Power Plant Control and Instrumentation The Control of Boilers and HRSG Systems

Power-plant control
and instrumentation
The control of boilers
and HRSG systems
David Lindsley
The Institution of Electrical Engineers
To my wife, Jo. Thanks for everything, especially your
patience and eagle-eyed spotting of errors during the
checking of this book.
Published by: The Institution of Electrical Engineers, London,
United Kingdom
2000: The Institution of Electrical Engineers
Reprinted 2005
This publication is copyright under the Berne Convention and the
Universal Copyright Convention. All rights reserved. Apart from any fair
dealing for the purposes of research or private study, or criticism or
review, as permitted under the Copyright, Designs and Patents Act, 1988,
this publication may be reproduced, stored or transmitted, in any forms or
by any means, only with the prior permission in writing of the publishers,
or in the case of reprographic reproduction in accordance with the terms
of licences issued by the Copyright Licensing Agency. Inquiries
concerning reproduction outside those terms should be sent to the
publishers at the undermentioned address:
The Institution of Electrical Engineers,
Michael Faraday House,
Six Hills Way, Stevenage,
Herts. SG1 2AY, United Kingdom
While the author and the publishers believe that the information and
guidance given in this work are correct, all parties must rely upon their
own skill and judgment when making use of them. Neither the author nor
the publishers assume any liability to anyone for any loss or damage
caused by any error or omission in the work, whether such error or
omission is the result of negligence or any other cause. Any and all such
liability is disclaimed.
The moral right of the author to be identified as author of this work has
been asserted by him/her in accordance with the Copyright, Designs and
Patents Act 1988.
British Library Cataloguing in Publication Data
A CIP catalogue record for this book
is available from the British Library
ISBN 0 85296 765 9
Printed in Great Britain by
Cambridge University Press, England
Preface
The aim of this book is to examine the control and instrumentation
systems of the drum-type boilers and heat-recovery steam generators
(HRSGs) that are used for the production of steam for turbines and indus-
trial processes. My intention is to provide information to assist the
designers, users and maintenance staff of such plants in understanding how
these systems function.
The end product of the steam plant may be electricity that is exported
to the grid, or it may be steam or hot water that is sent to a nearby process
plant, factory or housing complex, but in each case the general principles
of the control systems will be very similar. Nevertheless, the design of these
systems is a specialised task, an art as much as a science, and in this intro-
duction I aim to draw attention to the width and depth of knowledge that
it demands.
The knowledge base one needs to design a control system for a boiler
or HRSG is unusually wide. A power station is a complex entity,
embracing a wide range of what I refer to as primary disciplines-physics,
chemical engineering, thermodynamics, mechanical engineering and elec-
trical engineering. It also involves control technology and computing--the
secondary disciplines that combine two or more of the primary subjects.
The many machines in a power station operate together as an inte-
grated, highly interdependent system. In practice, engineers involved in
any area of power-station design, operation or maintenance must neces-
sarily have their skills focused on just one of the primary disciplines, neces-
sitating a high degree of training and experience in the relevant field: but
to work effectively they will also need at least a basic understanding of all
the others.
That may seem wide enough for anyone, but as soon as the focus
narrows onto one area, the control and instrumentation systems (C&I) of
x Priface
power stations, it becomes apparent that the subject is even more
demanding. Engineers working in this particular field must be proficient
in the highly complex areas of control theory and computers (hardware,
software or both-fast moving, ever changing subjects in their own rights),
but in addition they should have at least a rudimentary understanding of
the thermodynamics of steam generation and use, and of metallurgy,
chemistry and mechanical design. In addition it may be necessary to
understand how high-voltage heavy-current electrical systems work.
One of the problems faced by the industry is the need for control
engineers who understand, and are competent in, the very demanding
field of computer systems, as well as in the more traditional areas of engi-
neering. But, whereas the quantity and variety of information required by
the engineer has grown enormously over the past half-century, the period
allocated to graduate training has not expanded beyond the same four or
five years that I spent while I was being trained. And in my day computers
were specialised things that one might, perhaps, study after graduating.
Beside being complicated, computer technology is beguiling. It is
tempting, and intellectually satisfying, to sit at a keyboard tapping away
and generating words, formulae or pictures on the screen. If a mistake is
made, the thing simply doesn't work. At worst the system may 'crash',
necessitating a reboot-a process that may, at worst, result in the loss of
much carefully-constructed data. But that is all.
On the other hand, a computer controlling any power-station plant is
in command of a huge process involving explosive mixtures of gases, steam
at pressures and temperatures that become instantly lethal if anything goes
wrong, and massive roaring turbines driving generators that produce
megawatts of power at many tens of thousands of volts. A small mistake or
lack of attention to detail in such a case can have consequences that will
certainly be severe, probably very expensive and possibly tragic.
A power station is a complex thing, and its construction is a frantic,
long drawn-out process involving many people, sometimes hundreds of
them, working amid the difficulties of noise, dust and dirt, and extremes of
temperature. Heavy items are craned or manhandled into position under
a mess of cables and pipes, often with showers of sparks raining down from
welding and cutting operations high above. An instrument lovingly
installed on a pipe is all too often used as a foothold for a heavy-booted
rigger reaching up to install an item on another pipe. Instrument cubicles
are on occasion used as latrines by labourers who are caught short in the
middle of a task. Many a control desk designed with an eye for artistic
merit has come into violent contact with a massive steel girder being
moved into position-and emerged the worse off!
Preface x1
When one moves away from Europe or North America things become
even worse. (The expression 'debugging' takes on a new significance when
one has to extract a large and aggressive cockroach from an 1/0 card
rack.)
Across the world, cable trenches are dug and cables laid in them by elec-
tricians and labourers who have little or no understanding of electronics.
Expert supervision has its limits. Even if much careful attention has been
paid to defining earthing and screening requirements, all may be lost if the
wrong type of cable gland is used at a single point, or if armoured cable is
wrongly glanded. The fact that a malfunction has been caused by interfer-
ence is difficult enough to determine. Trying to discover why and where
the interference occurred in kilometres of cable trays and ducts snaking
their way through a vast site is often an impossible task.
So why should anybody in their right minds want to work in a field
that is often difficult, sometimes dangerous and always stressful?
Everybody will have their own answer but, for me, the magic of this field is
its huge scope and enormous challenges. In few other industries will one
have to apply one's mind to technologies that are so wide-ranging and
disparate as the thermodynamic processes of steam at 500C and the
operation of a high-speed data highway.
It is a varied, demanding and exciting field, and if in the course of
explaining its complexities I can lure into the power-station C&I field a
few people who might otherwise not have considered an engineering
career, then I shall be pleased.
So what I have tried to do in this book is to provide an outline of the
subject in a readable format. In doing this I have had to limit the depth of
the coverage. I make no apologies for glossing over some topics and for
simplifying some concepts. The experts in a particular field may well
quibble with my explanations, but I would maintain that if the ideas work
in practice, then that is an adequate starting point. It will always be
possible to refine the detail later on.
I must try to explain how I have approached the practical aspects of
boiler control and instrumentation. The rapid evolution of technology
makes it dangerous to define any details of implementation (a photograph
of today's state-of-the-art control room becomes very dated within only a
few years!). For this reason I have tried to concentrate on the overall prin-
ciples of each system, as I did with my earlier book on this subject, since
the principles of three-element feed-water control, as implemented in a
modern distributed control system, are virtually the same as those imple-
mented in a 40-year-old pneumatic system fulfilling the same function.
This time, however, in addition to information on system prin-ciples I
have tried to provide practical information on transmitters, analysers,
xii Priface
flame monitors, actuators and cabling. At the time of writing, develop-
ments in these areas seem to have reached something of a plateau, and I
can only hope that the information I have provided will not become out-
dated too soon. In any event, I believe that the matters relating to trans-
mitter and actuator installation and use will remain relevant well into the
conceivable future.
Finally I would like to thank the many individuals and organisations
who have made contributions to this book, either with direct contributions
of diagrams and technical articles, or by the provision of information. In
thanking the following, I do not wish to ignore all the others who have
helped me with this work: Balfour Beatty Ltd., Croydon, Surrey; B.I.C.C.
Components Ltd., Bristol Babcock Ltd., Kidderminster, Worcestershire;
British Standards Institution, London; Copes-Vulcan Ltd., Winsford,
Cheshire; Fireye Ltd., Slough, Berkshire; Howden Sirocco Ltd., Glasgow;
Kvaerner Pulping Ltd., Gothenburg, Sweden; Measurement Technology
Ltd., Luton, Bedfordshire; Mitsui Babcock Ltd., Crawley, West Sussex;
National Power plc, Swindon, Wiltshire; Rosemount Engineering
Company, Bognor Regis, West Sussex; Scottish Power plc, Glasgow;
Solartron Ltd., Farnborough, Hampshire; and Watson Smith Ltd.,
Leeds.
Diagra.Dltnatic sytnbols
In spite of the existence of many recognised standards for instrumentation
symbols [1], I have chosen to adopt a simple format which should be suffi-
cient to explain the concepts that I want to communicate to the reader.
These symbols would not be comprehensive enough to fully define the
requirements within a full-scale control-system design task (for example,
the controller symbol does not indicate whether or not auto/
manual facilities are required, or the form that these should take).
Nevertheless, I believe the diagrams will be easily understood by
engineers.
In the context of the controllers themselves, it is worth mentioning that
different terms are used in the USA and elsewhere to identify the same
function. In particular, the plant parameter that is measured and fed to a
controller is, in Europe, called the 'measured value,' while in the USA it is
referred to as the 'process variable'. Also, when referring to controllers, the
term 'reset' is often used in the USA instead of 'integral action'.
Abbreviations and term.s
used in this book
This book is addressed to people working across two very different
disciplines: power-plant and control systems. Technical terms and abbre-
viations that are easily understood by professionals in one field can be
bewildering to those who understand the other side, and so the following
list is provided in an attempt to help readers understand the abbreviations
and some of the terms that are used in the text and elsewhere in the
industry.
loo2
2oo2
2oo3
AC
ASCII
ADC
A/M
BSI
C&I
CCR
CCGT
CE
CHP
CMR
CPU
DAC
DC
DCS
voting
two-out-of-two voting
two-out-of-three voting
alternating current
American Standard Code for Information Interchange (a
standard defining the codes used for communication between
computers and between computers and their peripherals)
analogue-to-digital converter
auto/manual control facility
British Standards Institution
control and instrumentation
central control room
combined-cycle gas-turbine plant
European Community
combined heat and power (a type of plant that burns a fuel to
produce electricity and steam that is used either to heat a nearby
complex or by an industrial process)
continuous maximum rating (also MCR)
central processing unit
digital-to-analogue converter
direct current
distributed control system
xv1 Abbreviations and terms used in this book
deterministic
DV
EEPROM
EMC
EMI
FAT
FD
FDS
FWR
H/A
HP
HRSG
IC
ID
IEC
lEE
IEEE
IP
ISO
I/0
KKS
LAN
LED
load
LP
machine
MCB
MCR
mill
MTBF
MTTR
MV
P&ID
PCB
PF
A deterministic system is one in which events are dealt with in the
exact order in which they occur. With some systems, events are
dealt with by means which causes action to be taken in a sequence
that is dictated by external constraints (such as polling). Such a
system is not deterministic
desired value
electrically erasable programmable read-only memory
electromagnetic compatibility
electromagnetic interference
factory acceptance test
forced draught
functional design specification
feed-water regulator (control valve)
hand/automatic control facility
high pressure (the definition is relative: on major central-station
plant it is usually above 100 bar g)
heat-recovery steam generator
integrated circuit
induced draught
International Electro-technical Commission
Institution of Electrical Engineers
Institute of Electrical and Electronics Engineers
intermediate pressure (a relative definition, see HP above)
International Standards Organisation
input and output
Kraftwerk Kennzeichensystem (power station designation
system)
local-area network
light emitting diode
the flow of steam, in kg/s, that is produced at any given time by
the boiler or HRSG (sometime also the electrical load on the
generator, in MW)
low pressure (a relative definition, see HP above)
turbo-generator or alternator
miniature circuit breaker
maximum continuous rating (also CMR), typically, the highest
rate of steam flow that a boiler can produce for extended periods.
a device (also known as a pulveriser) that is used to crush coal into
fine powder before it is fed to the burners
mean time between failure
mean time to repair
measured value (also known as 'process variable')
piping and instrumentation diagram
printed circuit board
pulverised fuel (coal)
PLC
PSU
pulveriser
PV
RAM
RDF
RFI
ROM
RTC
SAT
SCAD A
TUV
UART
UL
UPS
VDU
WTE
Reference
Abbreviations and terms used in this book xvn
programmable-logic controller
power supply unit
a device (also known as a mill) that is used to crush coal into fine
powder before it is fed to the burners
process variable (also known as 'measured value')
random-access memory
refuse-derived fuel
radio-frequency interference
read-only memory
real-time clock
site acceptance test
supervisory, control and data-aquisition system
Technischer Uberwachungs Verein (German Technical Super-
visory Association)
universal asynchronous receiver /transmitter (an electronic de-
vice that controls communication with a peripheral)
underwriters' laboratories
uninterruptible power supply
visual display unit (also termed a 'monitor' or 'screen')
waste-to-energy (a type of plant where waste is burned to produce
electricity or heat for a district or industrial process)
ANSI/ISA-S5.1: Instrumentation symbols and identification. Instrument
Society of America, Research Triangle Park, North Carolina, USA, 1992
Cont e nt s
Pr e f ac e
Di a g r a mma t i c s y mbo l s
Abbrevi at i ons and t e r ms us e d i n t hi s book
1 The bas i c s o f s t e a m generat i on and us e
1.l Why an understanding of steam is needed
1.2 Boiling: the change of state from water to steam
1.3 The nature ofsteam
1.4 Thermal efficiency
1.5 The gas turbine and co mbined-cycle plants
1.6 Summary
2 The s t e a m and wa t e r ci rcui t s
2.1 Steam generation and use
2.2 The steam turbine
2.3 The condensate and feed-water system
2.4 The feed pumps and valves
2.5 The water and steam circuits of HRSG plant
2.6 Summary
3 The f uel , ai r and f l ue- gas ci rcui t s
3.1 The furnace
3.2 The air and gas circuits
3.3 Fuel systems
3.4 Igniter systems
3.5 Burner-management systems
3.6 Gas turbines in combined-cycle applications
3.7 Summary
4 Set t i ng t he de ma nd f or t he s t e a m generat or
4.1 Nature ofthe demand
i x
o
x n l
x v
1
1
2
4
9
lO
12
13
13
15
17
21
21
23
25
25
27
35
44
45
46
46
49
49
vi Contents
4.2 Setting the demand in power-station applications
4.3 The mast er demand in a power-station appl i cat i on
4.4 Load demand in combi ned heat and power plants
4.5 Waste-to-energy plants
4.6 Summar y
5 Combust i on and draught control
5.1 The principles of combust i on control
5.2 Worki ng with mul t i pl e fuels
5.3 The cont r ol of coal mi l l s
5.4 Dr aught control
5.5 Binary control of the combust i on system
5.6 Summar y
6 Feed-water control and i nst rument at i on
6.1 The principles of feed-water control
6.2 One, two and t hree-el ement control
6.3 Measuri ng and displaying the dr um level
6.4 The mechani sms used for feed-water control
6.5 Pumps
6.6 De-aerat or control
6.7 Summar y
7 St eam- t emperat ure control
7.1 Why st eam-t emperat ure control is needed
7.2 The spray-water at t emper at or
7.3 Temper at ur e control with tilting burners
7.4 Controlling the t emper at ur e of reheat ed steam
7.5 Gas recycling
7.6 Summar y
8 Control equi pment practi ce
8.1 A typical DCS configuration
8.2 Int erconnect i ons between t hesyst ems
8.3 Equi pment selection and envi ronment
8.4 Mechani cal factors and ergonomics
8.5 Electrical act uat ors
8.6 Hydraul i c act uat ors
8.7 Cabl i ng
8.8 El ect romagnet i c compat i bi l i t y
8.9 Reliability of Systems
8.10 Summar y
50
53
59
60
61
63
64
79
81
91
95
100
103
103
103
112
121
127
131
134
135
135
136
150
151
152
152
153
153
157
158
159
168
168
169
169
172
180
Contents vii
9 Re qui r e me nt s def'mition and e qui pme nt nome nc l at ur e
9.1 Overview
9.2 Defining the requirements
9.3 The KKS equipment identification system
9.4 Summary
10 Upgradi ng and r e f ur bi s hi ng s y s t e ms
10.1 The reasons behind the changes
10.2 Living with change
10.3 Making the decision to change
10.4 Arefurbi shment casest udy
10.5 Why refurbish?
10.6 Document i ng the present system configuration
10.7 Summary
Furt her readi ng
Index
183
183
183
186
195
197
198
199
201
202
204
205
205
209
213
Chapter 1
Th e b a s i c s o f s t e a m g e n e r a t i o n
a n d u s e
1.1 Wh y a n u n d e r s t a n d i n g o f s t e a m i s n e e d e d
St eam power is f undament al to what is by far the largest sector of t he elec-
t ri ci t y-generat i ng i ndust ry and wi t hout it t he face of cont empor ar y society
woul d be dr amat i cal l y di fferent from its present one. We woul d be forced
to rely on hydro-el ect ri c power plant, windmills, batteries, solar cells and
fuel cells, all of whi ch ar e capabl e of pr oduci ng onl y a fraction of t he elec-
tricity we use.
St eam is i mpor t ant , and t he safety and efficiency of its generat i on and
use depend on the appl i cat i on of cont rol and i nst rument at i on, often simply
referred to as C&I . The objective of this book is to provi de a bridge
bet ween t he discipline of power-pl ant process engi neeri ng and those of
electronics, i nst r ument at i on and cont rol engineering.
I shall st art by out l i ni ng in this chapt er t he change of state of wat er to
steam, followed by an overvi ew of t he basic principles of st eam generat i on
and use. Thi s seemi ngl y simple subject is ext remel y complex. Thi s will
necessarily be an overview: it does not pr et end to be a det ai l ed treatise and
at times it will simplify mat t ers and gloss over some details whi ch may
even cause t he t her modynami ci st or combust i on physicist to shudder, but
it should be underst ood t hat t he ai m is to provi de t he C&I engi neer wi t h
enough under st andi ng of t he subject to deal safely with pract i cal control-
system design, operat i onal and mai nt enance problems.
2 Power.plant control and instrumentation
1. 2 Bo i l i ng : t h e c h a n g e o f s t a t e f r o m w a t e r t o s t e a m
Whe n wat er is heat ed its t emper at ur e rises in a way t hat can be det ect ed
(for exampl e by a t her momet er ) . The heat gai ned in this way is called
sensible because its effects can be sensed, but at some poi nt t he wat er starts
to boil.
But here we need to look even deeper into the subject. Exact l y what is
meant by t he expression ' boiling' ? To st udy this we must consider the t hree
basic states of mat t er: solids, liquids and gases. (A pl asma, pr oduced when
t he at oms in a gas become ionised, is oft en referred to as the f our t h state of
mat t er, but for most pract i cal purposes it is sufficient to consi der onl y t he
t hree basic states.) I n its solid state, mat t er consists of many mol ecul es
t i ght l y bound t oget her by at t ract i ve forces bet ween them. Whe n the
mat t er absorbs heat t he ener gy levels of its mol ecul es increase and t he
mean di st ance bet ween t he mol ecul es increases. As mor e and mor e heat is
appl i ed these effects increase until t he at t ract i ve force bet ween the
mol ecul es is event ual l y over come and the particles become capabl e of
movi ng about i ndependent l y of each other. Thi s change of state from solid
to liquid is commonl y recogni sed as ' melting' .
As mor e heat is appl i ed to t he liquid, some of the mol ecul es gai n
enough ener gy to escape from t he surface, a process called evapor at i on
( wher eby a pool of liquid spilled on a surface will gr adual l y di sappear).
Wha t is happeni ng dur i ng the process of evapor at i on is t hat some of the
mol ecul es are escapi ng at fairly low t emperat ures, but as t he t emper at ur e
rises these escapes occur mor e rapi dl y and at a cert ai n poi nt t he liquid
becomes ver y agi t at ed, wi t h l arge quant i t i es of bubbl es rising to the
surface. It is at this t i me t hat the liquid is said to st art ' boiling' . It is in t he
process of changi ng state to a vapour, whi ch is a fluid in a gaseous state.
Let us consi der a quant i t y of wat er t hat is cont ai ned in an open vessel.
Her e, the air t hat bl anket s t he surface exerts a pressure on t he surface of
t he fluid and, as t he t emper at ur e of t he wat er is raised, enough energy is
event ual l y gai ned to over come t he bl anket i ng effect of t hat pressure and
t he wat er starts t o change its state i nt o t hat of a vapour (steam). Fur t her
heat added at this st age will not cause any furt her det ect abl e change in
t emper at ur e: t he energy added is used to change t he st at e of t he fluid. Its
effect can no l onger be sensed by a t her momet er , but it is still there. For
this reason it is cal l ed latent, r at her t hen sensible, heat. The t emper at ur e at
whi ch this happens is called t he ' boi l i ng point' . At nor mal at mospheri c
pressure t he boi l i ng poi nt of wat er is 100 C.
I f the pressure of t he air bl anket on t op of t he wat er were to be
increased, mor e ener gy woul d have to be i nt roduced to the wat er to enabl e
The basics o f steam generation and use 3
it to break free. I n ot her words, t he t emper at ur e must be raised furt her to
make it boil. To illustrate this point, i f t he pressure is increased by 10%
above its nor mal at mospheri c value, t he t emper at ur e of t he wat er must be
raised to j ust above 102 C before boiling occurs.
The st eam emergi ng from t he boiling liquid is said to be sat urat ed and,
for any given pressure, t he t emper at ur e at whi ch boiling occurs is called
t he saturation temperature.
The i nformat i on rel at i ng to st eam at any combi nat i on of t emperat ure,
pressure and ot her factors may be found in st eam tables, whi ch are
nowadays available in software as well as in t he mor e t radi t i onal paper
form. These tables were originally published in 1915 by Hugh Longbour ne
Cal l endar (1863-1930), a British physicist. Because of advances in
knowl edge and measur ement technology, and as a result of changi ng units
of measur ement , ma ny di fferent vari ant s of st eam tables are t oday in
existence, but t hey all enable one to look up, for any pressure, t he sat ura-
tion t emper at ur e, t he heat per unit mass of fluid, t he specific vol ume etc.
Under st andi ng st eam and t he st eam tables is essential in many stages
of t he design of power-pl ant cont rol systems. For exampl e, i f a desi gner
needs to compensat e a steam-flow measur ement for changes in pressure, or
to correct for density errors in a water-level measur ement , reference to
these tables is essential.
Anot her t er m rel at i ng to st eam defines t he quant i t y of liquid mi xed in
with t he vapour. I n t he UK this is called t he dryness f r act i on (in the USA t he
t er m used is steam qual i t y). What this means is t hat if each ki l ogram of t he
mi xt ur e cont ai ns 0.9 kg of vapour and 0.1 kg of water, t he dryness fract i on
is 0.9.
St eam becomes superheated when its t emper at ur e is raised above t he
sat urat i on t emper at ur e correspondi ng to its pressure. Thi s is achi eved by
collecting it from t he vessel in whi ch t he boiling is occurri ng, l eadi ng it
away from t he liquid t hr ough a pipe, and t hen addi ng mor e heat to it. This
process adds furt her ener gy to t he fluid, whi ch improves t he efficiency of
t he conversion of heat to electricity.
As stated earlier, heat added once t he wat er has st art ed to boil does
not cause any furt her det ect abl e change in t emperat ure. Inst ead it changes
t he state of t he fluid. Once t he st eam has formed, heat added to it contri-
butes to t he total heat of t he vapour. This is t he sensible heat pl us t he l at ent
heat p l u s t he heat used in increasing t he t emper at ur e of each ki l ogram of
t he fluid t hr ough t he number of degrees of superheat to whi ch it has been
raised.
I n a power plant, a maj or objective is t he conversion of energy locked
up in t he i nput fuel into ei t her usable heat or electricity. I n t he interests of
