(s) and G21(s) relates to the degree of coupling in the process. In the case of the boiler process, the order of the matrix is much larger, although some coupling terms are not very significant. When coupling terms are negligible, the control of each variable can be done on an independent loop basis. In this case, the process matrix reduces to a diagonal matrix, and the problem of control becomes one of several independent single-variable control loops.

Fig. 1.5 shows the concept of multivariable control where, in principle, the controller matrix should be multivariate in nature designed to compensate for the process interactions so as to effect overall responsive and stable performance. Prediction is the fundamental concept of dynamic control. The design of a control system with the correct predictive characteristics can be attempted provided that the process response characteristics are known.

FIGURE 1.5 Multivariable Controls

PROCESS MODELING

The development of a mathematical model describing the dynamic-response characteristics of the power plant process is a most demanding and time-consuming task. This development must proceed hand in hand with the analytical being confirmed by the experimental.

There is always a strong temptation for the control engineer to try to derive a model of the process from analysis of tests only, using the input-output black-box approach, whereby hopefully he need not get involved in the analysis of the basic-process physics. However, with this approach, an empirical fit of second or third order expressions to match the tested time responses would serve little more than to model the tested unit, and would not give us the ability to predict the dynamics of other units of different designs. Further, such an approach can easily lead to error since the observed changes might have occurred as a result of some inadvertent effect which might not have been suspected without analysis. For instance, in some typical tests not involving a change in spray valve, it was discovered that a part of the temperature changes occurred because of variations of spray flow due to variations in the pressure head across the valve. This effect might not have been detected