economics and t he envi r onment it is i mpor t ant to obt ai n t he highest
4 Power-plant control and instrumentation
possible level of efficiency in this conversion process. As we have already
seen, the greatest efficiency is obtained by maximising the energy level of
the steam at the point of delivery to the next stage of the process. When as
much energy as possible has been abstracted from the steam, the fluid
reverts to the form of cold water, which is then warmed and treated to
remove any air which may have become entrained in it before it is finally
returned to the boiler for re-use.
1. 3 Th e n a t u r e o f s t e a m
As stated in the Preface, the boilers and steam-generators that are the
subject of this book provide steam to users such as industrial plant, or
housing and other complexes, or to drive turbines that are the prime
movers for electrical generators. For the purposes of this book, such
processes are grouped together under the generic name ' power plant' . In
all these applications the steam is produced by applying heat to water until
it boils, and before we embark on our study of power-plant C&I we must
understand the mechanisms involved in this process and the nature of
steam itself.
First, we must pause to consider some basic thermodynamic processes.
Two of these are the Carnot and Rankine cycles, and although the C&I
engineer may not make use of these directly, it is nevertheless useful to
have a basic understanding of what they are how they operate.
1.3.1 The Carnot cycle
The primary function of a power plant is to convert into electricity the
energy locked up in some form of fuel resource. In spite of many attempts,
it has not proved possible to generate electricity in large quantities from
the direct conversion of the energy contained in a fossil fuel (or even a
nuclear fuel) without the use of a medium that acts as an intermediary.
Solar cells and fuel cells may one day achieve this aim on a scale large
enough to make an impact on fossil-fuel utilisation, but at present such
plants are confined to small-scale applications. The water turbines of
hydro-electric plants are capable of generating large quantities of electri-
city, but such plants are necessarily restricted to areas where they are
plentiful supplies of water at heights sufficient for use by these machines.
Therefore, if one wishes to obtain large quantities of electricity from a
fossil fuel or from a nuclear reaction it is necessary to first release the
energy that is available within that resource and then to transfer it to a
generator, and this process necessitates the use of a medium to convey the
The basics o f steam generation and use 5
energy from source to destination. Furt hermore, it is necessary to empl oy
a medi um t hat is readi l y avai l abl e and which can be used with relative
safety and efficiency. On pl ant Eart h, wat er is, at least in general, a
plentiful and cheap medi um for effecting such transfers. Wi t h the develop-
ment of t echnol ogy duri ng the t went i et h cent ur y ot her possibilities have
been considered, such as the use of mercury, but except for applications
such as spacecraft where entirely new sets of limitations and conditions
apply, none of these has reached active use, and st eam is universally used
in power stations.
The use of wat er and steam to provi de mot i ve power has a l ong
history. In the first cent ur y AD Her o of Al exandri a showed t hat steam
leaving via nozzles at t ached to a heat ed cont ai ner filled wi t h wat er woul d
cause the vessel to rotate, but in this simple machi ne (the aeolipile) the
st eam leaving the vessel was wast ed and for cont i nuous operat i on t he
process t herefore necessitated cont i nual l y repl aci ng the water. Wi t h the
nat ur e of Hero' s design, it was not a saimple task to refill the vessel while it
was in operat i on, but even i f a met hod had been found, using wat er in a
one- way process like this necessitates the provision of endless supplies of
t hat fluid. It was not until 1824 t hat a French engineer, Sadi Carnot ,
proposed a way to resolve this probl em. He used a cycle, where the transfer
medi um is par t of a closed l oop and the medi um is r et ur ned to its starting
conditions after it has done the work requi red of it.
Car not framed one of the t wo laws of t hermodynami cs. The first,
Joul e' s law, had rel at ed mechani cal energy to work: Carnot ' s law defined
the t emper at ur e relations appl yi ng to the conversion of heat energy into
mechani cal energy. He saw t hat i f this process were to be made reversible,
heat coul d be convert ed into work and then ext ract ed and re-used to make
a closed loop. In his concept (Figure 1.1), a piston moves freely wi t hout
encount eri ng any friction inside a cylinder made of some perfectly insu-
lating mat eri al . The piston is dri ven by a ' worki ng fluid'. The cylinder has
a head at one end t hat can be switched at will from bei ng a perfect
conduct or to being a perfect insulator. Out si de the cylinder are two bodies,
one of which can deliver heat wi t hout its own t emper at ur e ( T~ ) falling, the
ot her bei ng a bot t oml ess cold sink at a t emper at ur e (7-2) which is also
constant.
The operat i on of the system is shown graphi cal l y in figure 1.2, which
shows the pr essur e/ vol ume relationship of the fluid in the cyl i nder over t he
whol e cycle. As the process is a repeat i ng cycle its operat i on can be st udi ed
from any conveni ent st art i ng point, and we shall begin at the poi nt A,
where the cyl i nder head (at this time assumed to be a perfect conduct or of
heat), allows heat from the hot source to ent er the cylinder. The result is
t hat the medi um begins to expand, and i f it is allowed to expand freely,
Cylinder head
, l
. I . . . . . . . . . . . . . / I
W o r k i n g , , u i ]
Figure 1.1 Carnot's heat engine
Boyle's law (which states t hat at any t emper at ur e t he relationship
bet ween pressure and vol ume is constant) dictates t hat the t emperat ure
will not rise, but will stay at its initial t emper at ur e (Tl). This is called iso-
t hermal expansion.
When the pressure and vol ume of t he medi um have reached t he values
at poi nt B, the cylinder head is switched from being a perfect conduct or to
being a perfect insulator and t he medi um allowed to cont i nue its
expansi on wi t h no heat being gai ned or lost. Thi s is known as adiabatic
expansion. When t he pressure and vol ume of the medi um reach t he values
at poi nt C, t he cylinder head is switched back to being a perfect conductor,
but the external heat source is removed and replaced by the heat sink. The
piston is dri ven towards the head, compressing t he medi um. Heat flows
t hr ough t he head to the heat sink and when the t emper at ur e of the
medi um reaches t hat of the heat sink (at poi nt D), the cylinder head is
once again switched to become a perfect insulator and t he medi um is com-
pressed until it reaches its starting conditions of pressure and t emperat ure.
The cycle is t hen complete, havi ng t aken in and rejected heat while doing
external work.
1. 3. 2 T h e Ra n k i n e cycle
The Car not cycle postulates a cylinder with perfectly insulating walls
and a head whi ch can be switched at will from Being a conduct or to being
The basics of steam generation and use 7
A
I s o t h e r m a l
A d i a b a t i c n
I
o D o
I I
; ; ,
i !
' ~ I s o t h e r m a l ;
; i o
I !
|
Figure 1.2 The Carnot cycle
B
A d i a b a t i c
b C
I
I
I
V o l u m e
an insulator. Even with modifications to enabl e it to operat e in a worl d
where such things are not obt ai nabl e, it woul d have pr obabl y r emai ned a
scientific concept wi t h no pract i cal application, had not a Scottish
professor of engineering, William Ranki ne, proposed a modification to it
at t he begi nni ng of t he t went i et h cent ur y [I]. The concepts t hat Ranki ne
devel oped form t he basis of all t her mal power plants in use today. Even
todays combi ned-cycl e power plants use his cycle for one of t he two phases
of t hei r operation.
Fi gure 1.3 illustrates t he principle of t he Ranki ne cycle. St art i ng at
poi nt A again, t he source of heat is applied to expand t he medi um, this
t i me at a const ant pressure, to poi nt B, aft er whi ch adi abat i c expansi on
is agai n made to occur until t he medi um reaches t he conditions at poi nt
C. Fr om here, t he vol ume of t he medi um is reduced, at a const ant
pressure, until it reaches poi nt D, when it is compressed back to its initial
conditions.
o
o
T
A
B
B
I
J
I
' i
!
! ,
C
Votumo
Figure 1.3 The Rankine cycle
All of this may seem of onl y t heoret i cal interest, but it takes on a
pract i cal form in a t hermal power pl ant , where wat er is compressed by
pumps, t hen heat ed until it boils to pr oduce st eam which t hen expands
( t hr ough a t urbi ne or in some process) until it reverts to water. Thi s
oper at i on is shown in Fi gure 1.4 whi ch this t i me shows t emper at ur e
pl ot t ed against a quant i t y called ent r opy for the processes wi t hi n the boiler
and t ur bi ne of a power pl ant . ( Chapt er 2 describes in detail the functions
of t he vari ous i t ems of plant.) Ent r opy is a measure of t he port i on of the
ener gy in a syst em t hat is not avai l abl e for doi ng wor k and it can be used
to cal cul at e heat transfer for a reversible process.
In t he system shown in Fi gure 1.4, wat er is heat ed in feed heat ers (A to
B) using st eam ext r act ed from the turbine Wi t hi n t he boiler itself, heat is
used to furt her pr ewar m the wat er (in the economi ser) before it enters the
evaporat i ve stages (C) where it boils. At D superheat is added until the
condi t i ons at E are r eached at t he t ur bi ne inlet. The st eam expands in the
t ur bi ne to the condi t i ons at poi nt F, aft er which it is condensed and
r et ur ned to the feed heater. The energy in the st eam leaving the boi l er is
conver t ed to mechani cal ener gy in the turbine, whi ch t hen spins the
gener at or to pr oduce electricity.
E
Superheater
C ~ Evapo~tor
~ Feed heaters I <
F
A Condenser
Entropy
Figure 1.4 The Rankine cycle in a steam-turbine power plant
The di agr am shows t hat t he energy delivered to t he t urbi ne is
maxi mi sed i f poi nt E is at t he highest possible val ue and F is at t he lowest
possible value, and now we begin to see t he i mpor t ance of under st andi ng
these cycles if pl ant operat i on is to be underst ood and optimised. It
explains why t he t emper at ur e of t he st eam leaving t he boiler is super-
heat ed and why t he t urbi ne condenser operat es at very low pressures,
whi ch correspond with low t emperat ures.
1.4 T h e r m a l e f f i c i e n c y
The efficiency of a power pl ant is t he measur e of its effectiveness in con-
vert i ng fuel into electrical energy or process heat. This fact or sets t he cost
per uni t of electricity or heat generat ed, and in a net work of i nt ercon-
nect ed power stations it is this cost t hat det ermi nes t he revenue t hat will be
ear ned by t he plant. Al t hough several steps ma y be t aken to reduce losses,
some heat is inevitably lost in t he flue gases and in t he cooling wat er t hat
leaves t he condenser, and a realistic limit for t he efficiency of such a pl ant
is j ust over 40%. Al t hough it has long been underst ood that, for every unit
of money put into t he operat i on of t he plant, over hal f was bei ng lost, very
little could be done about this situation until devel opment s in materials
t echnol ogy br ought forward new opportunities.
One of the most dr amat i c power-pl ant devel opment s of t he second hal f
of t he t went i et h cent ury is t he realisation t hat by empl oyi ng one cycle in
combi nat i on wi t h anot her one, heat wasted in one could be use by t he
ot her to at t ai n enhanced efficiency, this is t he combi ned cycle.
1.5 The gas t ur bi ne and c o mbi ne d- c y c l e pl a nt s
The combined-cycle power station uses gas turbines to increase the effi-
ciency of t he power-generat i on process. Like many ot her machines that we
assume to be product s of the t went i et h century, t he gas t urbi ne isn' t t hat
new. I n fact, Leonar do da Vinci (1452-1519) sketched a machi ne for
ext ract i ng mechani cal energy from a gas stream. However, no practical
i mpl ement at i on of such a machi ne was considered until the ni net eent h
century, when George Brayt on proposed a cycle t hat used a combust i on
chamber exhaust i ng to t he atmosphere. I n 1872 Germany' s F. Stolze
pat ent ed a machi ne t hat ant i ci pat ed many features of a moder n gas-
t urbi ne engine, al t hough its performance was limited by t he constraints of
t he materials available at t he time.
Many ot her devel opment s across Europe cul mi nat ed in t he develop-
ment of an efficient gas t urbi ne by Frank Whi t t l e at the British Royal
Aircraft Establishment (RAE) in the early 1930s. Subsequent develop-
ment s at RAE led to viable axial-flow compressors, which could at t ai n
hi gher efficiencies t han t he centrifugal count erpart devel oped by Whittle.
All these gas turbines empl oyed t he Brayton cycle, whose pressure/
vol ume characteristic is shown in Figure 1.5. St art i ng at poi nt A in this
cycle air is compressed isentropically (A-B) before being fed into a com-
bustion chamber, where fuel is added and bur ned (B-C). The energy of the
expandi ng air is t hen convert ed to mechani cal work in a t urbi ne (C-D).
From C to D heat is rejected, and in a simple gas-turbine cycle this heat is
lost to t he atmosphere.
The rot at i on of the gas t urbi ne can be used to drive a generat or (via
suitable reduct i on gearing) but, when used in a simple cycle with no heat
recovery, t he t hermal efficiency of the gas t urbi ne is poor, because of the
heat lost to the at mosphere. The gases exhausted from the t urbi ne are not
only plentiful and hot (400-550C), but they also cont ai n substantial
amount s of oxygen (in combust i on terms, t he excess air level for the gas
t urbi ne is 200-300%). These factors poi nt to t he possibility of using the
hot, oxygen-rich air in a st eam-generat i ng plant, whose steam out put
drives a turbine.
Figure 1.5
B
; i
i
i
; i
;
;
; ~ . . ;
;
; I s e n t r o p i c ;
. - . ; . . . . . . . . . . . . . . . .
;
! !
I I
I I
, !
! I
The Brayton cycle
C
I s e n t r o p k :
I
!
I
I
I
D
V o l u m e
T h e use of s uch ot he r wi s e wa s t e d he a t i n a he a t - r e c ove r y s t e a m
g e n e r a t o r ( HRS G) is t he basis o f t he ' c ombi ne d- c yc l e ga s - t ur bi ne '
( CCGT) p l a n t whi c h has be e n a ma j o r d e v e l o p me n t o f t he pas t few
decades . Wi t h t he h e a t us ed t o g e n e r a t e s t e a m i n t hi s way, t he whol e p l a n t
be c ome s a b i n a r y u n i t e mp l o y i n g t he f eat ur es o f b o t h t he Ra n k i n e a n d t he
Br a y t o n cycles t o achi eve efficiencies t h a t ar e s i mpl y n o t possi bl e wi t h
ei t her cycl e on its own. I n fact , t he a d d i t i o n o f t he H R S G yi el ds a t h e r ma l
effi ci ency t h a t ma y be 5 0 % h i g h e r t h a n t h a t of t he gas t u r b i n e o p e r a t i n g
i n s i mpl e- cycl e mode .
On c e agai n, t he r e is n o t h i n g r eal l y n e w a b o u t t hi s concept Fr o m t he
mo me n t wh e n t he gas t u r b i n e b e c a me a pr a c t i c a l r eal i t y i t was ver y
obvi ous t h a t t he hot c o mp r e s s e d ai r i t e xha us t e d c o n t a i n e d h u g e a mo u n t s
o f heat . The r e f or e , t he c o mb i n e d cycl e was c ons i de r e d i n s ome d e p t h
a l mos t as s oon as t he gas t u r b i n e was r el eased f r om t he c ons t r a i nt s o f
mi l i t a r y appl i cat i ons . Howe ve r , be c a us e of t he i r use o f gases at e xt r e me l y
hi gh t e mp e r a t u r e s , ear l y ma c h i n e s suf f er ed f r om l i mi t e d bl ade life a n d
t h e y we r e t he r e f or e us ed onl y i n a ppl i c a t i ons wh e r e n o o t h e r s our ce o f
p o we r was r eadi l y avai l abl e. Wi t h i mp r o v e me n t s i n ma t e r i a l s t e c h n o l o g y
t hi s di f f i cul t y has been o v e r c o me a nd, nowa da ys , c ombi ne d- c yc l e pl a nt s
empl oyi ng gas t urbi nes form t he mai nst r eam of moder n power-st at i on
devel opment .
But whet her it is in a combi ned-cycl e pl ant or a simple-cycle power
station, our interest in this chapt er is in st eam and its use, and this vapour
will now be exami ned in mor e detail. We shall see t hat what seems a fairly
simple and commonpl ace t hi ng is, in fact, qui t e complex.
I n spite of its complexities it is i mpor t ant to tackle this subject in some
dept h, because t he power-pl ant cont rol and i nst r ument at i on engi neer will
need to deal wi t h t he physical par amet er s of st eam t hr ough t he various
stages of designing or using a pract i cal system.
1.6 S u mma r y
In t he above sections we have looked at the nat ur e of st eam and briefly
expl ai ned how it is deri ved and used in various part s of t he power station.
We have also studied simple and combi ned cycles, and seen t hat t he l at t er
provi de an oppor t uni t y of achi evi ng hi gher efficiencies, t her eby maxi-
mising t he revenue ear ned by t he plant.
I n t he following chapt ers we shall look at t he pl ant in mor e detail,
st art i ng wi t h t he wat er and st eam circuits and t hen movi ng on to discuss
t he combust i on process. Once t he pl ant is underst ood, t he principles of its
cont rol systems can be bet t er appreci at ed.
1.7 Re f e r e nc e s
1 RANKI NE, W.J.M.: ' A manual of the steam engine and other prime
movers' (Griffin, London, 1908)
Chapter 2
Th e s t e a m a nd wa t e r c i r c u i t s
2. 1 S t e a m g e n e r a t i o n a n d u s e
I n a convent i onal t her mal power pl ant , the heat used for st eam generat i on
may be obt ai ned by bur ni ng a fossil fuel, or it may be deri ved from the
exhaust of a gas t urbi ne. In a nucl ear pl ant t he heat may be deri ved from
t he radi oact i ve decay of a nucl ear fuel. In this chapt er we shall be
exami ni ng the wat er and st eam circuits of boilers and HRSGs , as well as
the st eam t urbi nes and t he pl ant t hat ret urns the condensed st eam to the
boiler.
In the t ype of pl ant bei ng consi dered in this book, t he wat er is
cont ai ned in t ubes lining the walls of a chamber which, in t he case of a
simple-cycle pl ant , is called the furnace or combust i on chamber. In a
combi ned-cycl e pl ant the t ubes form par t of the HRSG. In ei t her case, the
appl i cat i on of the heat causes convect i on current s to form in t he wat er
cont ai ned in the tubes, causing it to rise up to a vessel called t he drum, in
whi ch the st eam is separ at ed from the water. In some designs of pl ant t he
process of nat ur al ci rcul at i on is augment ed by forced circulation, the
wat er bei ng pumpe d t hr ough t he evapor at i ve circuit r at her t han al l owed
to ci rcul at e by convect i on.
Thi s book concent rat es on pl ant where a dr um is provi ded, but it is
wor t h ment i oni ng anot her t ype of pl ant where wat er passes from the
liquid to the vapour stage wi t hout the use of such a separat i on vessel. Such
' once-t hrough' boilers requi re feed-wat er and st eam- t emper at ur e cont rol
philosophies t hat differ qui t e significantly from those descri bed here, and
t hey are out si de the scope of this book.
Fi gure 2.1 shows a dr um boi l er in schemat i c form. Her e, the st eam gen-
erat i on occurs in banks of t ubes t hat are exposed to the r adi ant heat of
combust i on. Of course, wi t h HRS G pl ant no r adi ant energy is avai l abl e
I Drum ~,~
I - - r- - - r I
Figure 2.1 Schematic of a boiler
Feed
water
(since the combustion process occurs within the gas turbine itself) and the
heat of the gas-turbine exhaust is transferred to the evaporator tubes by a
mixture of convection and conduction. In this type of plant it is common
to have two or more st eam/ wat er circuits (see Figure 2.6), each with its
own steam drum, and in such plant each of these circuits is as described
below.
The steam leaves the drum and enters a bank of tubes where more
heat is taken from the gases and added to the steam, superheating it before
it is fed to the turbine. In the di agram this part of the plant, the super-
heater, comprises a single bank of tubes but in many cases multiple stages
of superheater tubes are suspended in the gas stream, each abstracting
additional heat from the exhaust gases. In boilers (rather than HRSGs),
some of these tube banks are exposed to the radiant heat of combustion
and are therefore referred to as the radiant superheater. Others, the con-
vection stages, are shielded from the radiant energy but extract heat from
the hot gases of combustion.
After the flue gases have left the superheater they pass over a third set
of tubes (called the economiser), where almost all of their remaining heat
is extracted to prewarm the water before it enters the drum.
Steam and water circuits 15
Fi nal l y t he last of t he heat in t he gases is used to war m t he ai r t hat is to
be used in t he process of bur ni ng t he fuel. (This ai r heat er is not shown in
t he di agr am since it is par t of t he ai r and gas pl ant whi ch is discussed in t he
next chapter. )
The maj or movi ng items of machi ner y shown in t he di agr am are t he
feed pump, whi ch delivers wat er to t he system, and t he fan whi ch provides
t he ai r needed for combust i on of t he fuel (in most plants each of these is
dupl i cat ed). I n a combi ned-cycl e pl ant t he pl ace of t he combust i on-ai r fan
and t he fuel firing system is t aken by t he gas t urbi ne exhaust.
Fi gure 2.1 shows onl y t he maj or items associated wi t h t he boiler. I n a
power-generat i on station, t he st eam passes to a t urbi ne aft er whi ch it has
to be condensed back to wat er, whi ch necessitates t he use of a heat
exchanger to ext ract t he last r emai ni ng vestiges of heat from t he fluid and
fully condense it into a liquid. Then, ent r ai ned ai r and gas has to be
r emoved from t he condensed fluid before it is r et ur ned to t he boiler.
The maj or r emai ni ng pl ant items f or mi ng par t of t he s t eam/ wat er
cycle will now be briefly described and t hei r operat i ons explained.
2.2 The s t e a m t ur bi ne
I n pl ant s using a t urbi ne, t he ener gy in t he st eam gener at ed by the boiler
is first conver t ed to kinetic energy, t hen to mechani cal rot at i on and finally
to electrical energy. On leaving t he t urbi ne t he fluid is fed to a condenser
whi ch compl et es t he conversi on back to wat er, whi ch is t hen passed to
f ur t her stages of processing before bei ng fed to t he feed pumps. I n t he
following par agr aphs, we shall exami ne this process (with t he except i on of
t he conversi on to electrical ener gy in t he al t ernat or).
I n t he t urbi ne, t he st eam is fed via nozzles onto successive rows of
blades, of whi ch al t er nat e rows are fixed to t he machi ne casing wi t h t he
i nt er medi at e rows at t ached to a shaft (Figure 2.2). I n this way t he heat
ener gy in t he st eam is conver t ed first to kinetic ener gy as it enters t he
machi ne t hr ough nozzles, and t hen this kinetic energy is convert ed to
mechani cal work as it i mpi nges ont o t he rot at i ng blades. Fur t her work is
done by t he react i on of t he st eam leaving these blades when it encount ers
anot her set of fixed blades, whi ch in t ur n redi rect it ont o yet anot her set of
rot at i ng blades. As t he st eam travels t hr ough t he machi ne in this way it
cont i nual l y expands, giving up some of its ener gy at each ri ng of blades.
The mome nt of rot at i on applied to t he shaft at any one ri ng of blades is t he
mul t i pl e of t he force applied to t he blades and mean di st ance of t he force.
Since each stage of rings abstracts ener gy f r om t he steam, t he force applied
at t he subsequent stage is less t han it was at t he precedi ng ri ng and,
[ Fixed blades
/ - -
t
Steam
Rotating blades ]
Figure 2.2 Turbine blading
therefore, to ensure t hat a const ant moment is appl i cd to the shaft at
each stage, the length of the blades in all rings after the first is made longer
t han t hat of the precedi ng ring. Thi s gives the t urbi ne its charact eri st i c
t aperi ng shape. The st eam enters the machi ne at the set of bl ades wi t h the
smallest di amet er and leaves it after the set of blades with the largest
di amet er. On the cont rol di agr ams present ed in this book, this is i ndi cat ed
by the usual symbol for a turbine, a rhomboi dal shape (Figure 2.3).
Turbi nes may consist of one or more stages, and in pl ant which uses
reheat i ng the st eam exiting the high-pressure or i nt ermedi at e stage of the
machi ne (the HP or I P stage, respectively) is r et ur ned to the boi l er for
addi t i onal heat to be added to it in a bank of tubes called the reheater. The
st eam leaving this stage of the boiler enters the final stage of the machine,
the low pressure (I,P) stage. Because the energy avai l abl e in the steam is
now much less t han it was at the HP stage, this par t of t he t urbi ne is char-
acterised by ext remel y long blades.
By the t i me it leaves the final stage of the t urbi ne, the steam has
exhaust ed al most all of the energy t hat was added to it in thc steam
generator, and it is therefore passed to a condenser where it is finally
St e a m in
\
Shaft coupling
the turbine to
the driven load
(generator etc.)
Steam out
Figure 2. 3 Symbolic representation o f a turbine
cool ed to convert it back to wat er which can be re-used in the cycle. The
condenser comprises a heat exchanger t hrough which cold wat er is circu-
lated. A simplified represent at i on of the compl et e circuit is shown in
Fi gure 2.4.
The cooling wat er t hat is pumped t hrough the condenser to abst ract
heat from the condensat e may itself be flowing t hough a closed circuit.
Alternatively, it may be dr awn from a river or the sea to which it is t hen
ret urned. In the l at t er cases, because of the heat received from the
condenser, care must be t aken to avoid undesi rabl e heat i ng of the river or
sea in the vicinity of t he discharge (or outfall).
In a closed circuit, the heat is released to the at mosphere in a cooling
tower. Wi t hi n these, the air t hat is used for cooling the wat er may ci rcul at e
t hrough the t ower by nat ural convection, or it may be fan-assisted. It is
usually desirable to minimise the format i on of a pl ume since, as well as
bei ng very visible, such pl umes can cause di st urbance to the near by envir-
onment by falling as a fine rain and possibly freezing on roads.
2.3 The c o n d e n s a t e and f e e d- wa t e r s y s t e m
Inside the plant, the st eam and wat er system forms a closed loop, wi t h
the wat er leaving the condenser bei ng fed back to the feed pumps for re-
use in the boiler. However, cert ai n ot her items of pl ant now become
1
8

P
o
w
e
r
-
p
l
a
n
t

c
o
n
t
r
o
l

a
n
d

i
n
s
t
r
u
m
e
n
t
a
t
i
o
n

o

I

,
o

C

o

m

o

o

0

o

u
l

I
u

~
o

m

~
.

o
~
_

~

~

r
.
-

o

I

I

I

I

I

I

J

=

C

8

I
L
l
:

I

/

_
/

,

s

I

v

,
,
m

~

=

Y
.

involved, because t he wat er leaving t he condenser is cold and cont ai ns
ent r ai ned air whi ch must be removed.
Ai r becomes ent r ai ned in t he wat er system at st art -up (when t he
various vessels ar e initially empt y), and it will appear duri ng nor mal
operat i on when it leaks in at those part s of t he cycle whi ch operat e below
at mospheri c pressure, such as t he condenser, ext ract i on pumps and low-
pressure feed heaters. Leakage can occur in these areas at flanges and at
t he sealing glands of t he rot at i ng shafts of pumps. Ai r ent r ai nment is ai ded
by two facts: one is t hat cold wat er can hold great er amount s of oxygen
(and ot her dissolved gases) t han can war m water; and t he ot her is t hat t he
low-pressure part s of t he cycle must necessarily correspond wi t h t he low-
t emper at ur e phases.
The presence of residual oxygen in t he feed-wat er supply of a boiler or
HRSG is hi ghl y undesirable, because it will cause corrosion of t he boiler
pi pework (part i cul arl y at welds, col d-worked sections and surface disconti-
nuities), great l y reduci ng t he serviceable life-span of t he plant. For this
reason great at t ent i on must be pai d to its removal.
Removal of dissolved oxygen is per f or med in several ways, and an
i mpor t ant cont ri but or to this process is t he deaer at or whi ch is shown in
Figure 2.4, l ocat ed bet ween t he condenser ext ract i on pump and t he boiler
feed-wat er pump.
2. 3. 1 The deaerator
The deaer at or removes dissolved gases by vigorously boiling t he wat er
and agi t at i ng it, a process referred to as ' stripping' . One t ype of deaer at or
is shown in Figure 2.5. I n this, t he wat er ent eri ng at t he t op is mi xed with
st eam whi ch is rising upwards. The steam, t aken di rect l y from t he boiler or
from an ext ract i on poi nt on t he turbine, heats a stack of met al trays and as
t he wat er cascades down past these it mixes wi t h the st eam and becomes
agitated, releasing t he ent r ai ned gases. The st eam pressurises t he
deaer at or and its cont ent s so t hat t he dissolved gases are vent ed to t he
at mosphere.
Mi ni mi si ng corrosion requires t he feed-water oxygen concent rat i on to
be mai nt ai ned below 0.005 ppm or less and al t hough t he deaer at or
provides an effective met hod of removi ng t he bulk of ent r ai ned gases it
cannot reduce t he concent r at i on bel ow about 0.007 ppm. For this reason,
scavenging chemicals ar e added to remove t he last traces of oxygen.
2.3.1.1 Chemical dosing
Volatile oxygen scavengers such as hydr azi ne (N2H4) and sodi um
sulphite (Na2SOs) have been used for oxygen removal (al t hough
W a t e r in
V e n t . ~ ] A V e n t
t r a y s t a c k
Steam,
' F " ] F ] F S t o r a g e v e s s e l
; u - - /
, I , ' 1 '
w a t e r p u m p
Figure 2.5 Principle of a deaerator
hydrazi ne is now suspect ed of bei ng carcinogenic). What ever their form,
the chemi cal scavengers are added in a concent rat ed form and it is
necessary to flush the injection pipes cont i nual l y or on a peri odi c basis to
prevent plugging. Similarly, bl owdown, a process of bl eedi ng wat er to
drai ns or a special vessel, is used to cont i nual l y or peri odi cal l y remove a
por t i on of t he wat er from the boiler, with aut omat i c or manual chemi cal
sampl i ng bei ng used to ensure t hat the correct concent r at i on is mai nt ai ned
in t he boi l er wat er.
From a cont rol and i nst rument at i on vi ewpoi nt , t he above chemi cal
dosing operat i ons are highly specialised and are t herefore usually
per f or med by equi pment t hat is suppl i ed as par t of a wat er - t r eat ment
pl ant package. The cont rol system (often based on a pr ogr ammabl e- l ogi c
cont rol syst em ( PLC) ) will generat e dat a and al arm signals for connect i on
to the mai n pl ant comput er - cont r ol system (frequent l y referred to as the
di st ri but ed cont rol syst em (DCS).)
Aft er the wat er has been deaer at ed and t reat ed, it is fed to feed pumps
whi ch deliver it back to t he boi l er at high pressure.
2.4 Th e f e e d p u mp s and v a l v e s
The feed pumps deliver wat er to the boi l er at high pressure, and t he flow
into the system is cont rol l ed by one or mor e feed-regul at i ng valves. The
feed pumps are general l y dri ven by electric motors, but small st eam
t urbi nes are also used (al t hough, clearly, these cannot be used at st art -up
unless a separat e source of st eam is avai l abl e for their operat i on).
The pr essur e/ f l ow charact eri st i c of pumps and the vari ous configura-
tions t hat are avai l abl e are discussed in Chapt er 6 but it shoul d be not ed
here t hat wi t h any pump the pressure tends to fall as the t hr oughput rises.
On t he ot her hand, due to the effect of friction, the resistance offered by
the boi l er system to the flow of wat er increases as the flow rat e increases.
(The syst em resistance is the mi ni mum pressure t hat is requi red to force
wat er into the boiler.) Ther ef or e t he pressure dr op across the val ve will be
highest at low flows.
It is wasteful to oper at e wi t h a pressure dr op t hat is significantly above
t hat at whi ch effective cont rol can be mai nt ai ned, bot h because this entails
an energy loss and also because erosion of val ve internals increases with
high pressure-drops. Wi t h fixed-speed pumps there is not hi ng t hat can be
done about this, but an i mpr ovement can be made if vari abl e-speed pumps
are used. These are mor e expensive t han their fixed-speed count erpart s,
but the increase in cost tends to be offset by the operat i onal cost savings
t hat can be achi eved (due to mor e efficient operat i on and reduced wear on
the valve). Such savings are i ncreased i f the pl ant operat es for prol onged
peri ods at low t hr oughput s and are most appar ent wi t h the l arger
boilers.
Fr om t he cont rol engineer' s vi ewpoi nt , vari abl e-speed pumps are an
at t ract i ve opt i on because t hey enabl e t he cont rol -syst em dynami cs to be
linearised over a wi de range of flows, l eadi ng to i mpr oved controllability.
However , the decision on their use will general l y be made by mechani cal
and process engineers, and will be based pur el y on economi c grounds.
2.5 Th e wa t e r and s t e a m c i r c ui t s o f HRSG pl ant
In the combi ned- cycl e pl ant the task of boiling the feed wat er and super-
heat i ng the st eam so pr oduced is achi eved by using the consi derabl e heat
Pressure-reducing and
desuperheating valve
D( ~ I ( Ni t Of
Steam turbo/Generator
.[' G e n e r a t o r
i .
B y p a s s v a l v e s
Condenser
" [
I G a s - t u r b i n e / G e n e r a t o r
i
I
!
i . . . . . . . . . . . . . . . . . . . . . . . . . . i
Figure 2.6 Gas turbines in a combined-cycle system
+
cont ent of t he exhaust from a gas turbine, sometimes wi t h and sometimes
wi t hout suppl ement ar y firing.
The vari et y of pl ant ar r angement s in use is very wi de and al t hough
the following description relates to onl y one confi gurat i on, it should enable
t he general nat ur e of these systems to be underst ood.
I n some plants t he gas and st eam turbines and t he gener at or are on
t he same shaft, others have separat e generat ors for t he gas and st eam
turbines. The installation shown in Fi gure 2.6 is of t he l at t er variety, and
t he di agr am shows j ust one gas t urbi ne and HRSG from several at this par-
t i cul ar plant.
St art i ng at t he condenser outlet, t he circuit can be t r aced t hr ough t he
ext ract i on pump and via the economi ser to t he deaerat or. From her e two
circuits are formed, one feeding t he LP section, t he ot her t he HP section.
These systems ar e of t he forced-ci rcul at i on t ype and ar e qui t e similar to
each ot her in layout, but t he st eam leaving t he HP side passes to a super-
heat er bank whi ch is positioned to receive t he hottest par t of the exhaust
from t he gas turbine. The superheat ed st eam goes to t he HP stage of t he
st eam t urbi ne and t he st eam leaving this stage goes to t he LP stage.
Sat ur at ed st eam from t he LP section of t he HRSG also enters t he t urbi ne
at this point. Bypass valves are empl oyed dur i ng st art -up and shut -down
and enabl e t he pl ant to operat e with only t he gas t urbi ne in service, under
whi ch condi t i on t he st eam from t he HP and LP stages is bypassed to t he
condenser
2.6 Summary
So far, we have studied t he nat ur e of steam, and t he pl ant and auxiliaries
t hat ar e empl oyed in t he process of generat i ng and using t he fluid. Now we
need to under st and t he mechani sms involved in obt ai ni ng t he heat t hat is
requi red to gener at e t he steam. This process involves t he fuel, air and flue-
gas circuits of t he plant, and all t he maj or equi pment requi red for cl ean
and efficient operation.
Chapt er 3 describes t he combust i on chamber (or furnace) and t he
pl ant and firing ar r angement s t hat are empl oyed in bur ni ng a vari et y of
fuels. I n addition, t he chapt er outlines how t he ai r requi red for combust i on
is obt ai ned, war med and distributed, and discusses t he characteristics and
limitations of t he pl ant involved in this process.
Chapter 3
Th e f ue l , ai r a nd f l ue - g a s c i r c u i t s
Havi ng looked at the st eam and wat er circuits of boilers and HRSGs , we
now move on to exami ne the pl ant which is involved in the combust i on of
fuel in boilers.
The heat used for generat i ng the steam is obt ai ned by burni ng fuel in a
furnace, or combust i on chamber, but to do this requires the provision of
air which is provi ded by a forced-draught (FD) fan (in l arger boilers, t wo
such fans are provi ded). Aft er the fuel has been burned, the hot product s of
combust i on are ext ract ed from the furnace by anot her fan, the i nduced-
dr aught (ID) fan, and fed to the chimney. Again, t wo I D fans are
provi ded on l arger boilers.
In this chapt er we shall exami ne not onl y the burners or ot her
equi pment used to bur n the fuel but also the fans and air heaters. Finally
we shall briefly exami ne how gas turbines are used in combi ned-cycl e
plant.
3. 1 T h e f u r n a c e
In boi l er pl ant the heat used for boiling the wat er is obt ai ned by burni ng
a fossil fuel (unlike the HRSG, where the heat is del i vered by the exhaust
of a gas turbine). Thi s process of combust i on is carri ed out in the furnace,
and comprises a chemi cal react i on bet ween the combust i bl e mat eri al and
oxygen. I f insufficient oxygen is avai l abl e some of' the combust i bl es will not
burn, whi ch is clearly inefficient and polluting. On the ot her hand, the
provision of too much oxygen leads to inefficient operat i on and to
corrosi on and undesi rabl e emissions from the stack due to the combi nat i on
of the surplus oxygen with ot her component s of the flue gases.
T h e oxyge n f or c o mb u s t i o n is p r o v i d e d i n ai r, wh i c h c ont a i ns a r o u n d
21% o f t he gas. Howe ve r , ai r al so c ont a i ns a r o u n d 77% ni t r oge n, a n d t he
c o mb u s t i o n pr ocess r esul t s i n t he p r o d u c t i o n o f n i t r o g e n di oxi de (NO2)
a n d ni t r i c oxi de ( NO) . Th e s e gases ( pl us ni t r ous oxi de, NzO) ar e col l ec-
t i vel y cal l ed n i t r o g e n oxi des, or NOx f or s hor t , a n d becaus e t h e y ar e of t en
b l a me d f or va r i ous d e t r i me n t a l effects on t he e n v i r o n me n t a h i g h l evel of
a t t e n t i o n mu s t be gi ven t o mi n i mi s i n g t he i r p r o d u c t i o n .
Un f o r t u n a t e l y , h i g h c o mb u s t i o n efficiencies i nva r i a bl y c o r r e s p o n d wi t h
t he p r o d u c t i o n o f h i g h levels o f n i t r o g e n oxi des, a n d t he r e f or e NOx
r e d u c t i o n i nvol ves car ef ul de s i gn o f t he bur ne r s so as t o yi el d a d e q u a t e
c o mb u s t i o n effi ci ency wi t h mi n i ma l s mo k e a n d c a r b o n mo n o x i d e ge ne r a -
t i on.
3.1.1 Fi ri ng arrangements
T h e c o mb u s t i o n o f oil, gas or pul ver i s ed coal is p e r f o r me d i n bur ner s .
Th e s e ma y be a r r a n g e d o n one wal l o f t he c o mb u s t i o n c h a mb e r ( whi ch is
t he r e f or e cal l ed ' f r ont - f i r ed' ) , or on f aci ng wal l s ( ' oppos e d fi red' ) or at t he
c or ne r s o f i t ( ' cor ner - f i r ed' or ' t a nge nt i a l ' ) , a n d t he char act er i s t i cs of c om-
bus t i on wi l l be ver y di f f e r e nt i n e a c h case. T h e b u r n e r s ma y be p r o v i d e d
wi t h i ndi vi dua l l y c ont r ol l e d fuel a n d ai r suppl i es, or c o mmo n c ont r ol ma y
be a ppl i e d f or al l t he bur ne r s , or t he y ma y be o p e r a t e d i n gr oups , e a c h
g r o u p h a v i n g d e d i c a t e d a n d s e pa r a t e l y c o n t r o l l e d suppl i es o f fuel a n d air.
Co mb u s t i o n o f r a w coal or o t h e r sol i d fuels s uch as mu n i c i p a l wast e,
cl i ni cal was t e or r ef us e- der i ved fuels is of t en c a r r i e d o u t i n f l ui di sed beds,
or on st oker s cons i s t i ng of mo v i n g gr at es or pl at f or ms .
T h e me t h o d s o f c ont r ol l i ng t hese var i ous a r r a n g e me n t s ar e ver y
di f f er ent . Wi t h f r ont - f i r ed or oppos e d- f i r e d boi l ers t he t e mp e r a t u r e o f t he
fl ue gases a n d t he r e s ul t i ng h e a t t r ans f er t o t he var i ous ba nks of super -
h e a t e r t ubes is a dj us t e d by b r i n g i n g bur ne r s i nt o servi ce or t a ki ng t h e m
o u t o f servi ce, a n d t hi s ma y be d o n e i ndi vi dua l l y or i n banks . Th i s me t h o d
pr ovi de s a s t e p- f unc t i on t y p e o f c ont r ol a n d fi ne a d j u s t me n t o f s t e a m t e m-
p e r a t u r e is p r o v i d e d by s pr a y- wa t e r a t t e mp o r a t i o n .
Cor ne r - f i r e d ( t a nge nt i a l ) boi l er s ar e a r r a n g e d i n s uch a way t h a t t he
b u r n i n g fuel ci r cul at es a r o u n d t he f ur na c e , f o r mi n g a l ar ge s wi r l i ng bal l o f
b u r n i n g fuel at t he cent r e. Wi t h t hi s t ype of boi l er t he ma n u f a c t u r e r s
us ual l y e mp l o y t i l t i ng me c h a n i s ms t o di r e c t t he f i r ebal l t o a h i g h e r or
l ower p o s i t i o n wi t hi n t he f ur na c e , a n d t hi s has a si gni f i cant effect on t he
t e mp e r a t u r e o f t he va r i ous ba nks o f s u p e r h e a t e r t ubes, a n d t he r e f or e on
s t e a m t e mp e r a t u r e . T h e ma x i mu m de gr e e of t i l t t h a t is avai l abl e wi t hi n
t he basi c de s i gn is t ypi cal l y -t-30 , a l t h o u g h t he de gr e e of mo v e me n t
e mp l o y e d i n pr a c t i c e is us ual l y r es t r i ct ed d u r i n g c ommi s s i oni ng.
The fuel, air and flue-gas circuits 27
The downside of tilting is t hat bur ne r s - - wi t h t hei r fuel and ai r
supplies, igniters, fl ame moni t ors e t c . - - a r e compl ex things, and tilting
t hem requires very careful engi neeri ng i f it is to be successful. Also, t he
tilting mechani sms must be rigorously mai nt ai ned i f t hey are to cont i nue
to operat e effectively over any l engt h of time.
The cont rol systems t hat regul at e bur ner tilting mechani sms must
ensure t hat exact l y t he same degree of tilt is applied to t he burners at all
four corners of t he furnace, since any mi sal i gnment will cause t he fireball
to ci rcul at e helically r at her t han as required.
3. 2 T h e a i r a n d g a s c i r c u i t s
The combust i on process requires t he provision of fuel and air in t he
correct rat i o to each other. This is known as t he stoichiometic ratio, and
under this condi t i on enough air is provi ded to ensure compl et e combust i on
of all t he fuel, wi t h no surplus or deficit. However, this is a t heoret i cal
ideal, and pract i cal considerations may necessitate operat i ng at a f uel / ai r
rat i o t hat is di fferent from t he stoichiometric value. I n addition, it must be
underst ood t hat t he efficiency of t he combust i on process will also be
affected by t he t emper at ur e of t he air provided.
I n t he following sections we shall see how air is delivered to t he
f ur nace at t he ri ght conditions of flow and t emper at ur e, st art i ng wi t h t he
auxi l i ary pl ant t hat warms t he air and movi ng on to t he types of fan
empl oyed in the dr aught plant.
3.2.1 The air heater
I n a simple-cycle plant, air is del i vered to t he boiler by one or mor e FD
fans and t he product s of combust i on are ext r act ed from it by I D fans.
Fi gure 3.1 shows this pl ant in a simplified form, and illustrates how t he
heat r emai ni ng in t he exhaust gases leaving t he furnace is used to war m
t he air being fed to t he combust i on chamber. This funct i on is achi eved in
an air heater, whi ch can be ei t her regenerat i ve, where an i nt ermedi at e
medi um is used to t ransfer t he heat from t he exhaust gases to t he i ncomi ng
air, or recuperat i ve, wher e a di rect heat t ransfer is used across a di vi di ng
partition.
One vari et y of regenerat i ve air heat er is t he Lj ungst r6m type, where
met al plates mount ed on a rot at i ng frame are passed t hr ough t he hot gases
and t hen to t he i ncomi ng air.
From a control engineer' s poi nt of view, an i mpor t ant consi derat i on is
the efficient combust i on of t he fuel, and here it is necessary to consider t he
Induced-
draught fan
Exhaust
gases to
chimney
Forced-draught fans
-
~ . j - /
Induced- S
draught fan
Figure 3.1 Draught-plant arrangement
Exhaust
gases to
chimney
losses and l eakages t hat occur in an air heater. Figure 3.2 shows how
vari ous leakages occur in a t ypi cal air heater: across the ci rcumferent i al ,
radi al and axial seals, as well as at the hub. These leakages are minimised
when the pl ant is first const ruct ed, but become great er as wear occurs
dur i ng pr ol onged usage. When the sheer physical size of the air heat er is
consi dered (Fi gure 3.3) it will be appr eci at ed t hat these leakages can
become significant.
3.2.2 Types o f f an
In addi t i on to t he FD and I D fans ment i oned above, anot her appl i cat i on
for l arge fans in a power-st at i on boiler is where it is necessary to over come
the resistance present ed by pl ant in the pat h of the flue gases to the stack.
A i r P reheater L eakage
The f uel , air and flue-gas circuits 29
Figure 3.2 Ai r heater leakage
Howden Sirocco Ltd. Reproduced with permission
Circ
/ea~
Figure 3.3 An air heater being lifted into position
I n some cases, envi r onment al legislation has enforced t he fitting of flue-
gas desul phuri sat i on equi pment to an existing boiler. Thi s involves t he use
of absorbers a n d / o r bag filters, plus t he at t endant duct i ng, all of whi ch
present addi t i onal resistance to the flow of gases. In this case this resistance
was not ant i ci pat ed when the pl ant was originally designed, so it is
necessary to fit addi t i onal fans to over come t he dr aught losses. These are
called ' boost er fans'.
What ever t hei r function, as far as t he fans themselves are concerned,
t wo t ypes are found in power-st at i on dr aught applications: centrifugal
(Figure 3.4) and axial-flow (Fi gure 3.5). In t he former, the bl ades are set
radi al l y on t he dri ve shaft wi t h the air or flue gas di rect ed to the cent re
and dri ven out war ds by cent ri fugal force. Wi t h axial-flow fans, t he air or
gas is dr awn al ong the line of the shaft by the screw act i on of t he blades.
Wher eas the bl ades of a cent ri fugal fan are fixed rigidly to t he shaft, t he
pi t ch of axial-flow fan bl ades can be adjusted. Thi s provi des an efficient
means of cont rol l i ng the fan' s t hroughput , but requires careful design of
t he associated cont rol syst em because of a phenomenon known as 'stall',
whi ch will now be described.
Figure 3.4 Centrifugal f an
Figure 3.5 Axialz[low f an
3. 2. 2. 1 The st al l condition
The angul ar relationship bet ween the air flow i mpi ngi ng on the bl ade of
a fan and the bl ade itself is known as the ' angle of attack' . In an axial-flow
fan, when this angle exceeds a cert ai n limit, the air flow over t he bl ade
separat es from t he surface and centrifugal force t hen t hrows the air
out wards, t owards the rim of the blades. Thi s act i on causes a bui l d-up of
pressure at the bl ade tip, and this pressure increases until it can be relieved
at the cl earance bet ween the tip and the casing. Under this condi t i on the
operat i on of the fan becomes unstable, vi brat i on sets in and the flow starts
to oscillate. The risk of stall increases i f a fan is oversized or i f t he system
resistance increases excessively.
For each setting of t he bl ades there is a poi nt on the fan charact eri st i c
beyond which stall will occur. I f these points are linked, a 'stall line' is
gener at ed (Fi gure 3.6) and if this is built into the pl ant cont rol system
(DCS) it can be used to war n the oper at or t hat the condi t i on is i mmi nent
and t hen to act i vel y shift oper at i on away from the danger region. The
act ual stall-line dat a for a given machi ne should be provi ded by the fan
manuf act ur er .
3.2.2.2 Centrifugal-fan surge
The stall condi t i on affects onl y axial-flow fans. However , centrifugal t~ns
are subj ect to anot her form of instability. I f t hey are oper at ed near the
peak of their pr essur e/ f l ow curve a small movement ei t her way can cause
the pressure to increase or decrease unpredi ct abl y. The poi nt at whi ch this
phenomenon occurs is known as the ' surge limit' and it is the mi ni mum
flow at which t he fan operat i on is stable.
g
I I I I I I I I I I
V o l u m e t r i c f ! C : ~
1 . . . . I
! 0 ~
Figure 3.6 The stall line of an axial-flow fan
The system desi gner needs to be aware of t he risk of surge occurring,
since it may be necessary to adapt the control-system design. However,
this is general l y not a probl em i f the fan is properl y designed in relation to
t he overall plant. Dur i ng the initial design of the control system, di al ogue
with the process engi neer or boiler designer will show whet her or not surge
prot ect i on will be requi red.
3.2.3 Final elements f or draught control
Reference has al r eady been made to the use of pitch-control in axial-flow
fans to regul at e the t hr oughput of t he machi ne. Ot her means of cont rol l i ng
flow are dampers, vanes or speed adj ust ment . Each of these devices has its
own characteristics, advant ages and disadvantages, and the selection of
t he cont rol l i ng device whi ch is to be used in a given appl i cat i on will be a
t r ade- of f bet ween the technical features and t he cost.
3.2.3.1 Types of damper
The simplest form of damper consists of a hi nged plate t hat is pivoted at
the cent re so t hat it can be opened or closed across the duct. This provides
a form of dr aught cont rol but it is not very linear and it is most effective
only near t he closed position. Once such a damper is mor e t han about 40-
60% open it can provide very little addi t i onal control. Anot her form of
damper comprises a set of linked blades across the duct (like a Venet i an
blind). Such muhi bl aded dampers are nat ur al l y mor e expensive and mor e
compl ex to mai nt ai n t han single-bladed versions, but t hey offer bet t er
l i neari t y of control over a wi der range of operation.
The task of designing a control system for opt i mum pert brmance over
the widest dynami c range will be simplified if t he relationship bet ween the
cont rol l er out put signal and the resul t ant flow is linear. Al t hough it is
possible to provide the requi red charact eri sat i on wi t hi n the cont rol system,
this will usually onl y be effective under aut omat i c control. Under manual
cont rol a severely nonl i near charact eri st i c can make it difficult for the
operat or to achieve precise adj ust ment .
It is possible to linearise the command- f l ow relationship under both
manual and aut omat i c control by the design of the mechani cal linkage
bet ween t he act uat or and the damper. However, this requires careful
design of the mechani cal assemblies and these days it is general l y consid-
ered simpler to build the requi red charact eri sat i on into t he DCS. This
appr oach provides a partial answer, but it should not be forgot t en t hat
such a solution is onl y effective under aut omat i c control.
3.2.3.2 Vane control
The second form of cont rol is by t he adj ust ment of vanes at t he fan inlet.
These vanes are cl earl y visible near t he cent re of Fi gure 3.7 (which shows a
cent ri fugal fan dur i ng manuf act ur e) . Such vanes are oper at ed via a
compl ex l i nkage whi ch rotates all t he vanes t hr ough t he same angl e in
response to t he c omma nd signal from t he DCS.
3.2.3.3 Variable-speed drives
Finally, cont rol of fan t hr oughput can be achi eved by t he use of variable-
speed mot ors (or drives). These may involve t he use of electronic control-
lers whi ch al t er t he speed of t he dri vi ng mot or in response to demand
signals from t he DCS or t hey can be hydraul i c couplings or variable-speed
gearboxes, ei t her of whi ch allows a fixed-speed mot or to dri ve t he fan at
t he desired speed. Vari abl e speed drives offer significant advant ages in t hat
t hey allow t he fan to operat e at t he opt i mum speed for t he requi red
t hr oughput of air or gas, whereas damper s or vanes cont rol t he flow by
restricting it, whi ch means t hat t he fan is at t empt i ng to deliver mor e flow
t han is requi red.
Figure 3.7 A centrifugal fan during manufacture
3.3 Fuel s y s t e ms
Fossil fuels t hat are bur ned in boilers can be used in solid, liquid or
gaseous form, or a mi xt ur e of these. Nat ural l y, t he handl i ng systems for
these types of fuels differ widely. Moreover, t he vari et y of fuels bei ng
bur ned is enormous. Solid fuels encompass a wide spect rum of coals as well
as wood, t he waste product s of industrial processes, muni ci pal and clinical
waste and refuse-derived fuels. ( The last ar e pr oduced by shreddi ng or
gri ndi ng domestic, commer ci al and industrial waste material. ) Li qui d
fuels can be heavy or light oil, or t he product s of industrial plant. Gas can
be nat ur al or manuf act ur ed, or t he by-product of refineries.
Each of these fuels requires specialised handl i ng and t reat ment , and
the cont rol and i nst rument at i on has to be appropri at e to t he fuel and t he
pl ant t hat processes it.
3. 3. 1 Coal f i r i n g ( pul ver i s ed f u e l )
Al t hough coal can be bur ned in solid form on grates, it is mor e usual to
break it up before feeding it to t he combust i on chamber. The t r eat ment
depends on t he nat ur e of t he coal. Some coals lend themselves to being
gr ound down to a ver y fine powder (called pulverised fuel (PF)) whi ch is
t hen carri ed to t he burners by a st ream of air. Ot her coals are fed to
i mpact mills whi ch use flails or hammer s to break up t he mat er i al before it
is propelled to t he burners by an ai r st ream. The t ype of mill to be used on
a part i cul ar pl ant will be det er mi ned by t he process engineers and it is the
task of t he cont rol engi neer to provi de a system whi ch is appropri at e. To
do this it is necessary to have some under st andi ng of how t he rel evant t ype
of mill operates.
Various types of pulverised-fuel mill will be encount ered, but two ar e
most commonl y used: t he pressurised vertical-spindle ball mill and t he
hori zont al -t ube mill.
3.3.1.1 Vertical-spindle ball mills
Fi gure 3.8 shows t he operat i ng principle of a typical ball mill, such as t he
Babcock ' E' mill. I n this device, t he coal t hat is di scharged from t he
storage hoppers is fed down a cent ral chut e ont o a table where it is crushed
by rot at i ng steel balls. Ai r is bl own into t he crushed coal and carries it, via
adjustable classifier blades, to t he PF pipes t hat t ransport it to t he
burners.
The air t hat carries the fine particles of coal to t he burners is supplied
from a fan called a ' pr i mar y- ai r fan' . This delivers air to t he mill, whi ch
t herefore operates under a pressure which is slightly positive wi t h respect
Se~
PF o u t l e t
Coal i nl et
P F o u t l e t
Cl assi fi er
~tationary
"inding ring
~rimary ai r
inlet
Gearbox
Figure 3.8 Pressurised ball mill
to t he at mospher e outside. Because of this and because of its ot her con-
st ruct i onal features, this t ype of mill is pr oper l y called a ' vertical-spindle,
pressurised ball mill'. The ai r-suppl y system for this t ype of mill is
discussed in mor e det ai l in Section 3.3.1.3.
3.3.1.2 Horizontal tube mills
In a t ube mill (Fi gure 3.9) the coal is fed into a cage t hat rot at es about a
hori zont al axis, at a speed of 18 to 35 rpm. Thi s cage contains a charge of
forged-steel or cast alloy balls (each of which is bet ween 25 mm and
100 mm in di amet er) whi ch are carri ed up the sides of the cage by the
rot at i on, until t hey event ual l y cascade down to the bot t om, onl y to be
car r i ed up again. The coal is pulverised by a combi nat i on of the i mpact
wi t h these balls, at t ri t i on of adj acent particles and crushing bet ween the
balls and the cage and bet ween one ball and another.
In this t ype of mill the crushed mi xt ure is dr awn out of the cage by a
fan, whi ch is called an exhauster. Because of this confi gurat i on, the t ube
Fuel to burners
Exhausters
Fuel to burners
Coal from feeder
A
Rotating drum containing
coal and ball charge
Figure 3.9 Horizontal tube mill
mill runs under a negative pressure, which prevents the fine coal dust
from escaping (as it tends to do wi t h a pressurised mill) However, the
exhausters have to handl e the di rt y and abrasive mi xt ure of coal and air
t hat comes t hrough the mill and t hey therefore requi re more frequent
mai nt enance t han the fans of a pressurised ball mill, whose function is
merel y to t ransport air from the at mosphere to the mill.
3.3 1.3 Air supply systems f or mills
As st at ed above, the crushed coal in a pressurised ball mill is propel l ed to
the burners by a st ream of war m air. Fi gure 3.10 shows the ar r angement
for doi ng this: cool air and heat ed air are mi xed to achieve the desired t em-
perat ure. Thi s t emper at ur e has to be high enough to part i al l y dr y the coal,
but it must not be so high t hat the coal coul d overheat (with the risk of the
coal / ai r mi xt ure igniting inside t he mill or even expl odi ng while it is being
crushed) The war m air is t hen fed to the mill (or a gr oup of mills) by
Hot air
~r
PF outlet
Tempering
air
Primary
air fan
Mill
Figure 3. 10 Primary air f an system f or a ball mill
means of yet anot her fan, called a ' pr i mar y air fan'. It shoul d be not ed
t hat the cool er of t he t wo air st reams is commonl y referred to as
' t emper i ng air' since, because it is obt ai ned from the FD fan exhaust it
may al r eady be slightly war m, and its funct i on is to t emper t he mixture.
Fi gure 3.11 shows the syst em t hat is used wi t h a t ube mill. Her e, hot
air and cold air are agai n mi xed to obt ai n t he correct t emper at ur e for the
air st ream but , because the mill in this case operat es under suction condi -
tions a pr i mar y air fan is not needed, and the col d air is obt ai ned directly
from t he at mosphere. The war med air mi xt ure is agai n fed to the mill as
' pr i mar y air' but in addi t i on a st ream of hot air is fed to the feeder for
t r anspor t at i on and dryi ng purposes.
3. 3. 2 Oi l -f i ri ng systems
I n compar i son wi t h coal, oil involves the use of much less capi t al plant.
On t he face of it, it woul d appear t hat all t hat is requi red is to ext ract the
oil from its st orage t ank and pump it to t he burners. But in real i t y life is
mor e compl i cat ed t han that!
Pr oper ignition of oil depends on t he fuel bei ng broken into small
dropl et s (at omi sed) and mi xed wi t h air. The at omi sat i on may be achi eved
Hot alr
Feeder
Raw coal and ho
Warm ai r
Ambi e nt
ai r
Figure 3. 11 Suction mill air supply system
by expelling t he oil t hr ough a small nozzle (a ' pressure j et ' ), or it may be
achi eved by t he use of compressed ai r or steam.
The fuel oil itself may be light (such as diesel oil or gas oil), or it ma y
be ext remel y viscous and tar-like (heavy fuel oil, commonl y ' Bunker C' ).
The handl i ng system must t herefore be designed to be appropri at e to t he
nat ur e of t he liquid. Wi t h t he heavi er grades of oil, pr ewar mi ng is
necessary, and to prevent it cooling and t hi ckeni ng t he fuel is cont i nual l y
ci rcul at ed to t he burners via a reci rcul at i on system (shown schemat i cal l y
in Fi gure 3.12). The l at t er process is sometimes referred to as 'spill-back'.
When a bur ner is not firing, t he oil circulates t hr ough t he pi pework ri ght
up to t he shut -off valve, whi ch is mount ed as close to t he oil gun as
possible.
Fr om t he poi nt of t he C&I engineer, t he cont rol systems involved wi t h
oil firing ma y i ncl ude any or all of t he following: cont rol l i ng t he t empera-
t ure of t he fuel, t he pressure of t he at omi si ng medi um, and t he
equalisation of t he fuel pressures at various levels on t he bur ner front.
L p C ,
p u mp
F-G
Oi l
heat er
HP
p u mp
Reci rcul at i on
J Bu r n e r 1 - %
- ~ Bu r n e r 2
Figure 3. 12 Simplified oil pumping, heating and recirculation system
~* Bu r n e r 3
~. Bu r n e r 4
3 . 3 . 3 Gas- f i ri ng syst ems
Al t hough i nher ent l y simpler t han ei t her oil or coal-fired systems, gas-
fired boilers have t hei r own complexities. Any escape of gas, part i cul arl y
into confi ned areas, presents considerable hazards, and great car e must
t herefore be t aken to guar d against leakage, for exampl e, from flanges and
t hr ough valves. But nat ur al gas is colourless, and any escape will t herefore
be invisible. Also, it is not safe to rely on odour to det ect leakages. By t he
time an odour has been det ect ed sufficient gas may have al r eady escaped
to present a hazard. It is t herefore necessary for gas-leak det ect ors to be
fitted al ong t he inlet pi pework wher ever l eakage coul d occur, and to
connect these to a comprehensi ve, cent ral al ar m system.
It is also necessary to prevent gas from seeping into the combust i on
chamber t hr ough leaking valves. I f gas does ent er undet ect ed into the
f ur nace dur i ng a shut -down period, it coul d collect in sufficient quantities
to be ignited ei t her by an acci dent al spark or when a bur ner is ignited. The
resulting explosion woul d almost cert ai nl y cause maj or damage and could
endanger lives. (It should be not ed t hat this risk is present with pr opane
igniters such as those used wi t h fuels ot her t han oil.)
The f u e l , ai r and f l ue-gas circuits 41
Prot ect i on against l eakage into the furnace t hr ough the fuel-supply
valves is achi eved by the use of ' doubl e-bl ock-and-bl eed' val ve assemblies
which provi de a secure seal bet ween t he gas inlet and the furnace. The
oper at i on of this system (see Fi gure 3.13) is t hat before a bur ner is ignited
bot h bl ock valves are closed and the vent is open. In this condi t i on any gas
whi ch may occupy the vol ume bet ween the t wo block valves is vent ed to a
safe pl ace and it can t herefore never devel op enough pressure to leak past
the second block valve. When st art -up of the bur ner is requi red, a sequence
of operat i ons opens the block valves in such a way t hat gas is admi t t ed to
the bur ner and ignited safely.
inlet pressure -
r educi ng system
Mai n
saf et y
s hut - of f
valve
Doubl e- bl ock- and- bl eed
assembly
Main
flow-
. | . val ve ~" Bur ner 2
Or
S
Bypass f l o w-
c ont r ol valve
Bur ner 4
Vent
Figure 3. 13 Simplified schematic o f gas-firing system
3. 3. 4 Was t e- t o- ener gy p l a n t s
Ther e has been a st eady devel opment of plants t hat i nci nerat e waste
mat er i al of various types and use t he heat thus pr oduced to gener at e elec-
tricity. Ear l y units suffered from t he unpredi ct abl e nat ur e of t he waste
mat er i al and t he severe corrosi on resulting from t he release of acidic
compounds dur i ng the combust i on process. But t he probl ems have been
l argel y overcome t hr ough t he appl i cat i on of i mpr oved combust i on systems
and by bet t er knowl edge of t he mat eri al s used in t he const ruct i on of t he
plant.
Waste mat er i al may be obt ai ned from any of several sources, i ncl udi ng
t he following:
muni ci pal ;
industrial;
clinical;
agri cul t ural .
The mat er i al may be bur ned aft er very basic t r eat ment (shredding etc.)
or it ma y be processed in some way, in whi ch case t he end result is t er med
refuse-derived fuel (RDF).
Several types of wast e-t o-energy pl ant are in existence, and we shall
look at one of t hem, so t hat its nat ur e and characteristics can be appre-
ciated. Ot her plants will differ in t hei r const ruct i on or technology, but
from an operat i onal poi nt of view t hei r f undament al characteristics will
probabl y be qui t e similar to those described below.
3.3.4.1 The bubblingfluidised-bed boiler
Fi gure 3.14 shows t he principles of a wast e-t o-energy pl ant based on t he
use of a bubbl i ng fluidised-bed boiler. First, t he i ncomi ng waste is sorted to
r emove oversized, bul ky or dangerous mat eri al . The r emai nder is t hen
carri ed by a system of conveyors to a ha mme r mill where it is broken down
until only manageabl e fragment s remai n. Aft er a separat or has r emoved
incombustible magnet i c items, t he waste is hel d in a storage building, from
wher e it is r emoved as requi red by a screw conveyor and t ransferred via
anot her conveyor to t he boiler. I mmedi at el y before ent eri ng t he boiler,
nonferrous met al s are r emoved by a separator.
The boiler itself comprises a vol ume of sand whi ch is kept in a fluidised
state by jets of air. A port i on of dol omi t e is added to t he sand to assist in
t he reduct i on of corrosion and to reduce any t endency of t he sand and fuel
to coalesce (a process known as ' slagging' ). Aft er t he sand/ dol omi t e
mi xt ur e has been heat ed by a system of st art -up burners, combustible
!

~
J

!

d
m

|

@

waste mat eri al added to it ignites. The heat released is used to gener at e
st eam in a way t hat is similar to convent i onal boilers such as those
described in Chapt er 2.
3.4 Igniter s ys t e ms
What ever t he mai n fuel of t he boiler may be, it is necessary to provide
some means of igniting it. A vari et y of igniters are used, but most moder n
systems compri se a means of generat i ng a hi gh-energy electric spark whi ch
lights a gas or light-oil supply whi ch in t ur n lights t he mai n fuel.
I n addi t i on to igniting t he fuel, t he igniter may sometimes be used to
ensure t hat t he fuel remai ns alight under conditions where it may
otherwise be extinguished. Thi s is referred to as provi di ng ' support ' for t he
mai n burner.
Like many aspects of power-station bur ner operations, the requi rement s
for igniters are defined in st andards such as those devel oped by t he
Nat i onal Fire Prot ect i on Association (e.g. NFPA 8502:95 [1]). In these
st andards, igniters are di vi ded into t hree categories each of whi ch is
defined in detail. I n essence, t he t hree classes have t he following charact er-
istics.
C l a s s 1: An igniter providing sufficient energy to raise the t emperat ure of
t he fuel and air mi xt ur e above the mi ni mum ignition t emperat ure, and to
support combustion, under any burner light-off or operating conditions. Such
igniters generally have a capacity of more t han 10% of the full-load capacity of
the mai n bur ner that it is igniting. This class is also referred to as a' Cont i nuous
Igniter'.
C l a s s 2: ( a l s o ref erred to a s a n ' i n t e r mi t t e n t i g n i t e r ' ) : Capable of lighting the
fuel only under a def i ned range o f l i g h t - o f f conditions. Such igniters have a capacity
generally bet ween 4% and 10% of the fiJll-load burner input and may also be
used to support combustion of the fuel at low loads or under a defined range of
adverse operating conditions.
C l a s s 3: Small igniters, generally applied to gas or oil burners. These igniters
are capable of lighting t he fuel only under a def i ned range o f conditions a n d may not be
used f o r support purposes. Two types of Class 3 igniter are defined: i nt errupt ed
igniters (not usually exceedi ng 4% of t he mai n bur ner fuel i nput energy),
whose operat i on is automatically stopped when a set time has expired after
t he first ignition; and di rect electric igniters which have enough energy to
ignite t he mai n fuel.
The t ype of igniter in use will define the met hods of operat i on of the
bur ner and the sequences t hat are to be empl oyed in the associated burner-
management system.
3.5 Burner-management systems
Safe operat i on of the bur ner and its associated igniter must be ensured
and in most cases this requires the use of a sophisticated bur ner - manage-
ment system (BMS). In outline, these systems include a means of
moni t ori ng the presence of the flame and a reliable met hod and pr ocedur e
for operat i ng the associated fuel valves in a sequence t hat provides safe
ignition at st art -up and safe shut -down, ei t her in the event of a fault or in
response to an oper at or command.
The pr ocedur e for lighting a bur ner depends, first, on checking t hat it
is safe to light it at all. This means that, if no ot her bur ner is firing, confir-
mat i on has been received t hat any fl ammabl e mixtures have been
exhaust ed from the furnace by means of a purge. Such a purge involves the
operat i on of FD and I D fans for a defined time, so t hat a cert ai n vol ume of
air has passed t hrough the furnace. (In a coal-fired boiler the flow rat e
t hr ough the furnace must be at least 40% of the full-load vol umet ri c air
f l o w . )
Once confi rmat i on has been received t hat the furnace purge is
compl et e (or i f ot her burners are al r eady firing), ignition of the bur ner will
depend on the successful operat i on of some form of igniter or pilot and,
once the mai n bur ner has been successfully lit, its operat i on must be con-
tinuously moni t ored, because an extinguished flame may mean t hat
unbur ned fuel is bei ng injected into the combust i on chamber. I f such fuel is
subsequent l y ignited it may explode.
Once a bur ner has ignited, the BMS must ensure t hat safe operat i on
continues, and i f any hazard arises the system must shut of f the burner,
and if necessary, trip the entire boiler.
On shut -down of a burner, steps must he taken to ensure t hat any
unbur ned fuel is cl eared from the pi pework. Thi s pr ocedur e is known as
scavenging, and in an oil bur ner it may involve bl owi ng compressed air or
st eam t hr ough the pi pework and bur ner passages. Such procedures are
defined in codes such as NFPA 8502-95.
Each component of the BMS is vital to the safety of the pl ant and to
the reliability of its operat i on, but the most onerous responsibility rests
with the flame detector: an electronic devi ce which is requi red to oper at e
in close proxi mi t y to hi gh-energy spark ignition systems, and in conditions
of ext reme heat and dirt. Moreover, it must provi de a reliable indication of
46 Power-pl ant control and instrumentation
t he pr e s e nc e or abs ence o f a p a r t i c u l a r f l ame i n t he pr es ence of ma n y
ot her s a n d i t mu s t d i s c r i mi n a t e b e t we e n t he e ne r gy o f t he f l ame a n d hi gh
levels o f r a d i a n t e ne r gy f r o m hot r e f r a c t or y ma t e r i a l s a n d pi pes. T h e
s i ght i ng of t he f l ame ma y al so be af f ect ed by c ha nge s i n f l ame p a t t e r n over
a wi de r a n g e o f o p e r a t i n g condi t i ons , a n d i t ma y al so be obs c ur e d by
s wi r l i ng s moke, s t e a m or dus t .
Safe o p e r a t i o n o f t he boi l er d e p e n d s on p r o p e r des i gn of t he BMS,
i n c l u d i n g t he f l ame s canner , a n d on car ef ul si t i ng o f t he s c a n n e r so t h a t i t
pr ovi de s r el i abl e a n d u n a mb i g u o u s de t e c t i on of t he r e l e va nt f l ame u n d e r
al l o p e r a t i o n a l condi t i ons . Af t er i ns t al l at i on, t he s ys t em c a n be e xpe c t e d t o
p e r f o r m safel y a n d r el i abl y only i f c o n s t a n t a n d me t i c ul ous a t t e n t i o n is pa i d
t o ma i n t e n a n c e . Th i s i mp o r t a n t ma t t e r is all t oo of t en i gnor e d, a n d t he
i nevi t abl e r esul t is t h a t t he s ys t em ma l f unc t i ons , l e a d i n g t o f ai l ur e t o i gni t e
t he fuel, wh i c h ma y i n t u r n de l a y s t a r t - up of t he boi l er. I n t he e xt r e me ,
ma l f unc t i ons c o u l d e ve n e n d a n g e r t he saf et y o f t he p l a n t i f t h e y r esul t i n
fuel bei ng a d mi t t e d t o t he c o mb u s t i o n c h a mb e r wi t h o u t be i ng p r o p e r l y
i gni t ed. A p r o p e r l y de s i gne d BMS will n o t al l ow t hi s t o h a p p e n , b u t i f
r e p e a t e d ma l f u n c t i o n s o c c u r i t is not u n k n o wn f or ope r a t or s t o i gnor e t he
wa r n i n g si gns a n d even t o ove r r i de saf et y syst ems. I n s uch cases i t is us ual
t o b l a me t he BMS a n d / o r t he f l ame moni t or s , whi c h c oul d be f ul l y f unc-
t i onal i f t h e y wer e n o t mi s us e d or ba dl y ma i n t a i n e d . Thi s i mp o r t a n t
subj ect is di s cus s ed i n gr e a t e r d e p t h i n Ch a p t e r 5.
3.6 Gas turbi nes in combi ned-cycl e appl i cati ons
I n t he c ombi ne d- c yc l e pl a nt , t he he a t us ed f or boi l i ng t he wa t e r a n d super -
h e a t i n g t he s t e a m is o b t a i n e d f r om t he e xha us t o f a gas t ur bi ne , as
des cr i bed i n Ch a p t e r 2. I n s uch pl ant , unl ess s u p p l e me n t a r y f i r i ng is used,
t he c o mb u s t i o n pr ocess occur s ent i r el y i n t he gas t ur bi ne . Wh e r e s uppl e-
me n t a r y f i r i ng is used t he r e l e va nt c ont r ol syst ems t ake on ma n y of t he
char act er i s t i cs o f t he oil- or gas- f i r i ng syst ems di scussed ear l i er i n t he
pr e s e nt c ha pt e r .
3.7 Summary
So far, we ha ve l ooked at t he o p e r a t i o n of t he boi l er a n d s t udi e d i n
out l i ne t he boi l er ' s s t e a m, wa t e r a n d gas ci r cui t s, a n d all t he ma j o r i t ems o f
p l a n t r e q u i r e d f or t hei r o p e r a t i o n . Wi t h t hi s u n d e r s t a n d i n g we c a n n o w
l ook at t he c ont r ol a n d i n s t r u me n t a t i o n syst ems as s oci at ed wi t h t he pl ant .
Th i s s ur vey wi l l be s t r u c t u r e d i n mu c h t he s a me wa y as t he p r e c e d i n g
c ha pt e r s , s t a r t i n g wi t h a n ove r vi e w o f a n i mp o r t a n t f u n d a me n t a l : t he
met hod by whi ch t he de ma nd for steam, heat or electrical power is
obt ai ned. Aft erwards, we shall see how this demand is t r ansmi t t ed to all
t he r el evant sections of t he pl ant so t hat t he requi rement s are properl y and
safely addressed.
1 NFPA 8502-95: Standard for the prevention of furnace explosions/
implosions in multiple burner boilers. National Fire Protection Association,
Batterymarch Park, Quincy, MA, USA, 1995
Chapter 4
S e t t i n g t h e d e ma n d f o r t h e s t e a m
g e n e r a t o r
4. 1 N a t u r e o f t h e d e ma n d
The st eam generat ed by the boi l er may be used to dri ve a t urbi ne in a
t her mal power-pl ant , or it may be delivered to an industrial process or a
di st ri ct -heat i ng scheme (or it may be provi ded for a mi xt ure of these uses).
Alternatively, the pr i mar y pur pose of the pl ant may be to i nci nerat e indus-
trial, domest i c or clinical waste, wi t h st eam bei ng generat ed as a val uabl e
by- pr oduct , to dri ve a t ur bo- gener at or or to meet a heat i ng demand. I n
each case, the fact or t hat pri mari l y det ermi nes the oper at i on of t he pl ant is
the amount of st eam t hat is requi red. Everyt hi ng else is subsi di ary to this,
al t hough it may be closely linked to it.
The det er mi nant t hat controls all the boiler' s operat i ons is called the
' mast er demand' . In t her mal power - pl ant the st eam is generat ed by
burni ng fuel, and t he mast er demand sets t he burners firing at a rat e t hat is
commensur at e wi t h the st eam product i on. Thi s in t ur n requires the FD
fans to del i ver adequat e air for the combust i on of the fuel. The air i nput
requi res the product s of combust i on to be expelled from the combust i on
chamber by the I D fans, whose t hr oughput must be rel at ed to t he st eam
flow. At the same time, wat er must be fed into the boiler to mat ch t he pro-
duct i on of steam.
As st at ed previously, a boi l er is a compl ex, mul t i vari abl e, i nt eract i ve
process. Each of the above par amet er s affects and is affect ed by all of the
others.
The way in whi ch the mast er demand operates is det er mi ned bot h by
t he general nat ur e of t he pl ant (is it a power station, an i nci nerat or or a
provi der of process steam?), and also by t he way in whi ch t he boiler is con-
figured wi t hi n t he cont ext of t he overall pl ant (is t here only one boiler
meet i ng t he demand, or are several combined?). The nat ur e of t he mast er
demand system depends on the t ype of pl ant wi t hi n whi ch t he boiler
operates, and it is t herefore necessary to exami ne it separat el y for each
t ype of application. I n t he following sections we shall deal wi t h t he mast er
demand as used in t he following classes of plant:
power stations;
combi ned heat and power (CHP) plants;
Wast e-t o-energy ( WTE) plants.
We shall see t hat al t hough all of these requi re t he boiler to be operat ed to
generat e steam, each has its own requi rement s and constraints.
4.2 Setting the demand in power-stati on applications
A boiler pr oduci ng st eam for an operat i ng t ur bo- gener at or has to ensure
t hat the machi ne cont i nual l y delivers t he requi red electrical ener gy to t he
load. Wi t h a combi ned-cycl e gas-t urbi ne pl ant it is frequent l y t he case t hat
t he power gener at ed by t he gas turbines is adjusted to meet t he demand,
with t he st eam t urbi ne maki ng use of all of t he waste heat from t he
turbines.
Wi t h all types of power-generat i ng plant, however, t he r equi r ement for
generat i on will be set, di rect l y or indirectly, by t he gri d-cont rol cent r e (or
t he ' cent ral di spat cher' ), and t he amount of power t hat is gener at ed will be
rel at ed to t he local or nat i onal demand at t hat time.
I n nat i onal networks, power stations are linked t oget her to generat e
electrical power in concert with one anot her. Toget her t hey must meet a
demand t hat is made up of t he combi ned needs of all t he users t hat are
connect ed to t he system (domestic, commerci al , agri cul t ural , industrial
etc.). The overall demand will var y from mi nut e to mi nut e and day to day
in a way t hat is systematic or r andom, di ct at ed by economi c, operat i onal
and envi r onment al factors. This pat t er n of use relates to t he entire
network, and t he fact t hat a l arge number of power generat ors and users
ar e linked via t he net work has little beari ng on t he overall demand,
al t hough t he ext r eme peaks and t roughs may well be smoot hed out. The
i nt erl i nki ng does, however, have operat i onal implications. For exampl e, a
sudden failure of one generat i ng pl ant will instantly t hrow an ext ra
demand on t he others.
Setting the demand f or the steam generator 51
I n a cold or t emper at e cl i mat e t he demand will be based pr edomi nat el y
on t he need for light, heat and mot i ve power. I n war mer climates and
devel oped areas it will also be det er mi ned by t he use of ai r-condi t i oni ng
and, possibly, desal i nat i on pl ant (for dri nki ng-wat er product i on).
Fi gure 4.1 shows how t he total electrical demand on the Uni t ed
Ki ngdom' s Gri d system varies from hour to hour t hr ough t he day, and
from a war m s ummer day to a cold wi nt er day. Clearly, in addi t i on to
bei ng affected by nor mal worki ng pat t erns, t he demand is det er mi ned by
the level of dayl i ght and t he ambi ent t emper at ur e, bot h of whi ch follow
basic systematic pat t erns but whi ch may also fluctuate in a very sudden
and unpredi ct abl e manner . Similar profiles can be devel oped for each
count r y and will be det er mi ned by cl i mat e as well as t he count ry' s indus-
trial and commer ci al infrastructure.
These days, t he demand for electricity in a devel oped nat i on is also
affected qui t e dr amat i cal l y by television broadcasts. Dur i ng a maj or
sporting event such as an i nt ernat i onal football mat ch, sudden upsurges in
demand will occur at hal f-t i me and full time, when viewers switch on t hei r
kettles. I n t he UK this can impose a sudden rise in demand of as much as
2 GW, whi ch is t he equi val ent to t he total out put of a reasonabl y l arge
5 0 0 0 0
,
2 0 0 0 0
1 0 0 0 0
! I I I
I i i {
i a i i
I I I I
. ! i I
i
i W i n t e r d a y
I
. " . , ~ ~ . ! I l l I i
. . - . ' - p ~ . ~ . . i l l i i I ~ 1 d e r n a n d )
i " , , ~ J I j i ! i T S = ,
~ - - ! j ! ! ! i w i n t e r d a y
~ / ? - " i " ! - ~ J i
I - - I I I - I I
i / i i i ~ i i
L i ~ . . , , / i J ' i : - =
I I i ! ( m i n i m u m
I ~ a ~ , , f ! ! I i ! d e m a n d )
i ! ! i
I i i i i I ,
I 0 3 . 0 0 0 6 . 0 0 0 9 . 0 0 1 2 . 0 0 1 5 . 0 0 1 8 . 0 0 2 1 . 0 0 2 4 . 0 0
! T y p i c a l
! s u m m e r d a y
N S u m m e r d a y
l i m e
Figure 4.1 Typical electrical demand in the United Kingdom
power station. Such a pat t er n of usage can be predi ct ed to wi t hi n a few
minutes, and audi ence predictions are rout i nel y fed to t he power-genera-
tion aut hori t i es on a daily basis to assist with t he provision of adequat e
supplies. But if t he result of t he mat ch requires ' ext ra time' pl ayi ng t here
will be two f ur t her peaks before t he pat t er n of consumpt i on ret urns to
nor mal an hour or so aft er t he end of t he mat ch. Thi s t ype of demand is
obviously not predictable.
The Gri d system has to be managed so t hat t he demand for electricity
is met wi t hi n st at ut ory limits at all times and under all conditions, and the
available gener at i ng pl ant has to be used in the most economi c manner .
Since t he privatisation of t he electricity supply i ndust ry in t he UK, t he
generat i on of electricity is based on t he demands of a t r adi ng system
known as t he Pool. Thi s is briefly described below, because t he operat i on of
t he Pool det ermi nes how each unit receives demand instructions. The
subject is of critical i mpor t ance because it governs t he operat i on of t he
power pl ant and, ul t i mat el y, its demise. Al t hough the following out l i ne is
based on t he UK Pool, ot her countries use systems based on similar princi-
ples.
4.2.1 Operation of the UK Pool
At 10 a. m. each morni ng, British generat i ng compani es who wish to
t r ade dur i ng t he following day submit bids on behal f of each of t hei r
plants, accompani ed by i nformat i on on t he capaci t y available from each
pl ant and its operat i onal paramet ers. These bids are t hen ranked nation-
ally in a form of a l eague table, with t he cheapest gener at or at t he top and
the most expensive at t he bot t om. This table is t er med t he ' meri t order' for
all t he gener at i ng units t hat are capabl e of being connect ed to t he system.
The details of each day' s meri t order are t ransmi t t ed to t he body respon-
sible for oper at i ng t he Gri d system, the Nat i onal Gri d Cont rol Cent re, and
to t he body responsible for t he t r adi ng systems.
The Nat i onal Gri d Cont rol Cent re det ermi nes a not i onal schedule ol
the generat i ng pl ant t hat is available and, on t he basis of this i nformat i on,
develops a system mar gi nal price (SMP) for every 30-mi nut e period of the
day in question. The SMP is combi ned wi t h a component whi ch reflects
the scarcity values of generat i ng plant, and from these factors is deter-
mi ned t he selling price for power, the ' pool selling price' . This takes intc
account t he operat i onal costs of the system, known as t he ' uplift costs'. The
uplift includes factors such as t he cost of mai nt ai ni ng a security mar gi n oJ
available power above t he demand, t he cost of anci l l ary pl ant t hat L,
requi red to mai nt ai n voltage and frequency, transmission constraints etc.
Setting the demand for the steam generator 53
The Nat i onal Gri d Cont rol Cent re exami nes the security of the trans-
mission net work t hat links all the part i ci pat i ng generat i ng pl ant and
consumers, and t hen plans its operations to ensure t hat the ent i re system
operates securely and efficiently.
At t he end of this process, t he Cent re issues instructions to all power
stations connect ed to t he network, setting the generat i on demand for t hei r
units on a mi nut e- by- mi nut e basis. It is these commands whi ch det er mi ne
the earni ngs of the plant, and wi t hi n any given pl ant t he boiler and
t urbi ne must respond to t hem in the most efficient and reliable manner
possible. The load al l ocat ed to a unit will be based both on the cost of gen-
erat i on and on t he ability of t he pl ant to respond to demand changes.
Under t he right conditions, a unit whose operat i onal costs are high but
whi ch responds quickly will be as likely to receive a load as one whi ch
generat es very cheapl y but is slow to respond to changes in demand.
4. 3 T h e ma s t e r d e ma n d i n a p o w e r - s t a t i o n
a p p l i c a t i o n
The response of a boi l er / t ur bi ne unit in a power station is det er mi ned by
t he dynami c characteristics of the two maj or items of plant. These differ
quite significantly from each other. The turbine, in very general terms, is
capable of respondi ng mor e quickly t han the boiler to changes in demand.
The response of t he boiler is det er mi ned by the t her mal inertia of its st eam
and wat er circuits and by t he characteristics of t he fuel system. For
exampl e, a coal -burni ng boiler, with its compl ex fuel -handl i ng plant, will
be much slower to respond to changes in demand t han a gas-fired one.
Also, t he t ur ndown of the pl ant (the range of st eam flows over whi ch it
will be capabl e of operat i ng under aut omat i c control) will depend on the
t ype of fuel being burned, with gas-fired units being i nherent l y capable of
operat i ng over a wi der dynami c range t han t hei r coal-fired equivalents.
The common factor in all these systems, however, is the mast er
demand which, in addi t i on to setting the firing rate, regulates the quant i t y
of combust i on air to mat ch t he fuel input, and t he quant i t y of feed wat er to
mat ch t he st eam product i on. I n t he present chapt er we shall exami ne t he
mast er system. Chapt ers 3, 5, 6 and 7 look at how t he commands from t he
mast er system are act ed upon by t he fuel, dr aught , feed-wat er and steam-
t emper at ur e systems.
The design of t he mast er system is det er mi ned by t he role whi ch t he
pl ant is expect ed to play, and here t hree options are available. The
demand signal can be fed pr i mar i l y to t he t urbi ne (boiler-following
control); or to the boiler (turbine-following control); or it can be di rect ed
to bot h (co-ordi nat ed uni t control). Each of these results in a di fferent
per f or mance of t he unit, in a manner t hat will now be analysed.
4. 3. 1 Boi l er- f ol l owi ng operation
Wi t h boiler-following control, t he power - demand signal modul at es t he
t urbi ne throttle-valves to meet t he load, while t he boiler systems ar e
modul at ed to keep t he st eam pressure constant. The principles of this
system ar e i l l ust rat ed in simplified form in Fi gure 4.2.
I n such a system, t he pl ant operat es with t he t urbi ne throttle-valves
par t l y closed. The act i on of openi ng or closing these valves provides t he
desired response to de ma nd changes. Sudden l oad increases ar e met by
openi ng t he valves to release some of t he stored ener gy wi t hi n t he boiler.
When t he de ma nd falls, closing t he valves increases t he stored ener gy in
t he boiler.
I n such a system t he t urbi ne is t he first to respond to t he changes. The
boiler cont rol system reacts aft er these changes have been made, increasing
or reduci ng t he firing to restore t he st eam pressure to t he set value.
Power/frequency
d e m a n d I
. . . . J Generator
Figure 4.2 'Boiler foUowing' system
4.3.2 Turbine-following operation
I n t he t ur bi ne - f ol l owi ng s ys t em ( Fi gur e 4.3), t he d e ma n d is fed di r e c t l y
t o t he boi l er a n d t he t u r b i n e t hr ot t l e- val ves ar e l eft t o ma i n t a i n a c o n s t a n t
s t e a m pr essur e. Pa r t i c ul a r l y i n t he case o f coal - f i r ed pl a nt , t hi s me t h o d o f
o p e r a t i o n offers sl ower r esponse, becaus e t he t u r b i n e o u t p u t is a dj us t e d
onl y af t er t he boi l er has r e a c t e d t o t he c h a n g e d d e ma n d . Howe ve r , t he
t ur bi ne - f ol l owi ng s ys t em enabl es t he uni t t o be o p e r a t e d i n a mo r e effi ci ent
ma n n e r a n d t u n i n g f or o p t i mu m p e r f o r ma n c e is easi er t h a n wi t h t he
boi l er - f oUowi ng syst em. I t is wo r t h c ons i de r i ng for l ar ge bas e- l oad p o we r
p l a n t ( wher e t he u n i t r uns at a fi xed l oad, us ual l y a h i g h one, for mos t of
t he t i me) , or wi t h gas- f i r ed p l a n t whe r e t he r es pons e is c o mp a r a t i v e l y
r api d.
4.3.3 Co-ordinated unit control
Wi t h c o - o r d i n a t e d u n i t l oa d c ont r ol ( Fi gur e 4.4), t he p o we r d e ma n d is
fed t o t he boi l er a n d t u r b i n e i n par al l el , wi t h var i ous c ons t r a i nt s bui l t i nt o
e a c h c h a n n e l t o r ecogni s e a n d al l ow f or a ny d y n a mi c cons t r ai nt s o f t he
r e l e va nt pl a nt . Th i s is a s ophi s t i cat ed t e c hni que , whi c h has c o me i nt o its
o wn wi t h t he d e v e l o p me n t o f power f ul , fast, a n d ver s at i l e c o mp u t e r
syst ems. I t c ombi ne s t he best f eat ur es of b o t h t he boi l er - f ol l owi ng a n d t he
L
6)
Figure 4.3 'Turbine foUowing' system
t I A
i Co-ordinatedunit -
Power, lrequency j i
demand b c o n t r o l l e r
, , L
. w
Turbine
Generator
r
Condense~
Boiler
feed F D
pump ~ fan
Figure 4.4 'Co -ordinated unit control' system
t urbi ne-fol l owi ng systems. However, its design demands considerable
knowl edge of t he characteristics and limitations of t he maj or pl ant items.
Also, commi ssi oni ng of this t ype of system demands great skill and care if
the full ext ent of t he benefits is to be obt ai ned. In part i cul ar, t he rate-of-
change of t he demand signals, as well as t he ext ent of t hei r dynami c range,
will need to be const rai ned to prevent undesirable effects such as the
stressing of pi pework because of excessively steep rat es-of-change of t em-
perat ure.
Because of t he nat ur e of its operat i on, the details of a co-ordi nat ed unit
load cont rol system have to be finely mat ched to t he confi gurat i on and
characteristics of t he pl ant to whi ch it is fitted. Figure 4.4 shows only the
overall principle of t he system: in practice, a wide vari et y of system config-
urat i ons are in use.
4.3.4 Relative performance
Of t he t hree options described above, t he co-ordi nat ed unit load control
system provides t he best possible response to changes in demand.
However, because it is so sophisticated its per f or mance is heavily
Setting the demand for the steam generator 57
dependent on t he accur acy of t he many pieces of i nformat i on on whi ch
its operat i on is based (such as t her mal rat e-of-change limits in t he
turbine). Unless it is regul arl y readjusted, it can suffer from an inability to
recognise and deal with t he st eady det eri orat i on in per f or mance t hat inevi-
t abl y occurs in each i t em of pl ant as it ages.
4.3.5 Comparing the response rates of the systems
As stated above, t he co-ordi nat ed unit load controller, when properl y
designed, commi ssi oned and mai nt ai ned, will provi de the best possible
response of t he unit wi t hi n t he constraints of t he pl ant itself.
Unfort unat el y, for many pract i cal reasons it is not universally used. I n
ol der pl ant this t ype of mast er confi gurat i on may not be available or
practical. I n ot her cases economi c or time constraints may have necessi-
t at ed the use of a simpler r~gime (the co-ordi nat ed unit system is expensive
to design, and its commi ssi oni ng requires much t i me and effort i f t he
maxi mum benefit is to be gai ned from its capabilities).
Wher e t he co-ordi nat ed unit system is not available, the choice lies
bet ween using boiler-following or turbine-following control. Al t hough
t hey both orchest rat e t he operat i on of t he boiler and t he t urbi ne to meet
changes in demand, t he performances of these configurations differ very
consi derabl y from each other, as is expl ai ned below.
4.3.5.1 Response of the boiler-following system
When a change occurs in its steam-flow demand, a boiler has to
overcome its own t her mal i nert i a before it can mat ch t he change.
Therefore, by using t he turbine' s ability to respond mor e rapidly, the
boiler-following system provides a bet t er response to load changes t han the
turbine-following system. Aft er t he t urbi ne has responded to t he change in
demand, t he boiler is commanded to follow on, correct i ng the steam-
pressure error as quickly as it can.
However, such rapi d response is only available for small-scale demand
changes, t hat is, changes t hat are within t he capaci t y of t he allowable
r ange of pressure-drop across the t hrot t l e valve. Also, t he rapi d response is
obt ai ned at a cost. When oper at ed in this way, the efficiency of t he unit is
inevitably reduced because of t he pressure t hat is dr opped across t he
t hrot t l e valves. The losses are reduced by decreasi ng this pressure drop,
but this also reduces t he scope for meet i ng sudden changes in demand.
Anot her probl em is t hat it is not easy to t une t he cont rol par amet er s of
a boiler-following system to obt ai n opt i mum overall performance, mai nl y
because of t he i nt eract i on bet ween st eam pressure and st eam flow t hat
occurs as t he t urbi ne and boi l er respond to changes.
Consi der what happens when a sudden rise in demand occurs. The first
response is for t he t hrot t l e valves to be opened. Thi s increases t he power
gener at ed by t he machi ne, but it also results in t he boiler pressure falling,
and when this happens t he boiler cont rol system reacts by i ncreasi ng t he
firing rate. Thi s is all ri ght as far as it goes since, qui t e correctly, it
increases t he boiler st eami ng rat e to meet t he increase in demand.
However, as t he firing change comes into effect and t he st eam pressure
rises, t he amount of power t hat is being gener at ed also increases. But as it
has al r eady been i ncreased to meet t he d e ma n d - - a n d in fact may have
al r eady done s o - - t h e power gener at ed can overshoot t he target, causing
t he t hrot t l e valves to st art closing again, whi ch raises t he boiler
pr e s s ur e . . . , and so on.
Vari ous met hods have been proposed to ant i ci pat e these effects, but
these t end to i ncrease t he compl exi t y of t he system, and t herefore its cost,
wi t h questionable l ong-t erm benefit.
4.3.5.2 Response of the turbine-following system
I n t he simplest version of t he turbine-following system t he boiler firing
rat e, and t he r at e of ai r and feed-wat er admission etc., are all fixed (or, at
least, held at a set value, whi ch ma y be adj ust ed f r om time to t i me by t he
boiler operat or), and t he t urbi ne t hrot t l e valves are modul at ed to keep t he
st eam pressure const ant . However, when t he fuel, air and wat er flows of a
boiler ar e hel d at a const ant value t he amount of st eam t hat is gener at ed
will not, in general , r emai n constant, mai nl y because of t he inevitable var-
iations t hat will occur in par amet er s such as t he calorific value of t he fuel,
t he t emper at ur e of t he feed wat er etc. I n the simple turbine-following
system, these vari at i ons ar e correct ed by modul at i on of t he t urbi ne
t hrot t l e valve to mai nt ai n a const ant st eam pressure, but this results in var-
iations in t he power gener at ed by t he turbine.
Because t he st eam- gener at i on rat e of its boiler is not aut omat i cal l y
adj ust ed to meet an ext ernal demand, a pl ant operat i ng under t he cont rol
of a simple t urbi ne-fol l owi ng system will gener at e amount s of power t hat
do not rel at e to t he short -t erm needs of t he grid system. Such a pl ant is
t herefore i ncapabl e of oper at i ng in a frequency-support mode, al t hough
this mode of oper at i on ma y be used where it is not easy, or desirable, to
adjust t he fuel i nput , for i nst ance in industrial wast e-i nci nerat i on plants.
Fi gure 4.3 shows t he preferred opt i on for ot her types of plant, wi t h t he
boiler firing rat e (and t he i nput of air and wat er) bei ng set by t he demand
on t he unit. Changes in this demand t herefore change the boiler' s firing
rate, and a cont rol l er t hen modul at es t he t urbi ne t hrot t l e valve to keep t he
st eam pressure const ant . Thi s t echni que closes t he loop ar ound t he power,
but it does so by di rect i ng t he changes in demand to t he boiler first,
rel yi ng on t he resulting changes in pressure to change t he amount of
power t hat is generat ed. As mi ght be expect ed, because of t he slow
react i on t i me of t he boiler, this results in a slower response to load changes
t han t hat of t he boiler-following system.
4. 4 Lo a d d e ma n d i n c o mb i n e d he a t a nd p o we r p l a n t s
Reference has al r eady been made to the use of gas turbines in combi ned-
cycle installations. This is a part i cul ar exampl e of a ' co-generat i on'
scheme: a t er m applied to dual -purpose plants where heat whi ch woul d
otherwise be wast ed from one process is used in anot her. I n t he case of
CCGT plant, t he heat exhaust ed from a gas t urbi ne is used to gener at e
steam. I n CHP plant, heat from a power station is used in anot her process.
The heat ma y be t aken from t he power pl ant as st eam ext r act ed f r om t he
turbine, or it may be t he heat abst ract ed from t he condensate.
Co-generat i on plants are ei t her ' t oppi ng' or ' bot t omi ng' systems. Wi t h
t he former, t he first pri ori t y is to generat e electricity, and as much use as
possible is made of t he heat t hat woul d otherwise be wast ed in t he process.
Wi t h t he latter, waste heat from some industrial process is used to gener at e
electricity vi a a st eam gener at or and turbine.
A st eam gener at or empl oyed in a CHP pl ant has to serve two masters:
t he need for heat, and t he demand for electricity generat i on. I n most cases
t he f or mer predomi nat es, because t he ent i re raison d'etre of t he pl ant was
probabl y t he need to serve a communi t y or an industrial plant, and t he
plant' s ability to gener at e electricity is of secondar y i mpor t ance (even
t hough, as a spin-off, it is ext remel y valuable).
For this reason, t he devel opment of a t rul y effective mast er - demand
signal for a CHP pl ant is much mor e compl ex t han it is wi t h a pl ant whose
only funct i on is to gener at e electricity. The needs of all t he users have to be
t aken into consideration, as must t he cost of t he steam, heat and electricity
t hat is produced. Fur t her mor e, it is possible t hat t he way in whi ch t he
mast er demand is confi gured may need to be modi fi ed at some t i me over
t he life of t he pl ant because of changes in fuel prices or alterations in t he
requi rement s of t he industrial, commer ci al or domest i c complexes whi ch
benefit from t he process.
The wide r ange of possibilities of i nt erconnect i ng t he various systems
in CHP pl ant gives rise to very diverse met hods of organising t he mast er
demand. Confi guri ng a mast er - demand signal t hat takes all t he requi re-
ment s into account ought not to be a significant probl em, beari ng in mi nd
t he power and flexibility t hat is offered by t he moder n DCS, but t he diffi-
cul t y is to obt ai n enough dat a on these requi rement s, and t hen to ensure
t hat t he i nformat i on is correct. Qui t e often it seems that, even if t he
options mi ght have been consi dered at some time, reconciling t he various
requi rement s has proved to be i nt ract abl e and so a cheap and simple com-
promi se has been empl oyed. This may be reasonabl y effective, and t he
pl ant t hat is so devel oped continues to gener at e heat and power for days on
end, t he response to changes in demand seems adequat e, and t he opera-
tional st af f ar e unpr epar ed to al t er anyt hi ng for fear of rocki ng t he boat.
None of this alters t he fact t hat the expendi t ure of a few mor e days (or
even weeks) of effort in front -end definition could have yielded, over the
operat i onal life of t he plant, efficiency and per f or mance i mprovement s
t hat woul d have ampl y recovered t he cost of devel opment .
4.5 Waste-to-energy plants
The design of mast er - demand signals of wast e-t o-energy plants, as
described in Chapt er 3, requires very careful at t ent i on. The requi rement s
in this ar ea are somewhat similar to those rel at i ng to CHP plants, t he
reconci l i at i on of differing operat i onal requi rement s.
The design of the system may favour consumpt i on of the waste
mat eri al , wi t h electricity generat i on t reat ed as a useful and revenue-
ear ni ng by-product , or it ma y t ry to maxi mi se the power-generat i on cap-
ability of t he plant. In bot h cases, however, it is i mpor t ant to recognise t he
special characteristics of t he plant, in part i cul ar, t he boiler response.
Because of t he nat ur e of its compl ex fuel -handl i ng system (see Figure
3.14), a wast e-t o-energy pl ant cannot be expect ed to be very responsive to
demand changes. Ther ef or e it is largely i mpract i cal to consider t he appli-
cat i on of advanced cont rol logic to t he mast er - demand system for this t ype
of pl ant (al t hough various at t empt s have been made to do so). The most
cost-effective solution is to appl y a simple boiler-following system as
described in Section 4.3.1. The reduct i on in efficiency is negligible (and
even somewhat academi c, since not onl y is t he fuel in this t ype of pl ant
easily obt ai ned, but also t he user is pai d for consumi ng it!). Also, t he diffi-
culties of t uni ng the system (due to the i nt eract i on bet ween t he st eam
gener at or and t he st eam user) are less of a probl em in this t ype of installa-
tion, because of the very di fferent dynami c responses of the t urbi ne and
boiler. The difference bet ween the slow response of t he boiler and the
qui ck response of t he t urbi ne also simplifies decoupl i ng one from t he ot her
in t he opt i mi sat i on process.
These factors ease the selection of a mast er system in wast e-t o-energy
plants. A basic boiler-following design provides speed and simplicity of
commissioning, and usual l y performs adequat el y.
4.6 Summary
In this chapt er we have seen how a ' mast er demand' signal is generat ed
in respect to the nat ur e of the duties t hat the pl ant is designed to
undert ake. Thi s signal is responsible for ensuring t hat t he boiler reacts to
changes in demand, and it must also co-ordi nat e the operat i on of each of
the subsi di ary systems. The mai n areas involved in this process are t he
combust i on and dr aught systems, the feed-wat er system and the st eam-
t emper at ur e cont rol system.
In the next chapt er we shall see how the combust i on and dr aught
systems of a boi l er react to the demands of the mast er signal to pr oduce the
requi red firing rat e and how the suppl y of air keeps in step wi t h the
changes to pr oduce t he correct conditions for the combust i on of the fuel.
I n addition, we shall see how the inlet and exhaust fans are regul at ed to
mai nt ai n the correct pressure in the furnace while all this is going on.
Chapter 5
Co mbus t i o n and draught cont rol
Whe n consi deri ng fired boilers and heat - r ecover y st eam generat ors it is
cl ear t hat in t he areas of t hei r st eam and wat er circuits t here are ma ny
similarities bet ween t hem ( al t hough t he HRSG may have t wo or mor e
pressure systems). But when t he systems for cont rol l i ng t he heat i nput ar e
exami ned, t he t wo types of pl ant t ake on al t oget her di fferent charact eri s-
tics. The reason for this is f undament al : wi t hi n t he HRSG, no act ual
combust i on process is involved, since all t he heat i nput is deri ved from t he
gas-t urbi ne exhaust (except wher e suppl ement ar y firing is i nt r oduced
bet ween t he gas t ur bi ne and t he HRSG) . The subject of combust i on
cont rol , whi ch we shall be exami ni ng in this chapt er, is t herefore onl y
rel evant to fired plant.
Nat ural l y, in a fired boiler t he cont rol of combust i on is ext r emel y
critical. I n or der to maxi mi se operat i onal efficiency combust i on must be
accurate, so t hat t he fuel is consumed at a rat e t hat exact l y mat ches t he
demand for st eam, and it must be execut ed safely, so t hat t he ener gy is
rel eased wi t hout risk to plant, personnel or envi ronment . ( The amount of
ener gy involved in a power pl ant is considerable: in each second of its
oper at i on a l arge boiler releases ar ound a billion joules, and in a process of
this scale t he results of an er r or can be catastrophic. )
I n this chapt er we shall see how t he combust i on process is cont rol l ed to
meet t he t wo objectives defined in t he previous par agr aph. We shall also
exami ne t he subsidiary systems t hat mai nt ai n t he correct operat i onal con-
ditions in t he fuel -handl i ng pl ant of coal-fired boilers.
5.1 The pri nci pl es of combus t i on control
In Chapt er 3 we saw t hat the t heoret i cal l y perfect combust i on of a fuel
requires the provision of exact l y the right amount of air needed tbr
compl et e combust i on of the fuel. For the boiler as a whol e this means t hat
the total amount of air bei ng delivered to the combust i on chamber at any
instant mat ches the total amount of fuel ent eri ng t hat chamber at t hat
time. For an individual bur ner it means t hat the fuel and air bei ng
del i vered to the bur ner are always in st ep wi t h one another.
On the surface, therefore, it woul d appear t hat the mat t er of combust i on
cont rol merel y involves keeping the fuel and air i nput s in step with each
other, accordi ng to the demands of the master, and i f this were t rue this
role woul d be adequat el y addressed by a st rai ght forward flow rat i o con-
troller. Unfort unat el y, when the realities of practical pl ant are involved,
the si t uat i on once agai n becomes far more compl ex t han this simple
analysis woul d suggest.
When the relationship bet ween the fuel and air flowing at any instant
into the furnace is chemi cal l y ideal for combust i on, the relationship
bet ween the t wo flows is known as the stoichiometric f uel / ai r ratio.
However , as st at ed earlier, it is usually necessary to oper at e at a f uel / ai r
ratio t hat is different from this t heoret i cal l y opt i mal value, general l y with
a cert ai n amount of excess air. All the same, even t hough mor e t han the
t heoret i cal amount of air has to be provi ded, any overprovi si on of air
reduces the efficiency of the boiler and results in undesirable stack
emissions, and must t herefore be limited.
The reduct i on in efficiency is due to losses which are composed of the
heat wast ed in the exhaust gases and the heat whi ch is t heoret i cal l y
avai l abl e in the fuel, but which is not burned. As the excess-air level
increases, the heat lost in the exhaust gases increases, while the losses in
unbur ned fuel reduce (the short age of oxygen at the lower levels increasing
the degree of i ncompl et e combust i on t hat occurs). The sum of these t wo
losses, plus the heat lost by radi at i on from hot surfaces in the boiler and its
pi pework, is identified as the total loss.
Fi gure 5.1 shows t hat operat i on of the pl ant at the poi nt identified at
'A' will correspond wi t h mi ni mum losses, and from this it may be assumed
t hat this is the poi nt to which the operat i on of the combust i on-cont rol
system shoul d be t arget ed. However , in pract i ce air is not evenl y distrib-
ut ed within the furnace. For exampl e, operat i onal considerations requi re
t hat a suppl y of cooling air is provi ded for idle burners and flame
monitors, to prevent t hem being damaged by heat from near by active
burners and by general radi at i on from the furnace. Air also enters the
Combustion and draught control 65
u .
Total losses
s S s
\ i . "
" \ i . . - " "
\ ..i, "
S S S ~0
s " s " i ~ . Unburned loss
i "
os ~ . . . . . .
k l ~ e e s s 02
A
Figure 5.1 Heat losses in a f urnace
combust i on chamber t hrough leaks, observat i on ports, soot -bl ower ent ry
points and so on. The sum of all this is referred to as ' t r amp air' or ' setting
leakage' . I f this is included in the total bei ng supplied to the furnace, and i f
t hat total is appor t i oned to the total amount of fuel bei ng fired, the impli-
cat i on is t hat some burners (at least) will be depri ved of the air t hey need
for the combust i on of their fuel. In ot her words, the correct amount of air
is being provi ded in total, but it is going to places where it is not avai l abl e
for the combust i on process.
Oper at i on of the firing system must t ake these factors into account ,
and from t hen on the system can appor t i on the fuel and air flows. I f these
are mai nt ai ned in a fixed relationship with each ot her over the full range
of flows, the amount of excess air will be fixed over the entire range.
5.1.1 A simple system: "parallel control"
The easiest way of mai nt ai ni ng a relationship bet ween fuel flow and air
flow is to use a single act uat or to position a fuel-control val ve and an air-
cont rol damper in parallel wi t h each ot her as shown in Figure 5.2. Here,
the openi ng of an air-control damper is mechani cal l y linked to the
openi ng of a fuel cont rol valve to mai nt ai n a defined relationship bet ween
T
I
F u e l f l o w ~"~
Figure 5. 2 Simple 'parallel' control
Master
demand
,t
Fue/va/ve A/r control damper
fuel flow and ai r flow. Thi s system is empl oyed in very small boilers, and
a var i ant allows a non-l i near relationship bet ween valve openi ng and
damper openi ng to be det er mi ned by t he shape of a cam, wi t h a range of
cams offering a vari et y of relationships.
Al t hough this simple system may be qui t e adequat e for very small
boilers bur ni ng fuels such as oil or nat ur al gas, its deficiencies become
i ncreasi ngl y appar ent as t he size of t he pl ant increases.
One l i mi t at i on of t he system is t hat it assumes t hat t he amount of fuel
flowing t hr ough t he valve and t he quant i t y of ai r flowing past t he damper
will r emai n const ant for a given openi ng of t he respective devices. I n
practice, i f a valve or damper is held at a given opening, t he flow past it
will change as t he applied pressure changes. Fur t her mor e, t he flow will
also be affect ed by changes in t he characteristics of t he fuel and air,
not abl y t hei r densities.
Anot her pr obl em is t hat t he response times of t he fuel and air systems
ar e never identical. Therefore, i f a sudden l oad-change occurs and t he two
cont rol l i ng devices ar e moved to pr edet er mi ned openings, t he flows
t hr ough t hem will react at di fferent rates. Wi t h an oil-fired boiler, a
sudden i ncrease in demand will cause t he fuel flow to increase quickly, but
t he air system will be slower to react. As a result, i f t he f uel / ai r rat i o was
correct before t he change occurred, t he firing conditions aft er t he change
will t end to become fuel-rich until t he air system has had t i me to cat ch up.
Thi s causes charact eri st i c puffs of black smoke to be emi t t ed as unbur ned
fuel is ejected to t he chi mney.
On a l oad decrease t he reverse happens, and t he mi xt ur e in t he combus-
tion chamber becomes air-rich. The resulting hi gh oxygen cont ent coul d
lead to corrosion damage to t he met al work of t he boiler, and to
unaccept abl e flue-gas emissions.
5.1.2 Fl ow ratio control
The first appr oach to overcomi ng t he limitations of a simple ' parallel'
system is to measure t he flow of t he fuel and t he air, and to use closed-loop
controllers to keep t hem in t rack wi t h each other, as shown by the two con-
figurations of Figure 5.3.
I n each of these systems t he mast er demand (not shown) is used to set
t he quant i t y of one par amet er being admi t t ed to t he furnace, while a con-
troller mai nt ai ns an adjustable relationship bet ween t he two flows (fuel
and air). Ei t her of t he flows can be selected to be t he one t hat responds
di rect l y to t he mast er and, in Section 5.1.2.1, we shall see t he di fferent
effects t hat result when fuel flow or ai r flow is used in this way.
I n t he system shown in Figure 5.3a a gain block or amplifier in one of
t he flow-signal lines is used to adjust t he ratio bet ween t he two flows. As
the gai n (g) of this block is changed, it alters the slope of the fuel-flow/air-
flow charact eri st i c, changi ng t he amount of excess air t hat is present at
each flow. Note t hat when t he gain is fixed, the amount of excess air is t he
same for all flows, as shown by t he hori zont al line.
I n practice, this situation woul d be impossible to achieve, since some
air i nevi t abl y leaks into t he furnace, with the result t hat t he amount of
excess air is proport i onal l y great er at low flows t han high flows. Thi s
causes t he excess-air line to curve hyperbol i cal l y upwards at low flows
(much as is shown in Figure 5.3b). Practical bur ner requi rement s demand
t hat t he quant i t y of air should always be slightly gr eat er t han t hat whi ch
t he t heoret i cal stoichiometric rat i o woul d dictate. The charact eri st i c
woul d t herefore not pass t hr ough the origin of t he gr aph as is shown in
Figure 5.3a.
Fi gure 5.3b shows a di fferent cont rol ar r angement worki ng wi t h t he
same idealised pl ant (i.e. one with no ai r leaking into the combust i on
chamber ) . Here, instead of a gai n function, a bias is added to one of t he
signals. The effect of this is t hat a fixed surfeit of air is always present and
this is proport i onal l y l arger at t he smaller flows, with t he result t hat t he
amount of excess ai r is largest at small flows, as shown. Changi ng t he bias
signal (b) moves t he curve bodily as shown.
Each of these cont rol configurations has been used in pract i cal plant,
al t hough t he version wi t h bias (Figure 5.3b) exacerbat es t he effects of
t r amp ai r and t herefore tends to be confined to smaller boilers. The
ar r angement shown in Figure 5.3a t herefore forms t he basis of most
pract i cal f uel / ai r ratio cont rol systems.
T
. , , " / " Excess ai r
. . . . . . . . ~ . _ , - - . " . . . . . . . . . . . . . . . ~ . . . . . . . . . . . . . . .
Fuel fl ow
Fuel fl ow Ai r flow
I I
Ai r control damper
( a )
T
. . ~ " '..x~ Excess air
Fuel fl ow
+
II
Ai r control damper
Figure 5.3
( b )
Fuel~air ratio control
a Gai n adj ust ment of f uel / ai r rat i o
b Bias adj ust ment of f uel / ai r rat i o
In these illustrations it has been assumed t hat the mast er demand is fed
to the fuel valve, leaving the air-flow cont rol l er to mai nt ai n the f uel / ai r
rat i o at the correct desired value. When this is done, the confi gurat i on is
known as a ' fuel lead' system since, when the l oad demand changes, the
fuel flow is adj ust ed first and the cont rol l er t hen adjusts the air flow to
mat ch the fuel flow, after the l at t er has changed.
It doesn' t have to be done this way. Instead, the mast er demand can be
rel ayed to the air-flow controller, whi ch means t hat the task of mai nt ai ni ng
the f uel / ai r ratio is t hen assigned to the fuel controller. For obvi ous reasons
this is known as an ' air-lead' system.
5.1.2.1 Comparing the "fuel-lead' and 'air-lead' approaches
Of the t wo alternatives descri bed above, the fuel-lead version will provi de
bet t er response to l oad changes, since its act i on does not depend on the
sl ower-respondi ng pl ant t hat supplies combust i on air to the furnace.
However, because of this, the system suffers from a t endency to pr oduce
fuel-rich conditions on l oad increases and fuel-lean conditions on decreases
in the load. Oper at i ng in the fuel-rich region raises the risk of unbur ned
fuel bei ng ignited in an uncont rol l ed manner, possibly causing a furnace
explosion. Wher eas operat i ng with t oo much excess air, while not raising
t he risk of an uncont rol l ed fire or an explosion, does cause a vari et y of
ot her probl ems, including back-end corrosi on of the boi l er structure, and
undesi rabl e stack emissions.
The air-lead system is slow to respond because it requires the dr aught
pl ant to react before the fuel is increased. Al t hough this avoids the risk of
creat i ng fuel-rich condi t i ons as the l oad increases, it remai ns pr one to such
a risk as the l oad decreases. However , the hazar d is less t han for the fuel-
l ead system.
A furt her limitation of these systems (in either the fuel-lead or air-lead
version) is t hat t hey offer no prot ect i on against equi pment failures, since
these cannot be det ect ed and correct ed wi t hout special precaut i ons bei ng
taken. For exampl e, in the fuel-lead version, if the fuel-flow t ransmi t t er
fails in such a way t hat it signals a l ower flow t han the amount t hat is
act ual l y bei ng del i vered to the furnace, the f uel / ai r rat i o cont rol l er will
at t empt to reduce the suppl y of combust i on air to mat ch the erroneous
measurement . Thi s will cause the combust i on conditions to become fuel-
rich, wi t h t he at t endant risk of an explosion. Conversely, i f the fuel-flow
t ransmi t t er in the air-lead system fails low, the fuel cont rol l er will at t empt
to compensat e for the appar ent loss of fuel by injecting mor e fuel into t he
furnace, wi t h similar risks.
Th e s e ar e j u s t s ome o f t he f ai l ur e char act er i s t i cs whi c h t he basi c s ys t em
de s i gn c a n n o t addr ess. Al t h o u g h t he sel f - di agnost i c f eat ur es i n c o r p o r a t e d
i n mo d e r n t r a n s mi t t e r s c a n be a r r a n g e d t o r ai se a n a l a r m a n d t r i p t he
bur ne r s , or o p e r a t e t he p l a n t i n a p r o t e c t e d mo d e , unt i l t he f aul t has be e n
c or r e c t e d, i t woul d be pr e f e r a bl e t o e mp l o y a s ys t em whi c h has g r e a t e r
i n h e r e n t abi l i t i es t o deal wi t h f ai l ur es b o t h i n t he p l a n t a n d i n its c ont r ol
a n d i n s t r u me n t a t i o n e q u i p me n t .
T h e so- cal l ed ' cr oss- l i mi t ed' c o mb u s t i o n c ont r ol s ys t em addr esses t hese
fact ors i n a ver y c o mp r e h e n s i v e way, as des cr i bed i n t he f ol l owi ng sect i on.
5. 1. 3 Cross-limited control
Fi gur e 5. 4 shows t he pr i nci pl es o f t he cr os s - l i mi t ed c o mb u s t i o n c ont r ol
syst em. I n d i v i d u a l f l ow- r at i o cont r ol l er s (7, 8) ar e p r o v i d e d f or t he fuel
a n d ai r syst ems, r espect i vel y. I g n o r i n g f or t he mo me n t t he sel ect or uni t s
(5, 6) a n d t he f u e l / a i r r at i o a d j u s t me n t bl ock (4), i t wi l l be seen t h a t t he
ma s t e r d e ma n d si gnal is fed t o e a c h o f t hese cont r ol l er s as t he desi r ed-
va l ue si gnal , so t h a t t he del i ver y o f fuel a n d ai r t o t he f ur na c e c o n t i n u a l l y
ma t c h e s t he l oad. Because fuel fl ow a n d ai r fl ow ar e e a c h me a s u r e d as p a r t
o f a cl osed l oop, t he s ys t em c o mp e n s a t e s f or a n y c ha nge s i n e i t he r of t hese
flows t h a t ma y be c a us e d by e xt e r na l fact ors. For t hi s r e a s on i t is
s ome t i me s r ef er r ed t o as a ' f ul l y me t e r e d ' syst em. T h e effect o f t he f ue l ] a i r
r at i o a d j u s t me n t bl ock (4) is t o mo d i f y t he ai r - f l ow si gnal i n a c c o r d a n c e
wi t h t he r e q u i r e d f u e l / a i r r el at i ons hi p.
So far, t he c o n f i g u r a t i o n pe r f or ms s i mi l ar l y t o t he basi c syst ems s h o wn
i n Fi gur e 5.3. T h e di f f er ence be c ome s a p p a r e n t wh e n t he ma x i mu m a n d
mi n i mu m sel ect or s ar e b r o u g h t i nt o t he pi ct ur e. Re me mb e r i n g t he
p r o b l e ms o f t he di f f er i ng r es pons e- r at es o f t he fuel a n d ai r s uppl y syst ems,
cons i der wh a t h a p p e n s wh e n t he ma s t e r d e ma n d si gnal s u d d e n l y r equest s
a n i ncr eas e i n fi ri ng. As s ume t ha t , p r i o r t o t h a t i ns t a nt , t he fuel a n d ai r
cont r ol l er s ha ve be e n k e e p i n g t he i r r es pect i ve c o n t r o l l e d va r i a bl e i n st ep
wi t h t he d e ma n d , so t h a t t he fuel -fl ow a n d modi f i e d ai r - f l ow si gnal s ar e
e a c h e qua l t o t he d e ma n d si gnal . Wh e n t he ma s t e r d e ma n d si gnal
s u d d e n l y i ncr eases, i t n o w be c ome s l a r ge r t h a n t he fuel -fl ow s i gnal a n d it is
t he r e f or e i g n o r e d by t he mi n i mu m- s e l e c t o r bl ock whi c h i ns t e a d l at ches
o n t o t he mo d i f i e d ai r - f l ow s i gnal ( f r om i t e m 4). T h e fuel c ont r ol l e r n o w
assumes t h e r ol e of f u e l / a i r r at i o cont r ol l er , ma i n t a i n i n g t he boi l er ' s fuel
i n p u t at a va l ue t h a t is cons i s t ent wi t h t he ai r b e i n g del i ver ed t o t he
f ur nace. T h e ai r fl ow is me a n wh i l e be i ng i ncr eas ed t o me e t t he n e w
d e ma n d , si nce t he ma x i mu m- s e l e c t o r bl ock (6) has l a t c he d o n t o t he r i si ng
ma s t e r si gnal .
Fuel flow

2
Master
demand
A/r #ow
,
9 ~ 10
Air control damper
Fuel valve
Figure 5.4 Basic cross-limited control system
On a decrease in load, t he system operat es in t he reverse manner . The
mi ni mum-sel ect or block locks ont o t he collapsing mast er and quickly
reduces t he fuel flow, while t he maxi mum- sel ect or block chooses t he fuel-
flow signal as t he demand for the air-flow cont rol l er (8), whi ch t herefore
starts to operat e as t he f uel / ai r rat i o controller, keeping t he air flow in step
with t he fuel flow.
Analysis of t he system will show t hat it is much bet t er able to deal with
pl ant or C& I equi pment failures. For exampl e, i f t he fuel valve fails open,
t he air cont rol l er will mai nt ai n adequat e combust i on ai r to meet t he
quant i t y of fuel bei ng supplied to t he combust i on chamber. This may
result in overfi ri ng but it cannot cause fuel-rich conditions to be cr eat ed
in t he furnace. Similarly, i f t he fuel-flow t r ansmi t t er fails low, al t hough the
fuel cont rol l er will still at t empt to compensat e for t he appar ent loss of
fuel, t he ai r flow cont rol l er will ensure t hat adequat e combust i on ai r is
supplied.
The system cannot compensat e for all possible failures, but it provides
a much hi gher level of prot ect i on t han any of t he simpler systems
described earlier, and when coupl ed with self-checking diagnostics and
pr oper faul t -det ect i on techniques it provides a high degree ofsat~ty.
5.1.3.1 Using gas analysis to vary the fuel~air ratio
I n t he systems shown in Figures 5.3 and 5.4, t he relationship bet ween the
fuel and air quant i t i es is manual l y adjusted, ei t her t he gai n or the bias is
al t ered to change t he combust i on conditions. Wi t h such systems, if t he
adj ust ment fact or is set wrongl y, or if changes outside t he system di ct at e
t hat the f uel / ai r rat i o should be altered, no provision exists for aut omat i c
correct i on, and t he ri ght combust i on conditions can only be rest ored by
manual i nt ervent i on. To i mprove per f or mance and safety, some form of
aut omat i c recogni t i on and correct i on of these factors woul d be prefer-
able.
I f t he f uel / ai r rat i o is incorrect, combust i on of t he fuel will be affected
and t he results will be observable in t he flue gases. Thi s indicates t hat an
effective way of opt i mi si ng t he combust i on process is to change t he f uel / ai r
rat i o aut omat i cal l y in response to measur ement s of t he flue-gas cont ent .
For all fossil-fuelled boilers, t he oxygen cont ent of t he flue gases
increases as t he excess-air quant i t y is increased, while t he car bon dioxide
and wat er cont ent decreases. The car bon monoxi de cont ent of t he boiler' s
flue gases is a di rect i ndi cat i on of t he completeness of the combust i on
process and systems based on the measur ement of this par amet er have
long been recognised as an effective mechani sm for i mprovi ng combust i on
per f or mance in coal and oil-fired boiler pl ant [1]. However, experi ence
indicates t hat t he use of this gas as a cont rol l i ng par amet er is less advant a-
geous in boilers fired on nat ur al gas [2].
Meas ur ement of the flue-gas oxygen cont ent oft en provides a good indi-
cat i on of combust i on per f or mance, but it must be appreci at ed t hat t he
presence of ' t r amp air' due to leakages into t he combust i on chamber can
l ead to anomal ous readings. I n t he presence of significant leakage,
reduci ng t he ai r / f uel ratio to mi ni mi se t he flue-gas oxygen cont ent can
result in t he burners being st arved of air. This is an ar ea where systems
based on car bon monoxi de measur ement s provi de bet t er results since t he
car bon monoxi de cont ent of the gases is a di rect i ndi cat i on of combust i on
per f or mance and is unaffect ed by the presence of t r amp air.
A system which adjusts the f uel / ai r ratio in relation to the flue-gas
oxygen cont ent is shown in Fi gure 5.5. The oxygen measur ement is fed to a
cont rol l er (5) whose out put adjusts the f uel / ai r ratio by var yi ng the multi-
pl yi ng fact or of a gai n block (8).
The t ransmi t t ers used for measuri ng flue-gas oxygen are usual l y based
on the use of zi rconi um probes, whose conduct i vi t y is affected by the
oxygen cont ent of the at mosphere in which t hey are installed. True two-
wire 4 - 2 0 mA analysers are now avai l abl e (Figures 5.6 and 5.7), and are
bot h accur at e and reliable.
The flue gases leave the combust i on chamber t hrough ducts of consider-
able cross-sectional ar ea and it is inevitable t hat a significant degree of
stratification will occur in the gases as t hey flow to the chimney. Air
ent eri ng the furnace t hrough the registers of idle burners will t end to
pr oduce a hi gher oxygen cont ent in the gases flowing al ong one area of the
duct t han will be present in anot her area, where fewer burners may be
idle.
It is t herefore necessary to t ake consi derabl e care t hat any gas analysis
provides a t rul y represent at i ve sampl e of t he average oxygen content, and
this demands t hat great care should be exercised over the selection of the
Flue-gas
oxygen
Boiler oontent Fuel flow
load
6 7
F 1
Ai rfl ow
Ai r ~ damper
Figure 5.5 Oxygen trimming of fuel~air ratio
Figure 5. 6 An in-situ oxygen analyser
Fisher Rosemount Ltd. Reproduced by permission
l oc a t i on o f t he anal ys er . Wi t h l a r ge r duct s i t ma y be neces s ar y t o p r o v i d e
sever al anal yser s. T h e si gnal s f r o m t hese c a n be c o mb i n e d , or t he o p e r a t o r
c a n be gi ve n t h e f aci l i t y t o sel ect one or mo r e o f t h e m f or use.
A be t t e r o p t i o n is n o w avai l abl e. T h e p o we r a n d fl exi bi l i t y of mo d e r n
c o mp u t e r - b a s e d c ont r ol syst ems al l ows f or t r ul y i nt e l l i ge nt s a mp l i n g t o be
a ppl i e d, wh e r e t he s ys t em r ecogni ses t he d y n a mi c st at us o f t he pl a nt , s uch
as wh i c h b u r n e r s ar e b e i n g fi red, a n d a u t o ma t i c a l l y selects t he anal ys er
s i gnal t o be used, or i nt e l l i ge nt l y mi xes t he a na l ys e r si gnal s t o opt i mi s e per -
f or ma nc e . T h e i ns t a l l a t i on o f s uc h a s ys t em r equi r es car ef ul obs e r va t i on o f
FLUE GAS FLOW I I ~.~O*,~,d~
/ ~ to flue duct I
I I
/ ~ 1N NI~" ~ 1 ~
connection I
connection
i , ,
Calibration gas
Input connection
I Conduit
connection
Figure 5.7 Installation o f an in-situ oxygen analyser
t he pl ant per f or mance over an ext ended peri od and t he devel opment and
subsequent appl i cat i on of a suitable system based on those observations.
Al t hough such techniques ar e possible. Despite t he considerable
advances t hat have been made in gas-analyser t echnol ogy over t he past
few years, f uel / ai r rat i o t r i mmi ng on t he basis of gas analysis is still t r eat ed
wi t h some reservation. It is general l y accept ed t hat t he measur ement s ma y
occasionally fail or be mi sl eadi ng and for this reason it is usual to allow
manual i nt ervent i on in t he absence of reliable oxygen control. I n Fi gure
5.5 this facility is pr ovi ded by t he ha nd/ a ut o station (7). I n addition, a
ma x i mu m/ mi n i mu m l i mi t er block (6) restricts t he amount of adj ust ment
t hat is permi t t ed, to const rai n t he effects of anomal ous or invalid measure-
ment s or i ncorrect cont rol actions.
This system also characterises t he set-value signal for t he oxygen con-
troller over t he boiler' s l oad range by means of a funct i on block (4),
provi di ng for hi gher excess-air operat i on at low loads. The i ndi cat i on of
boiler l oad ma y be obt ai ned from ei t her st eam flow or air flow, and the
exact shape and par amet er s of t he oxygen versus load charact eri st i c will
be defined by t he boiler desi gner or process engineer.
In practice, facilities may also be i ncorporat ed to allow t he oper at or to
adjust t he system by biassing t he load signal upwards or downwar ds at any
given poi nt to yi el d bet t er combust i on wi t h reduced stack emissions.
Because t he oxygen cont ent of air is 21% by vol ume (or roughl y 23%
by weight), a given change in oxygen cont ent represents appr oxi mat el y
five times t hat change in t erms of excess air. Since it is i ndeed airflow t hat is
bei ng cont rol l ed, t he oxygen loop must recognise t he presence of this high-
gai n component , and t he gain of t he cont rol l er (5) should be set at a kick-
of f low val ue (typically 0.25, or a proport i onal band of 400%). The t i me
constants of t he fuel / ai r/ fl ue-gas system are long, and t he i nt egral t er m of
t he oxygen cont rol l er will t herefore also t end to be long.
5.1.3.2 Combining oxygen measurement with other parameters
The use of an oxygen- t r i m signal on its own can be misleading, for t he
reasons not ed earlier, and bet t er per f or mance can be obt ai ned by
combi ni ng oxygen t ri m wi t h t he opaci t y of t he flue gases, since reduci ng
t he ai r flow event ual l y results in the product i on of visible smoke. However,
it is usually undesi rabl e to operat e a boiler in t he regi on wher e smoke is
bei ng pr oduced, and an i mpr ovement is to adjust t he air flow on t he basis
of anot her par amet er , such as car bon monoxide.
Figure 5.8 shows how the car bon monoxi de and oxygen measur ement s
can be combi ned to t ri m t he f uel / ai r ratio. Basically, t he system comprises
two gas-analysis controllers (6 and 10) whose set-value signals are det er-
mi ned in rel at i on to the boiler load (via funct i on generat ors 5 and 7).
However, t he set val ue for t he oxygen cont rol l er is also t r i mmed by t he
out put of t he car bon monoxi de cont rol l er (the two signals bei ng combi ned
in s ummat or 9). Ha nd/ a ut o facilities enabl e t he system to operat e with
bot h analysers in command, or wi t h onl y oxygen t ri m in service (the CO
cont rol l er bei ng on manual at ha nd/ a ut o station 8), or with fully manual
f uel / ai r rat i o adj ust ment ( hand/ aut o station 12 bei ng on manual , to
isolate bot h gas-analysis controllers).
I n anot her var i ant of this system, ei t her of t he two flue-gas analysis con-
trollers can be selected for operat i on, ei t her by manual i nt ervent i on or
aut omat i cal l y by means of a maxi mum-sel ect i on function.
5.1.3.3 Using carbon-in-ash measurements
I n boilers bur ni ng solid fuels, t he car bon cont ent of t he ash has t radi t i on-
ally been used to provide an i ndi cat i on of the completeness of combustion,
Boiler
load
Flue-gas
carbon monoxide
content
5
~
7
Figure 5.8
F
"I
1
f 18
Flue-gas
oxygen
comer#
D
11 12
Combined CO and 02 trimming o f fuel/air ratio
1
13
14
r
Ai r control damper
since any carbon remai ni ng in the ash indicates t hat i ncompl et e combus-
tion has occurred. Until comparat i vel y recently, accurate online carbon-
in-ash sampling was not possible and measurement of this par amet er
required manual sampl i ng and analysis. With the emergence of online
analysers t he picture has changed, and tests have indicated that online
measurement can play a useful part in optimising the combust i on process
[3]. I n addition, analysis of unbur ned carbon can indicate whet her the
coal mills (pulverisers) require adjustment. However, the long transfer-
time constants of the combust i on process coupled with the comparat i vel y
slow response of the instruments and problems of stratification [4] suggest
t hat this t echni que is onl y useful for l ong-t erm correct i on of firing, wher e
relatively stable l oad condi t i ons can be mai nt ai ned for ext ended periods.
5. 1. 4 Mul t i pl e- burner systems
The systems t hat have been described so far are based on t he adj ust ment
of t he total quant i t y of fuel and ai r t hat is admi t t ed to t he combust i on
chamber . Thi s appr oach ma y suffice wi t h smal l er boilers, wher e adjust-
ment of a single fuel valve and ai r damper is reasonable, but l arger units
will have a mul t i pl i ci t y of burners, fuel systems, fans, damper s and com-
bust i on-ai r supplies. I n such cases pr oper consi derat i on has to be given to
t he di st ri but i on of ai r and fuel to each bur ner or, i f this is not practical, to
small groups of burners. Agai n, suitable st andar ds have been devel oped by
t he NFPA for t he design of t he pl ant and cont rol systems of such boilers
[5].
The concept of i ndi vi dual l y cont rol l i ng ai r registers to provi de t he
correct f uel / ai r rat i o to each bur ner of a mul t i bur ner boiler has been
i mpl ement ed, but in most pract i cal situations t he expense of t he instru-
ment at i on cannot be justified. Oil and gas burners can be oper at ed by
mai nt ai ni ng a defi ned rel at i onshi p bet ween t he fuel pressure and t he dif-
ferent i al pressure across t he bur ner air register ( r at her t han pr oper flow
measurement s), but even wi t h such economi es t he capi t al costs ar e high
and t he payback low. The need to provi de a modul at i ng act uat or for each
ai r register adds f ur t her cost.
A mor e pract i cal opt i on is to cont rol t he ratio of fuel and air t hat flows
to groups of burners. Fi gure 5.9 shows how t he principles of a simple cross-
l i mi t ed system ar e appl i ed to a mul t i bur ner oil-fired boiler. The pl ant in
this case comprises several rows of burners, and t he flow of fuel oil to each
row is cont rol l ed by means of a single valve. The combust i on air is
supplied t hr ough a common wi ndbox, and t he flow to t he firing burners is
cont rol l ed by a single set of secondary-ai r dampers.
I n most respects t he ar r angement closely resembles t he basic cross-
l i mi t ed system shown in Fi gure 5.4, wi t h t he oil flow i nferred from t he oil
pressure at t he row. A funct i on gener at or is used to conver t t he pressure
signal to a fl ow-per-burner signal, whi ch is t hen mul t i pl i ed by a signal
represent i ng t he number of burners firing in t hat row, to yield a signal
represent i ng t he total amount ofoi l flowing to t he burners in t he group.
The system operat es in exact l y t he same way as t he basic confi gurat i on
of Fi gure 5.4, and it is r epeat ed for each row of burners, so t hat t he rat i o of
t ot al fuel-oil flow to t ot al ai r flow ent er i ng t he boiler is mai nt ai ned at t he
desi red value. The mast er de ma nd and t he oxygen- t r i m signals are fed to
all t he rows to keep t he firing rat e in step wi t h t he l oad demand and t he
flue-gas oxygen cont ent at t he correct level.
This basic confi gurat i on is not restricted to oil-fired boilers. It can also
be used wi t h gas-fired pl ant and it can be applied to systems burni ng a
mi xt ur e of fuels, wi t h suitable modifications as will now be described.
5.2 Working wi t h mul t i pl e fuel s
The cont rol systems of boilers bur ni ng several di fferent types of fuel have
to recognise t he heat -i nput cont ri but i on being made at any time by each of
t he fuels, and t he ar r angement s become mor e compl i cat ed for every addi -
tional fuel t hat is to be considered.
O i l p r e s s u r e
I a t b u r n e r g r o u p
i =
'
I
I N u m b e r o f i
I
I
I
i t
I
I
i
i
i
i
I
I
I
I
i
i
I
i
I I
i
I I
I I
I I
I B u r n e r - g r o u p o i l v a l v e
M a s t e r S e c o n d a r y - a i r
i d e m a n d f l o w t o b u r n e r F l u e - g a s
g r o u p o x y g e n c o n C n t
,
m g r o u p i- i
i i . i I '
===l~ ~ T~gut~huoi r
I B u r n e r f u e l
i s y s t e m [ - Z 1
I S e c o n d a r y - a i r c o n t r o l
i d a m p e r
Figure 5.9 A control system f or multiple burners (one burner group shown)
Fi gure 5.10 shows a system for a boiler burni ng oil and gas. The similari-
ties to the simple cross-limited system are very apparent , as are the
commonal i t i es wi t h the fuel-control par t of the muhi bur ner syst em (shown
wi t hi n the chai n- dot t ed ar ea of Fi gure 5.9).
The cross-limiting funct i on is performed at the mi ni mum-sel ect or
bl ock (5) whi ch cont i nuousl y compar es the mast er demand wi t h the
quant i t y of combust i on air flowing to the common wi ndbox of t he bur ner
group. The gai n block (6) translates the air flow into a signal represent i ng
the amount of fuel whose combust i on can be suppor t ed by the avai l abl e
secondar y air.
Master Gas Gas
demand Secondary temperature pressure
air flow
Oil pressure
at burner group
1 ( ~ Number of
, T
oil burners
firing in
group
pressure/flow
or gas
per
unit of
5 e air 7 ,
Burner-group oil valve
Burner gas F
pressure/flow
Gas equi val ent ~ Number of
of oil firing, gas burners
firing in group
~ r 14
Oil equivalent
of gas firing
Burner-group gas valve
f 1
9
f l
Figure 5.10 Controlling multiple fuels (one burner group shown)
The selected signal (the load demand or t he available air) ul t i mat el y
forms t he desired value of bot h t he gas and oil closed-loop controllers. But,
before it reaches t he rel evant cont rol l er a value is subt ract ed from it, whi ch
represents t he heat cont ri but ed by the ot her fuel (convert ed to t he same
he a t / m s value as t he fuel bei ng controlled). The conversion of oil flow to
equi val ent gas flow is performed in a funct i on gener at or (10), while t he
ot her conversion is per f or med in anot her such block (14). Each of t he t wo
s ummat or units (11 and 13) algebraically subtracts t he ' ot her-fuer signal
from t he demand.
Not e that, in t he case of this system, the gas pressure signal is compen-
sated against t emper at ur e variations, since the pressure/ fl ow relationship
of t he gas is t emper at ur e- dependent .
As before, each fuel-flow signal represents t he flow per burner and so it
has to be multiplied by t he number of burners in service in or der to
represent the total fuel flow.
These di agr ams are hi ghl y simplified, and in pract i ce it is necessary to
i ncorporat e various features such as interlocks to prevent overfiring and to
isolate one or ot her of t he pressure signals when no bur ner is firing t hat
fuel. (This is because a pressure signal will exist even when no firing is
t aki ng place.)
5.3 The c ont r ol of coal mi l l s
So far, we have looked at boilers where the i nput of fuel can be measured
and where its flow can be regul at ed by means of one or mor e valves. Wi t h
boilers burni ng coal, the mill (or pulveriser) system must be t aken into
consideration. The mills have al r eady been described in Chapt er 3, now
we shall look at how t hey are controlled.
But first it has to be underst ood that, because t he mill has to meet
defined per f or mance guarant ees, t he control st rat egy to be applied in a
given installation must be devel oped in association with t he manuf act ur er
of t he mill. Once t hat st rat egy has been agreed it must be applied to each
of t he mills t hat feed the boiler. The demand is fed in parallel to all t he mill
sub-systems, with facilities for biassing t he signal to any one of t hem wi t h
respect to the others.
5.3.1 The 'load line"
The dr op in pressure experi enced by air flowing t hr ough a mill will be
det er mi ned by t he geomet r y of t he mill, t he amount of coal in it and t he
vol ume of air flowing t hr ough it. Fi gure 5.11 shows schemat i cal l y t hat a
MI L L E M P T Y
MI L L F UL L
Low pressure drop
High pressure drop
MI L L E MP T Y MI L L FULL
Low pressure drop High pressure drop
Figure 5.11 Effect of coal load and airflow on cross-mill differential pressure
hi gh pressure-drop across t he mill ma y be t he result of a hi gh coal l oad in
t he mill or a hi gh ai r flow t hr ough it, or a combi nat i on of both. The air-
flow rat e will bear a square-l aw relationship to t he di fferent i al pressure
across t he mill, and t he di fferent i al pressure across a restriction such as a
flow nozzle or an orifice pl at e will also have a square-l aw relationship wi t h
t he air flow. Fr om this, it can be appreci at ed t hat t he charact eri st i c curve
rel at i ng t he mill di fferent i al pressure and t he pr i mar y- ai r di fferent i al
pressure will be a st rai ght line. Thi s is called t he ' l oad line' and is specific to
a given design of mill oper at i ng under defined conditions. The manuf ac-
t ur er will define t he correct load-line par amet er s and scales for a given
design of mill.
5.3.1.1 Load control strategies f o r pressurised mills
Wi t h pressurised mills, some cont rol systems oper at e on t he principle of
compar i ng t he t wo differential-pressure signals and modul at i ng t he feeder
speed to keep t he rel at i onshi p bet ween t he two in t rack wi t h t he load line,
as shown in Fi gure 5.12. The met hods of var yi ng t he speed of t he feeder
i ncl ude vari abl e-rat i o gearboxes or vari abl e-speed motors.
Master demand
~ s I - p r e $ s u r 8
acrossm#l
Differential-pressure
across flow-nozzle in
PA line
0
P A ~ control
Figure 5.12
()
7
4
0
MiU-differential / PA-differential control system
?
F]
The speed of t he feeder is sometimes fed back to t he mast er system as
an i ndi cat i on of coal flow, to provi de a degree of closed-loop operation. It
is not a perfect solution, since a change in the calorific value of t he coal
cannot be det ermi ned by this system. But, in t he absence of reliable and
fast systems for measuri ng the heat i nput from coal, it is as good as can be
achieved.
Al t hough the system described above provides an adequat e met hod of
control, it cannot deal with changes in the pri mary-ai r (PA) flow caused
by external factors. Therefore, if t he PA flow changes, the system must
wait for t he resulting change in steam pressure before a correction can be
made.
An approach to overcomi ng this limitation is to provide closed-loop
control of t he pri mary-ai r flow, as shown in Figure 5.13. Here, because the
system detects and i mmedi at el y reacts to changes in PA flow, and adjusts
the flow-control damper to compensate, disturbances to steam product i on
are minimised. Again, a feeder-speed signal, representing fuel flow, is fed
Q
Mas ter d eman d
P A f l o w c ontrol
Figure 5.13
Di fferenti el -pres s ure
ac ros s fl ow-nozzl e h,
P A l i ne
Di fferenti al -pres s ure
ac ro s s mi l l
Closed-loop control of PA flow
F eed er - s p eed control l er
back to t he mast er system to provide closed-loop correct i on of speed
changes, whi ch woul d otherwise i nt roduce di st urbances to t he st eam
pressure.
Both of these systems adjust the feeder speed after t he PA flow has been
changed, and this can lead to del ayed response to changes in demand.
Fi gure 5.14 shows a system t hat adjusts t he feeder speed in parallel with t he
PA flow. Thi s also shows some pract i cal refinements: a mi ni mum- l i mi t
block t hat prevent s t he PA flow from being reduced below a predet er-
mi ned limit, and a mi ni mum selector block whi ch prevent s t he coal feed
being i ncreased above t he availability of pr i mar y air (the bias uni t sets t he
mar gi n of ai r over coal).
5.3.1.2 Load control systems f o r suction mills
I n br oad terms, the l oad-cont rol strategies for suction mills follow similar
principles to those of t he pressurised mills as described above. A very
Master demand
Minimum limit
o n P A f l o w
demand
Limits feeder speed
w~th respect to
available PA flow
I r
F1
Bias
' V
F1
Differential-pressure
across flow-nozTJe in
PA line
?
r l
Feeder speed PA flow
Figure 5.14 Parallel control of feeder speed and PA f l ow
Mast er demand
Figure 5.15
0
Feeder speed
Simple suction-mill control
8
Exhaust er f l ow
Mast er demand
r
7q
Ai r pressure at
AV fl ow to mill mi l l inlet
?
Feeder speed E xhaus ter
flow-control
damper
Figure 5.16 Improved control of suction mill
Mill inlet damper
simple t echni que is to adjust t he speed of t he coal feeder in parallel wi t h
the flow t hr ough the exhauster, as shown in Figure 5.15. Here again, the
feeder speed is ret urned to t he mast er system to correct for speed variations
t hat woul d otherwise di st urb t he steam pressure.
Thi s system provides open-l oop operat i on of the mill and, once again,
i mproved performance can be achieved by the use of a closed loop ar ound
t he air flowing t hrough t he mill, as shown in Figure 5.16. I n this system, an
addi t i onal cont rol loop mai nt ai ns a const ant air pressure at t he mill inlet.
Wi t h these systems, it is again necessary to feed back to t he mast er
system a signal t hat represents the i nput of fuel from t he mill to t he
combust i on chamber. Feeder speed provides this function, and t hereby
minimises steam-pressure disturbances.
5.3.2 Mi l l temperature control
It is very i mpor t ant t hat t he t emperat ure of t he air in t he mill should be
mai nt ai ned wi t hi n close limits. For many reasons, i ncl udi ng i nadequat e
dryi ng of the coal, combust i on efficiency will be reduced if t he t emperat ure
is too low, while too hi gh a t emperat ure can result in fires or explosions
occurri ng in t he mill. The control techniques for bot h pressurised and
suction mills involve mi xi ng hot and cold air streams to achieve t he correct
t emperat ure. However, whereas pressurised coal mills require t he use of
two dampers for this purpose (one controlling the flow of hot air, t he ot her
the cold air) in a suction mill only one damper needs to be adjusted, to
admi t more or less cold air into t he stream of hot air being dr awn into t he
mill by the exhausters.
Figure 5.17 shows a t emperat ure control system for a pressurised mill,
with one act uat or provi ded for the hot-air damper and anot her for the
t emperi ng-ai r damper.
Sometimes the two dampers are linked mechani cal l y and positioned by
a single actuator. The use of two separate actuators adds cost, but allows
for a great er degree of operat i onal flexibility since it allows the openi ng of
each damper to be biassed with respect to t he ot her from the central
Mi l l e x i t temperature
+
IOO~
!
o~
Hot-ak" demper
I
~o Ternpedng-al r damper
4 position
i I ~
J
i I
4
4
4
4 I
10o~
i i
controt~r ootOot sign,,t
T e m p e r ~ i r Ho t - a i r
damper damper
Figure 5. 17 Mi l l temperature control
cont rol room. Thi s enables the oper at or (or a sophisticated control
system) to optimise mill per f or mance whilst still mai nt ai ni ng t he mill tem-
per at ur e at t he correct value.
5. 3. 3 Controlling multiple mills and multiple f ue l s
Lar ge coal-fired boilers are provi ded with several coal mills, each of
whi ch has its own cont rol subsystem as described above, and in addi t i on
t hey i nvari abl y bur n ot her fuels as well as coal.
Fi gure 5.18 shows t he mill-control system of such a pl ant in simplified
form. It is present ed here to illustrate a r equi r ement t hat is an i nt egral
par t of boiler cont rol systems: the need to handl e applications where a
single cont rol l er sends commands to several subloops in parallel, and
wher e any of t he subloops may be isolated at will from t he controller.
Here, each mill feeds a gr oup of burners (say six), and each of these
groups may also fire fuel-oil. Since, at any given time, any mill group may
be out of service, oper at i ng at a fixed t hroughput , or otherwise requi ri ng
i ndependence from t he ot her groups, the overall loop gain will change,
and this is addressed by t he gai n-compensat i on block (item 4) in the
mast er - demand signal line. The demand signal from this block is fed to
each group via individual ha nd/ a ut o stations, one for each mill group
(i t em 10).
The out put of each of these stations event ual l y becomes t he desired
val ue for t he rel evant pr i mar y- ai r flow cont rol l er (17), but first t he heat
cont ri but i on from any oil burners firing in t hat group must be t aken into
consideration. 3"his i nput is derived from a measur ement of t he oil pressure
at t he burners in t he group (1), convert ed to represent t he oil flow per
bur ner (by means of funct i on block 2) and t hen multiplied (4) by the
number of oil burners in t hat group t hat are firing at t he time. The
resulting signal is t hen convert ed (9) to represent t he amount of coal t hat
woul d equat e to t hat quant i t y ofoi l , and this is subt ract ed from the mast er
de ma nd (block 12) to represent the amount of coal firing t hat is needed
from the group. Thi s firing demand is prevent ed from falling below a safe
pr edet er mi ned val ue ( mi ni mum- l i mi t block 15).
By account i ng for t he oil firing, the openi ng of t he pr i mar y- ai r damper
is i mmedi at el y adj ust ed if an oil bur ner trips, or if one is br ought into
service, to compensat e for t he change, wi t hout wai t i ng for t he heat -i nput
effects to be det ect ed via the mast er-pressure controller.
5.3.3.1 The challenge of hand/auto changeover
The heat i nput from a l arge coal mill can be as much as 100 MW, but
t he mechani cal design of t he mill and its auxiliaries is such t hat it can var y
Oil pressure Master demand
Mi l l
di f f erent i al
pr essur e

t5
Set mi l l
l oad l i ne
r
Feeder speed
cont r ol l er
Figure 5.18
Number o f
bur ner s fi ri ng
Pr i mar y- ai r
di f f erent i al Pr i mar y ai r
pr essur e t emper at ur e
sa-J
. , , . . .
Temperat ure
compensat i on
,
o~w
I h~r r ~r
4
~ . ~ O#
bur ni ng at
, t hi s gr oup
Coal
e q ~ of
oi l bur ni ng at
thi s gr oup A
Gai n
compensat i on
, ( number o f
" ~ " mi l l s i n
To ot her mi l l gr oups
5
Mi l l gr oup mast er
HanOVAuto st at i on
12
Mi ni mum
pr~nary-a~
15
Pr i mar y- ai r
d a mp e r
actuator
A comprehensive mill control system (one millgroup shown, excluding
temperature control)
t he t hr oughput by onl y a compar at i vel y small amount , cert ai nl y no mor e
t han 50%. Therefore, t he i nt r oduct i on of one mill to t he heat i nput of such
a boiler amount s to a step change of as much as 50 MW, and t he change in
t hr oughput t hat can be smoot hl y modul at ed is also 50 MW. Such l arge
step-changes requi re efficient modul at i on of any ot her fuels t hat are bei ng
fired at t he same time.
These factors make it i mpract i cal to consi der st art i ng up mor e t han
one mill at a t i me and requi re t he facility of allowing any mill to be
oper at ed under manual or aut omat i c cont rol , i ndependent of t he others.
Thi s brings about a severe chal l enge to t he DCS software.
The mast er de ma nd is fed in paral l el to several subloops, one for each
mill group. On st art -up of t he pl ant all of these will be under manual
cont rol . Whe n t he mill has r eached a t hr oughput of roughl y 50% of its
capaci t y, or when ot her conditions det er mi ne t hat aut omat i c cont rol is
now possible, t he oper at or will switch t he mast er demand into service. The
difficulty is t hat up to t hat instant, t he system cannot be made aware of
whi ch mill gr oup is about to be t ransferred to respond to t he mast er signal,
and each group ma y be oper at i ng at a ver y di fferent t hr oughput from any
other.
Whi l e a loop is bei ng t ransferred from manual to aut omat i c cont rol
(or vice versa), it is i mpor t ant t hat t he pl ant is not subjected to a sudden
di st urbance. At t he moment of changeover, t he ' hand' and ' aut omat i c'
signals must be equal. Thi s is called ' bumpless transfer' , and it can be
achi eved by provi di ng t he oper at or wi t h indications of bot h signals so t hat
t hey can be made equal before changeover is initiated. However, such a
system woul d not be accept abl e in most cases, since t he process of
changi ng from one mode of cont rol to anot her should be as quick and
simple as possible, and should not requi re t he oper at or to undul y di st urb
t he operat i on of t he plant.
To achieve what is know as ' procedureless, bumpless t ransfer' from
manual to aut omat i c control, a common t echni que is to make t he con-
t rol l er out put follow (or ' t rack' ) t he manual demand, so t hat when t he
system is switched to aut omat i c the signal to t he act uat or is not subjected
to a sudden change.
Thi s is easy enough wi t h a single cont rol l er positioning a single
act uat or, but what happens when one cont rol l er commands several
subloops as shown in Fi gure 5.18? It is cl earl y impossible to force t he
mast er cont rol l er out put to adopt a val ue t hat cannot be known ahead of
time, or to change t he out put of t he cont rol l er i f it is al r eady modul at i ng
one or mor e mills.
This probl em is frequent l y not recognised by DCS vendors who have
little or no experi ence of boiler control, and it can be qui t e difficult to
expl ai n it to t hem. But under st andi ng it and resolving it are absolutely
essential i f t he system is to be expect ed to operat e smoot hl y and wi t h
mi ni mal oper at or i nt ervent i on. Various solutions have been developed,
such as ' freezing' t he mast er demand while t he t ransfer is effected and
gr adual l y r ampi ng one signal up or down to mat ch t he other. It is
i mpor t ant , however, t hat t he DCS vendor should be able to demonst r at e
t he solution offered wi t hi n t hei r system, and t hat t hey should be able to
demonst r at e its use on an existing power plant.
5. 3. 3. 2 Compl exi t y o f screen di spl ays
I n considering t he oper at or displays associated wi t h a system such as t hat
shown in Fi gure 5.18, at t ent i on should be given to t he vast amount of infor-
mat i on t hat must be provided. The di agr am given her e is necessarily
simplified, and excludes the ma ny interlocks and ot her functions t hat are
requi red in reality. When a pract i cal pl ant is considered it soon becomes
appar ent t hat accommodat i ng t he amount of i nformat i on and cont rol
facilities can l ead to very cl ut t ered display screens.
Clearly, t he mill groups are car bon copies of each other, var yi ng onl y
in respect to t he t ag numbers of each i t em and t he dynami c i nformat i on
rel at i ng to each ar ea of t he plant. It is t herefore reasonable to display onl y
one group at a t i me on t he screen, allowing it to be st art ed, adjusted or
stopped as requi red. However, to avoid maki ng any mistakes, t he oper at or
should be very cl earl y and unambi guousl y i nformed of whi ch group is
displayed at any time. Also, a mast er display should enabl e t he oper at or to
view t he status of t he ent i re set of mills feeding t he boiler.
The devel opment of these oper at or displays is t herefore unusual l y
demandi ng and i f insufficient t i me or money is al l ocat ed to t he perfor-
mance of this task t he results can be at best unwi el dy and at worst
dangerous.
5.4 Draught control
I n Chapt er 3 we saw t hat , in a fired boiler, t he air requi red for combust i on
is provi ded by one or mor e fans and t he exhaust gases ar e dr awn out of t he
combust i on chamber by an addi t i onal fan or set of fans. On boilers wi t h
ret ro-fi t t ed flue-gas desul phuri sat i on plant, addi t i onal booster fans ma y
also be provided. The cont rol of all these fans must ensure t hat an
adequat e supply of air is available for t he combust i on of t he fuel and t hat
t he combust i on chamber operat es at t he pressure det er mi ned by t he boiler
designer. In a fluidised-bed boi l er the air must also provi de the pressure
r equi r ed to mai nt ai n t he bed in a fluid state.
All of the fans also have to cont ri but e to the provision of anot her
i mpor t ant f unc t i on- - pur gi ng of the furnace in all condi t i ons when a
collection of unbur ned fuel or combust i bl e gases coul d ot herwi se be acci-
dent al l y ignited. Such operat i ons are requi red pri or to l i ght -off of the first
bur ner when t he boi l er is bei ng started, or after a trip.
The cont rol systems for the fans have to be designed to meet the requi re-
ment s of st art -up, nor mal oper at i on and shut -down, and to do so in the
most efficient manner possible, because the fans may be physically l arge
and requi re a l arge amount of power for their oper at i on (several MW in
some cases). In addi t i on, as we saw in Chapt er 3, the per f or mance con-
straints of t he fans, such as surge and stall, have to be recognised, if
necessary by t he provi si on of special cont rol functions or interlocks.
Chapt er 3 also descri bed t he met hods of cont rol l i ng the t hr oughput of
the fans, i.e. pi t ch-cont rol , dampers, vanes or speed adj ust ment . I n the
present chapt er we shall exami ne how these elements are adj ust ed to
address the oper at i onal requi rement s of the boiler.
5.4.1 Maintaining the furnace draught
Apar t from suppl yi ng air to suppor t combust i on, the FD fans have to
oper at e in concer t wi t h the I D fans to mai nt ai n the furnace pressure at a
cert ai n value. The heavy solid line of Fi gure 5.19 shows the pressure profile
t hr ough the vari ous sections of a typical bal anced- dr aught boi l er system.
It shows t he pressure from the poi nt where air is dr awn in, to the poi nt
where the flue gases are exhaust ed to the chimney, and demonst rat es
how the combust i on chamber operat es at a slightly negative pressure,
whi ch is mai nt ai ned by keepi ng the FD and I D fans in bal ance wi t h each
other.
I f t hat bal ance is di st urbed the results can be ext remel y serious. Such
an i mbal ance can be br ought about by the acci dent al closure of a damper
or by the sudden loss of all flames. It can also be caused by mal oper at i on of
t he FD and I D fans. The dashed line on the di agr am shows the pressure
profile under such a condi t i on, whi ch known as an ' implosion' . The results
of an i mpl osi on are ext remel y serious because, even t hough the pressures
involved may be small, the surfaces over whi ch t hey are appl i ed are very
l arge and t he forces exert ed become enormous. Such an event woul d
al most cert ai nl y result in maj or st ruct ural damage to the plant.
P r e s s u r e
p r o f i l e
Wi n d b o x
\
\
E l ~ E x h a u s t
g a s e s t o
c h i m n e y
i i i i ~ i , ' i ~ i i
i i i t i i ~ i i i !
i i i i i ! i i I
i i i i i i N u m l a l I i i i
operation ,
y . - . . - : : L _ i i - - i i . . . . ~ t -tL~-
Implosion
C O l IdJJJOt~
Figure 5.19 Draught profile of a boiler and its auxiliary plant
5.4.2 Fan control
The t hr oughput of t wo fans operat i ng t oget her can be regul at ed by a
common cont rol l er or by i ndi vi dual controllers for each fan. Al t hough a
single cont rol l er cannot ensure t hat each fan delivers the same flow as its
part ner, this confi gurat i on is much simpler to t une t han the al t ernat i ve,
wher e t he t wo controllers can i nt eract wi t h each ot her and make opt i mi sa-
tion ext remel y difficult. Whi chever opt i on is used, t he cont rol system must
be designed to pr ovi de sufficient air to suppor t combust i on.
In the simplest case, the fan or fans will be dri ven by a cross-limited
syst em (see Fi gure 5.4), but wi t h mul t i bur ner installations the flow must be
cont rol l ed for each bur ner or gr oup of burners. The system shown in
Fi gure 5.9 shows how this is ar r anged by regul at i ng t he secondar y air flow
to each bur ner group. In such cases this air suppl y is dr awn from a
common wi ndbox whi ch is mai nt ai ned at a pressure whi ch may be fixed or
var yi ng wi t h boi l er t hr oughput .
Fi gure 5.20 shows how such a cont rol system can be i mpl ement ed. The
desi red-val ue signal for t he pressure cont rol l er is deri ved from st eam flow,
so t hat the pressure in the wi ndbox will change over t he boi l er l oad range,
to a charact eri st i c t hat will be defined by the process engineer. The
S team f l ow
~ x pres s ure
(or windbox/rurnaoe
differential pres s ure)
<
Set r ~r ~t ~ml
pres s ure
Figure 5.20 Controlling the windbox pressure
/
V
I=D fan vane actuators
maxi mum- sel ect or uni t ensures t hat t he pr essur e- demand signal cannot
fall bel ow a pr edet er mi ned mi ni mum value. The measur ed val ue for the
cont rol l er can be based on a measur ement of t he wi ndbox pressure or the
wi ndbox- t o- f ur nace differential pressure (whi ch is what the boi l er designer
woul d pr obabl y require).
5.5 Bi nary c ont r ol o f t he c o mb u s t i o n s y s t e m
So far, we have consi dered only t he modul at i ng systems involved wi t h t he
combust i on plant. I n practice, these systems have to operat e in concert
wi t h bi nar y cont rol systems such as interlocks and sequences. The purpose
of an interlock is to co-ordi nat e t he operat i on of different, but i nt errel at ed
pl ant items: t ri ppi ng one set of fans i f anot her set trips, and so on. The
purpose of a sequence system is to provide aut omat i c st art -up or shut-
down of t he plant, or of some par t of it.
The logic for interlock operat i ons will be defined by t he boiler designer
and will probabl y have to compl y with some local, nat i onal or i nt erna-
tional st andard. The systems ar e very specific to t he part i cul ar plant, and
no at t empt will t herefore be made in this book to define these, because t he
objective here is to provide a general overview of boiler cont rol systems.
However, one topic t hat we shall look at is bur ner management since,
like modul at i ng loops, this t ype of system is very dependent on t he correct
operat i on of i nput and out put transducers.
5. 5. 1 Fl ame monitoring
The requi rement s for a comprehensi ve bur ner - management system
(BMS) have al r eady been discussed in Chapt er 3, and at t ent i on was
dr awn t here to the i mpor t ance of fl ame moni t ori ng.
Moni t ori ng t he status of a fl ame is not easy. The det ect or must be able
to di scri mi nat e bet ween t he flame t hat it is meant to observe and any ot her
in t he vicinity, and bet ween t hat flame and t he hot surfaces wi t hi n t he
furnace. The det ect or must also be able to provide reliable det ect i on in t he
presence of t he smoke and st eam t hat may be swirling ar ound t he flame.
To add to t he problems, t he det ect or will be requi red to operat e in t he hot
and di rt y envi r onment of t he bur ner front, and it will be subjected to addi -
tional heat r adi at ed from t he f ur nace into whi ch it is looking.
Wi t h t hei r at t endant BMSs, flame scanners of a boiler ar e vital to t he
safety and prot ect i on of t he plant. I f insufficient at t ent i on is pai d to t hei r
selection, or i f t hey ar e badl y installed or commissioned, or i f t hei r mai nt e-
nance is neglected, t he results can be, at best, annoyi ng. The probl ems will
i ncl ude nuisance trips, pr ot r act ed st art -up of t he boiler and t he creat i on of
hazar dous conditions t hat coul d have serious safety implications.
Fi gure 5.21 shows a typical flame det ect or and t he swi vel -mount i ng
t hat enables its sighting angle to be adjusted for opt i mum performance.
A fl ame scanner is a compl ex opto-electronic assembly, and moder n
scanners i ncorporat e sophisticated technologies to i mprove fl ame recog-
nition and discrimination. Al t hough t he electronics assembly will be
Heat-insulating nipple
Air-purge connection
Swivel mounting
\
!
/
Electronics enclosure
Figure 5.21
/
Alternative air-purge connection
Typical.flame scanner
q) Fireye l,td. Reproduced by permission
designed to oper at e at a high t emper at ur e (typically 65 C), unless great
care is t aken this value could easily be exceeded and it is theretbre
i mpor t ant to t ake all possible precaut i ons to reduce heat conduct i on and
radi at i on ont o the electronic component s. The illustration shows how a
heat-insulating nipple is used to prevent undue heat bei ng conduct ed from
the boiler st ruct ure to the electronics enclosure. It also shows t wo purge-
air connect i ons t hat are provi ded bet ween the electronics enclosure and
the swivel mount . Ei t her of these connect i ons may be used, the ot her being
bl anked off.
5.5.1.1 The requirements f o r purge air
The purge air t hat is supplied to the scanner serves two purposes: it
provides a degree of cooling and it prevents dust, oil and soot from being
deposi t ed on the optical parts of the unit. The air should be avai l abl e at
each burner, even if the bur ner itself is not operating.
It shoul d t herefore be obvi ous t hat t he air used for purgi ng should be
cool, dr y and clean, and that it should be avai l abl e at all times. But, in
many cases these requi rement s are ignored, and the performance of the
i nst rument is t her eby i nevi t abl y degraded.
Purge air can be obt ai ned from the i nst rument -ai r supply, or it can be
provi ded by dedi cat ed blowers. In some cases it is taken from the FD fan
discharge. Each of these is viable, provi ded the requi rement s out l i ned
above have been t horoughl y considered. It is also i mpor t ant t hat the
presence of t he pur ge- ai r supply should be moni t or ed and its loss trans-
mi t t ed to t he DCS, because failure of t he air supply coul d result in
expensive and possibly i rreparabl e damage to the scanners. Moder n
scanners i ncl ude self-monitoring circuits t hat will war n of overheat i ng.
The scanner system should be fail-safe, as a failed system represents t he
loss of a critical link in t he plant' s safety chain. I f it is overri dden, t he
oper at or can become used to oper at i ng wi t hout it in place, and such lapses
can event ual l y creat e a severe hazard.
Fi gure 5.22 shows an installation whi ch clearly demonst rat es exampl es
of neglect, i ncl udi ng a broken pur ge- ai r connect i on and a badl y misa-
ligned scanner. Unfort unat el y, in spite of t he critical i mpor t ance of
reliable fl ame moni t ori ng, it is not too difficult to find such exampl es on
oper at i ng power plant.
Figure 5.22 Example of aflame-scanner installation
5. 5. 1. 2 Fl ame spectra
The spect r um of r adi at i on from a fl ame is det er mi ned by ma ny factors,
i ncl udi ng t he t ype of fuel bei ng bur ned and t he design of t he burner. The
i nt ensi t y of t he fl ame tends to be low for gas and high for coal and oil. The
fl ame will also flicker and, in general , l ow-NOx burners will demonst r at e a
l ower flicker f r equency t han gun- t ype burners. Oi l and coal flames t end to
pr oduce a hi gher degree of i nfrared radi at i on, whereas a gas fl ame is rich
in ul t ravi ol et radi at i on. Radi at i on in t he visible par t of t he spect r um will
also depend on these factors, but these days t he t endency is to use det ect ors
whose response is biassed t owards ei t her t he i nfrared or t he ul t ravi ol et end
of t he spect rum, since emissions in these ranges provi de bet t er i ndi cat i on of
a fl ame t han visible radi at i on, whi ch can be plentiful and misleading.
Each t ype of fuel also produces by-product s of combust i on, whi ch
affect t he t r anspar ency of t he fl ame and t herefore t he bl anki ng effect it has
on adj acent flames or on any flames on t he opposite side of t he furnace. Oil
and coal flames t end to obscure i nfrared radi at i on, while gas flames
pr oduce wat er vapour whi ch obscures ul t ravi ol et radi at i on.
Tabl e 5.1 shows one manuf act ur er ' s advice on t he t ype of fl ame
scanner to use in various applications. It is not i nt ended t hat this table
should be r egar ded as bei ng absolute or rigorous. I n cert ai n ci rcumst ances
a gi ven t ype of fl ame scanner will provi de bet t er or worse per f or mance
t han woul d appear to be i ndi cat ed from t he table. Reput abl e manuf ac-
t urers will be pleased to provi de application-specific gui dance. At t he
design stage this advice will be based on previous experi ence of similar
installations. For a retrofit on an existing plant, t he manuf act ur er should
be asked to car r y out a comprehensi ve site survey, using various types of
scanner, while t he burners are st art ed, oper at ed under various loads, and
stopped. Several tests ma y be requi red, and a survey may last for several
days. The gr eat er the at t ent i on t hat is pai d to this study, t he bet t er will be
t he per f or mance of t he final installation.
5. 5. 1. 3 Burner-management systems and pl ant safety
The design of t he BMS will ai m to address critical safety issues, and t he
sequences for a given t ype of boiler or bur ner will be defined in conj unct i on
wi t h t he pl ant designer, bear i ng in mi nd t he requi rement s of applicable
codes such as NFPA 8502-95. I n fact t he NFPA st andar d defines in some
det ai l t he exact sequences involved in lighting-off, moni t or i ng and
r unni ng- down operat i ons of burners, and shows how these ar e to be linked
wi t h t he pl ant i nt erl ock systems (for exampl e, ensuri ng t hat t he furnace
has been pur ged before any at t empt can be made to initiate a bur ner light-
Table 5.1 Flame-scanner application guide
( Fireye Ltd. Reproduced by permission)
Boi l er t ype Fuel t ype Discrimination capabi l i t y
Infrared Ultraviolet
Front-fired Gas M H
Oil H H
Coal H H
Gas/ oi l M H
Gas/ coal M H
Oi l / coal H H
Coal / oi l / gas M H
Corner-f i red Gas L H
Oil H H
Coal H H
Gas/ oi l L H
Gas/ coal L H
Oi l / coal H H
Coal / oi l / gas L H
Opposed- f i r ed Gas L H
Oil M M
Coal M M
Gas/ oi l L M
Gas/ coal L M
Oi l / coal L M
Coal / oi l / gas L M
H = high, M = medium, L = low
off sequence). For these reasons, t he sequences will not be described here.
However, at t ent i on will be pai d to cert ai n safety-related aspects of BMSs.
Safety requi rement s are very comprehensively defined in every applic-
able standard. For example, NFPA 8502-95 describes t he events and
failures which should be recognised in t he design of the system. The UK
Heal t h and Safety Executive (HSE) has described [6] in considerable
detail t he requi rement s for t he safe design of a software-based system
(defined as a pr ogr ammabl e electronic system (PES)). However, in
practice it is very difficult for the nonspecialist to det ermi ne whet her or not
a system is adequat el y fail-safe. Even using the checklists provi ded in the
HSE document can be inadequate. For example, one i t em in the checklist
asks:
l O0 Power-plant control and instrumentation
Have adequate precautions been specified to protect against electrical inter-
ference in the environment of the PES with regard to:
(i) inherent design oft he PES;
(ii) installation practices (e.g.: separation of power and signal cables);
(iii) Electromagnetic compatibility (EMC) test programme, including con-
ducted interference on power supplies, electro-static discharges and ra-
diated interference?
In t hat it is difficult to say what is or is not adequate in this context, this
is a subjective assessment. For example, one system surveyed by the aut hor
found t hat the effectiveness of very comprehensi ve shielding in the DCS of
a pl ant had been negat ed by the provision of poorl y designed access
doors.
One solution is to assume t hat a pr ogr ammabl e system will occasionally
generat e i ncorrect commands and to therefore ensure t hat all its opera-
tions are cont i nuousl y shadowed by another, i ndependent , system. I f a
di screpancy occurs bet ween the actions of the t wo systems a trip should be
initiated or the rel evant sequence prevent ed from bei ng carri ed out.
5.6 Summary
Havi ng l ooked at the cont rol systems appl yi ng to the combust i on and
dr aught plant, in the next chapt er we shall t urn our at t ent i on to the feed-
wat er systems.
1 ANSON, D, CLARKE, W. H. N., CUNNI NGHAM, A. T. S. and
TODD, P.: ' Carbon monoxide as a combustion control parameter' , J Inst
Fuel, 1971, xl l v (363)
2 DUKELOW, S. G.: ' The control of boilers' (Instrument Society of
America, Research Triangle Park, NC, USA, 1991)
3 ESKENAZ1, D. V., I ) ' AGOSTI NI , M., LEVY, E. K. et al . : ' On-line
measurement of unburned carbon' , EPRI Heat-rate Improvement conference,
EPRI, Palo Alto, CA, 1989
4 HURT, R. H., LUNDEN, M. M., BREHOB, E. G., and MALONEY,
D. J.: 'Statistical kinetics for pulverized coal combustion' . Proceedings of
the 26th international symposium on Combustion, Naples, Italy, 1996
5 NFPA 8502-95 ' Standard for the prevention of furnace explosions/
implosions in multiple burner boilers'. National Fire Protection Association,
Batterymarch Park, Quincy, MA, USA
Combustion and draught control 1 O1
6 ' Programmable electronic systems in safety related applications'. Health
and Safety Executive, Library and Information Services, Broad Lane,
Sheffield, $3 7HQ, UK
Chapter 6
Fe e d- wat e r c ont r ol and
i n s t r u me n t a t i o n
6.1 The p r i n c i p l e s o f f e e d- wa t e r c o nt r o l
The objective of a feed-wat er cont rol system ma y seem simple: it is to
supply enough wat er to t he boiler to mat ch t he evaporat i on rate. But as is
so oft en t he case wi t h boilers, this t urns out to be a surprisingly compl ex
mission to accomplish. Ther e are difficulties even in maki ng t he basic
drum-l evel measur ement on whi ch t he cont rol system depends. The design
of t he cont rol system is t hen f ur t her compl i cat ed by t he ma ny i nt eract i ons
t hat occur wi t hi n t he boiler system and by t he fact t hat t he effects of some
of these i nt eract i ons ar e great er or smaller at various points in the boiler' s
l oad range.
The control-system designer' s task is to devel op a scheme t hat provides
adequat e cont rol under t he widest pract i cabl e range of operat i onal condi-
tions, and to do so in a manner t hat is bot h safe and cost-effective. To do
this it is necessary to under st and t he det ai l ed mechani sms of t he feed-wat er
and st eam systems and to be fully aware of t he operat i onal requi rement s.
I n all but t he smallest and simplest boilers, each of t he i nt errel at ed
factors has to be t aken into account , and it is insufficient to rely on simple
responses to t he t hree par amet er s whi ch seem to be r el evant to t he supply
of feed water: st eam flow, feed-wat er flow and t he level of wat er in t he
dr um.
6.2 One , two and t hree- el ement cont rol
The level of wat er in t he dr um provides an i mmedi at e i ndi cat i on of t he
wat er cont ai ned by t he boiler. I f t he mass flow of wat er into t he system is
great er t han the mass flow of st eam out of it, the level of wat er in the
dr um will rise. Conversely, if the st eam out put is great er t han the feed
inflow, the level will fall.
As stated in Chapt er 2, the purpose of the dr um is not onl y to separat e
the st eam from the wat er but also to provi de a st orage reservoir t hat allows
short -t erm i mbal ances bet ween feed-wat er suppl y and st eam pr oduct i on to
be handl ed wi t hout risk to the plant. As the level of wat er in the dr um
rises, the risk increases of wat er bei ng carri ed over into the st eam circuits.
The results of such ' carry-over' can be cat ast rophi c: cool wat er i mpi ngi ng
on hot pi pework will cause ext reme and localised stresses in the met al and,
conversely, i f the level of wat er falls there is a possibility of the boiler being
damaged, par t l y because of the loss of essential cooling of the furnace
water-walls.
Therefore, the t arget of the feed-wat er cont rol system is to keep the
level of wat er in the dr um at appr oxi mat el y the mi dpoi nt of the vessel.
Gi ven this objective, it woul d appear t hat the simplest solution woul d
appear to be to measure the level of wat er in the dr um and to adjust the
delivery of wat er to keep this at the desired va l ue - - f e e di ng more wat er
into the dr um i f the level is falling, and less i f the level is rising.
Unfort unat el y, the level of wat er is affected by transient changes of the
pressure wi t hi n the dr um and the sense in which the level varies is not
necessarily rel at ed to the sense in which the feed flow must be adjusted. In
ot her words, it is not sufficient to assume t hat simply because the level is
increasing the feed-wat er flow must be decreased, and vice versa.
This st range situation is due to effects known as ' swell' and ' shrinkage' .
Boiling wat er comprises a t urbul ent mass of fluid cont ai ni ng many st eam
bubbles, and as the boiling rat e increases the quant i t y of bubbl es t hat is
generat ed also increases. The mi xt ure of wat er and bubbl es resembles
foam, and t he vol ume it occupies is di ct at ed bot h by the quant i t y of wat er
and by the amount of the st eam bubbl es wi t hi n it. I f the pressure within
the system is decreased, the sat urat i on t emper at ur e is also l owered and the
boiling rat e t herefore increases (because the t emper at ur e of the mi xt ure is
now hi gher in relation to the sat urat i on t emper at ur e t han it was before the
pressure change occurred). As the boiling rat e increases, the densi t y of the
wat er decreases, but since the mass of st eam and wat er has not changed
the decrease in density must be accompani ed by an increase in the vol ume
of t he mixture.
By this mechani sm the level of wat er in the dr um appears to rise, a phe-
nomenon referred to as 'swell'. The rise of level is misleading: it is not
indicative of a real increase in the mass of wat er in the system, which
woul d requi re the suppl y of wat er to be cut back to mai nt ai n the status
quo. In fact, i f the dr op in pressure is the result of the st eam demand
Feed-water control and instrumentation 105
suddenl y increasing, the wat er suppl y will need to be i ncreased to mat ch
the i ncreased st eam flow.
' Shri nkage' is t he opposi t e of swell: it occurs when the pressure rises.
The mechani sm is exact l y the same as t hat for swell, but in the reverse
direction. Shri nkage causes the level of wat er in the dr um to fall when the
st eam flow decreases, and once agai n the del i very of wat er to the boi l er
must be rel at ed to the act ual need r at her t han to the possibly mi sl eadi ng
i ndi cat i on pr ovi ded by the drum-l evel transmitter.
I f a slow change of st eam flow occurs, all is well because the pressure
wi t hi n the system can be controlled. It is when rapi d st eam-fl ow changes
happen t hat probl ems occur since, due to swell or shrinkage, the dr um-
level i ndi cat i on provides a cont r ar y i ndi cat i on of the wat er demand.
Fol l owi ng a sudden increase in st eam demand, whi ch causes the
pressure to dr op (and t herefore t he dr um level to rise), a simple level con-
troller woul d respond by reduci ng the flow of feed water. Equally, a
sudden decrease in st eam flow, whi ch woul d be accompani ed by a rise in
pressure and an at t endant fall in the dr um level, woul d cause a level con-
troller to increase the flow of wat er. Bot h actions are, of course, in the
i ncorrect sense.
The effects of swell and shrinkage, in addi t i on to bei ng det er mi ned by
the rat e of change of pressure, also depend on the relative size of the dr um
and the pressure at which it operates. I f the vol ume of the dr um is l arge in
relation to the vol ume of the whol e system the effect will be smal l er t han
otherwise. I f the system pressure is low the effect will be l arger t han wi t h a
boiler operat i ng at a hi gher pressure, since the effect of a given pressure
change on the densi t y of the wat er will be great er in the low-pressure
boiler t han it woul d if t he same pressure change were to occur in a boi l er
operat i ng at a hi gher pressure.
Faced wi t h this situation, designers of cont rol systems have responded
by i mpl ement i ng a vari et y of solutions. The simplest of these is a ' t wo-
el ement ' system, since it is based on the use of t wo process measur ement s in
pl ace of the single drum-l evel measur ement used above.
6.2.1 Two-element f eed-wat er control
Remember i ng t hat the basic r equi r ement of a feed-wat er cont rol system
is to mai nt ai n a const ant quant i t y of wat er in the boiler, it is appar ent t hat
one way of addressing the pr obl em woul d be to mai nt ai n the flow of wat er
into the system at a val ue whi ch mat ches the flow of st eam out of it. One
version of this system is shown in Fi gure 6.1. Here, the flow is cont rol l ed by
an easily recognised device, a valve. We shall look at valves in mor e dept h
later, but for the moment assume t hat the version used in t he di agr am
mai nt ai ns the rate of water flowing t hr ough the valve at a figure which is
directly proport i onal to t he demand signal from t he controller (i.e. if t he
demand signal varies linearly from 0 to 100%, t he flow rate also changes
linearly between 0 and 100%). Such a valve is said to have a ' linear charac-
teristic' and in t he system shown this is empl oyed in conjunction with a
t ransmi t t er t hat produces a signal proport i onal to steam flow. Used
together, these two devices keep t he paramet ers in step. I f t he t ransmi t t er
produces a signal which is equal to the steam flow at all loads and if t he
flow t hrough t he valve is mat ched with this signal at every poi nt in t he
flow range, a controller gai n of uni t y will ensure that, t hr oughout t he
dynami c range of t he system, t he flow of water will always be equal to the
flow of steam.
S t e a m f l ow 0 - 2 0 k~/ s
,
I i i
4 m A = 0 k g / s
2 0 mA = 2 0 kgl s
FIC
Figure 6.1
4 m A
2 0 mA
F e e d f l ow 0 - 2 2 kg/ s
Feed-water control based on feed-flow measurement only
= 0 % open = O k g / s
= 1 0 0 % ope n = 2 2 kg/ s
Nat ural l y, scaling factors of t he t r ansmi t t er and t he valve must be
t aken into account . I f t he r ange of t he flow t r ansmi t t er is di fferent f r om
t he valve' s flow-control range, t he cont rol l er gai n will need to be adj ust ed
accordi ngl y, and in pract i cal systems this is always necessary.
I n or der to provide an adequat e operat i onal mar gi n of confidence, t he
range of t he cont rol valve is always designed to be gr eat er t han t he flow
range of t he boiler. For exampl e, in a boiler pr oduci ng 20 kg/ s of steam,
t he valve may be sized to deliver 22 kg/ s of wat er when it is fully open. I n
this exampl e, wi t h a l i near valve characteristic, an openi ng of approxi-
mat el y 91% will be needed to pass a flow of 20 kg/s.
I n this case, i f t he steam-flow t r ansmi t t er produces an out put of 100%
at 20 kg/ s flow, t he cont rol l er gai n must be such t hat a measur ed val ue of
100% produces an out put of 91%. This is a proport i onal band of 110 (i.e. a
gai n of 20/22) and i f this gai n is assigned to t he cont rol l er t he feed flow will
mat ch t he st eam flow over t he ent i re r ange of boiler l oad (assuming t hat
t he valve charact eri st i c is linear, t hat t he flow t r ansmi t t er out put is 4 mA
at zero flow, and t hat zero flow of wat er occurs with a valve signal of
0%).
The pr obl em wi t h this system is t hat it onl y mat ches t he steam- and
feed-flow rates. If, at t he outset, t he dr um level is below t he desired value,
t hat is wher e it will stay, because i f everyt hi ng is set up correct l y t he feed
into t he boiler will always mat ch t he st eam flowing out of it, and t here is
no mechani sm for i nt roduci ng t he small surfeit of feed over steam, or t he
slight deficit, t hat is needed to correct the drum-l evel error.
It is i mpor t ant to consi der t he pract i cal real i t y of what woul d happen
i f things were not to be set up correctly. I n this situation, i f t here is a small
setting er r or in t he cont rol l er gain, or i f t he feed valve passes mor e or less
wat er t han it should at t he given opening, or i f t he st eam flow t r ansmi t t er
is slightly out of Calibration, t he dr um level will i nt egrat e up or down at a
rat e det er mi ned by t he scale of t he error, and not hi ng will correct for this
undesi rabl e state of affairs.
I n ot her words, t he system cannot correct t he dr um level i f this
par amet er deviates f r om t he desired val ue ei t her because of an initial er r or
or because of small errors in t he steam-flow measur ement or nonlinearities
bet ween t he valve demand and t he act ual flow t hr ough it. I n t he exampl e
given above, t he exact gai n r equi r ed is 0.909 09 . . . . Therefore, i f t he con-
troller gai n were to be set to 0.91 as suggested above, t he feed-wat er flow
woul d be slightly gr eat er t han t he st eam flow, and t he dr um level will
gr adual l y increase.
To count er these effects it is necessary to add a feedback el ement , con-
sisting of anot her cont rol l er whi ch will act to correct for any mi smat ch
S t e a m f l ow Dr um l e v e l
LC
Fe e d- wa t e r r egul at i ng v a l v e
Figure 6.2 Basic two-element feed-water control system
bet ween the act ual and desired dr um levels. Fi gure 6.2 shows one vari et y
of such a ' t wo-el ement ' system.
In such a system, because the dr um vol ume and the st eam and feed
flows form an i nt egrat i ng system, wi t h the dr um level i nt egrat i ng any
st eam-fl ow/ feed-fl ow mi smat ch, it is unnecessary to empl oy an addi t i onal
i nt egrat i on function in the controller. Therefore the drum-l evel cont rol l er
shoul d be of the pr opor t i onal - onl y variety.
The correct gai n for this cont rol l er can be det er mi ned from a
knowl edge of the swell and shrinkage effects wi t hi n the boiler. I f these are
not known t hey can be det er mi ned by test. A suitable test woul d be to
change the st eam flow as rapi dl y as possible by, say, 10% of the maxi mum
evapor at i on rat e of the boiler, while keepi ng the feed flow in step wi t h the
st eam flow. (This can be achi eved by hooki ng a feed-flow signal into the
system while t empor ar i l y disabling the drum-l evel controller).
To see how the i nformat i on on the boiler' s swell charact eri st i c can be
used to help with cont rol l er tuning, let us exami ne a t wo-el ement system
where t he range of t he steam-flow t ransmi t t er is r anged as above ( 0-
20 kg/s), and t he feed valve is agai n sized to deliver 22 kg/ s when it is
100% open. Assume t hat t he drum-l evel t r ansmi t t er is r anged to pr oduce
4 mA when t he wat er level has dr opped below t he setpoint by 250 mm,
and t hat it is 20 mA when t he level is 250 mm above t he setpoint (i.e. a
range of 500 mm) . Finally, assume t hat a test as described above has det er-
mi ned t hat t he swell resulting from a sudden 10% change of st eam flow
raises t he dr um level by 80 mm.
I f t he drum-l evel cont rol l er is to exact l y count er act t he effect of swell,
it must pr oduce an out put t hat cancels out t he step change in t he st eam
flow, whi ch was 10%. The cont rol l er out put must t herefore change by
10% when t he i nput error changes by 16%, whi ch means t hat t he gain
must be 0.625 ( 10 + 16).
When t he steam-flow and pressure changes have settled out and t he
wat er level has r et ur ned to t he setpoint, t he level-controller out put will
agai n become zero. The valve openi ng will t hen revert to t racki ng slow
changes in t he st eam flow, as described earlier.
This analysis depends on t he swell effect being const ant over t he boiler
load range, whi ch may or ma y not be true, but it provides a pract i cal
met hod of t uni ng this t ype of system, and will pr oduce a fairly good perfor-
mance over a wide range of conditions. Theoret i cal l y, bet t er results coul d
be obt ai ned by car r yi ng out tests to det er mi ne the swell effect at various
15oints in the load r ange and i nt roduci ng a nonl i near funct i on wi t hi n t he
level cont rol l er to compensat e for t he differences across the range. But this
is r at her compl i cat ed for what is essentially a simple system and in any
case per f or mance is likely to be limited by the ot her serious deficiencies
wi t hi n t he system, whi ch we shall exami ne in t he next section, whi ch
discusses a mor e comprehensi ve system, known as t hree-el ement control.
6.2.2 Three-element f eed-wat er control
Thr oughout t he above analysis, reference has been made to the feed-
wat er valve charact eri st i c being linear and t he valve being sized to
pr oduce a fixed flow when it is 100% open. However, the flow t hr ough a
valve depends bot h on its openi ng and on t he pressure dr op across it. I n a
feed-wat er system, the pressure dr op across t he valve varies from instant to
instant, and t he flow t hr ough it at any given openi ng will t herefore vary.
For reasons given earlier, in a simple t wo-el ement system based on dr um
level, the inclusion of an i nt egrat i on el ement in t he level cont rol l er is unde-
sirable. Ther ef or e t he varyi ng flow results in the level cont rol becomi ng
offset, to restore the steam-flow/feed-flow balance. This offset is undesir-
abl e, si nce i t needl essl y er odes t he saf et y ma r g i n p r o v i d e d by t he pr es ence
of t he d r u m.
On e me t h o d of c o r r e c t i n g for t he e r r or p r o d u c e d by t he f eed val ve is
t he a d d i t i o n o f a t h i r d e l e me n t t o t he s y s t e m - - a me a s u r e me n t o f feed-
wa t e r flow.
Th e r e ar e var i ous ways o f i mp l e me n t i n g such a syst em, o n e o f wh i c h is
s hown i n Fi gur e 6.3. He r e , t he o u t p u t o f t he d r u m- l e v e l c ont r ol l e r is
t r i mme d by a si gnal r e pr e s e nt i ng t he di f f er ence b e t we e n t he f eed- f l ow a n d
s t eam- f l ow si gnal s. A ga i n bl ock (4) is i n t r o d u c e d t o c o mp e n s a t e f or a n y
di f f er ence b e t we e n t he r a nge s of t he t wo t r a ns mi t t e r s . I n mo s t cases t he
s t eam- f l ow a n d feed-fl ow si gnal s wi l l cancel out , a n d t he d r u m- l e v e l con-
t r ol l er wi l l be mo d u l a t i n g t he f eed fl ow t o keep t he l evel at t he s et poi nt . I n
t hi s case, i t is r e a s ona bl e t o a p p l y a n i nt e gr a l t e r m i n t hi s cont r ol l er , as
s hown.
I n a n o t h e r i mp l e me n t a t i o n of t hi s f a mi l i a r s ys t em s hown i n Fi gur e 6.4,
a ' cas cade c ont r ol ' t e c h n i q u e is appl i ed. T h e dr um- l e ve l c ont r ol l e r ( i t em 5,
Feed flow Steam flow
I J 4
Drum level
LC
5
Figure 6.3 One type of three-element feed-water control system
proport i onal only), compares t he measur ed level signal wi t h a set val ue
and produces a bi pol ar out put proport i onal to any error. This trims a
modi fi ed steam-flow signal, whi ch is act i ng as t he desired value for a
closed-loop feed-wat er cont rol l er (7). As previously, a gain block (4)
adjusts for any range difference bet ween t he steam-flow and feed-flow
transmitters.
These ar e not t he onl y ways of i mpl ement i ng t hree-el ement control.
Several vari ant s of t he system are in common use, each wi t h its own advan-
tages and disadvantages. However, each system has one fact or in common,
t he use of steam-flow, feed-flow and drum-l evel measurement s. The appli-
cat i on of t he feed-flow measur i ng el ement compensat es for any variations
in feed-flow, whet her these are due to t he pump characteristics or ot her
factors, and t he t hr ee- el ement system is t herefore r ecommended wher ever
accur acy of cont rol is requi red.
Feed flow Steam flow Drum level
~
3
LC
5
I
Figure 6.4 Alternative (cascade) three-element feed-water control system
As mi ght be expect ed, a t hree-el ement system is mor e expensive t han a
single-element or t wo-el ement system, but it is al ways wort hwhi l e to
exami ne the real cost differences before deciding on whi ch system to use.
For exampl e, the use of a steam-flow signal may not, in fact, add any addi-
tional cost, because this measur ement may be needed elsewhere, e.g. in the
combust i on cont rol system.
However, the presence of a feed-flow measur ement in the system does
necessarily add a significant cost burden. The t ransmi t t er itself is a
compl ex and not inexpensive i t em and it may also requi re the provision of
a pr i mar y el ement (such as a Venturi, flow nozzle or orifice plate) whi ch
will add furt her cost and requi re the provision of adequat e lengths of
straight pi pewor k upst r eam and downst r eam of the device.
The capi t al cost of the system needs to be carefully considered against
t he backgr ound of a det ai l ed knowl edge of the operat i onal r+gime of the
plant. All too often, insufficient rational t hought is appl i ed to these factors.
The use of a small dr um in the boiler offers a not insignificant reduct i on in
the cost of a pl ant , but its use also reduces the operat i onal safety margi ns
and t herefore requires increased accur acy of feed-wat er control. Yet, t he
cost-cutting exercises t hat lead to the use of a small dr um are very often
ext ended to t he cont rol system as well, furt her reduci ng the margi n for
error. I f safety margi ns are not to be dangerousl y eroded, careful consid-
erat i on must be appl i ed to the det er mi nat i on of pract i cal design rules for
feed-wat er cont rol systems. The cont rol implications shoul d be given j ust
as much consi derat i on as the process or mechani cal aspects.
6.3 Me a s ur i ng and di s pl ayi ng t he d r u m l evel
From the above, it is appar ent t hat the pr i mar y obj ect i ve of the feed-
wat er cont rol system of a dr um- t ype boiler or HRS G is to mai nt ai n t he
dr um level at the correct value. We shall now look at how this par amet er is
measured. It is an area where the probl ems are unexpect edl y complex.
Figure 6.5 shows one met hod of measuri ng the dr um level. Thi s
connects the differential-pressure t ransmi t t er directly to the dr um, via
isolating valves. Not e the ' const ant -head' reservoir connect ed to the upper
t appi ng point. Because the impulse pi pework to the t ransmi t t er is outside
the heat ed zone of the boiler, any st eam wi t hi n it will t end to condense,
and the pressure appl i ed to the HP port of the t ransmi t t er will t herefore be
the st eam and wat er pressure plus the pressure due to the wei ght of this
condensate.
The l at t er will depend on the vol ume of condensat e t hat has collected,
and this will be t i me-dependent . The pi pe will be full of st eam after the
S t e a m
Densi ty o f
s t e a m = ps
H
J i i i ! i i i i Y - . . . . .
~ Densi ty o f
w a t e r = pw
C o n s t a n t - h e a d
r e s e r v o i r
D e n s i t y o f h3
c o n d e n s a t e i n I
r e f e r e n c e l e g = Pr
L e v e l /
tra n s m itte r
Figure 6. 5 Measuri ng the drum level without a wat er column
t r ansmi t t er impulse lines have been ' bl own down' on start-up, but it will
aft erwards start to fill with condensate. The const ant -head vessel is left
del i berat el y unl agged, so t hat t he st eam in it condenses, mai nt ai ni ng t he
pipe to t he t r ansmi t t er full of condensate.
The differential-pressure appeari ng at the t ransmi t t er ports is a
funct i on of several variables: t he wat er level in t he dr um, t he densities of
this wat er and t he st eam above it and t he densi t y of the wat er in t he pipes
to t he t ransmi t t er. I n addition, the deri vat i on of a level signal needs to t ake
into account the densi t y of t he fluid used dur i ng t he initial cal i brat i on of
t he t ransmi t t er. Finally, it is also necessary to recognise t he r equi r ement
t hat t he lowest dr um- wat er level at whi ch t he system will be allowed to
operat e will be at some poi nt above t he level of t he lower t appi ng point.
The differential pressure at the t r ansmi t t er is defined by*
DP = (I/ p, ) x [(H x p,) - [(L x Pw) + (H - L) x p,)]]
where
DP = the differential pressure at the t ransmi t t er (at 20 C)
H = t he distance bet ween the bottom and top tappings
L = the height of wat er in the dr um above the bot t om tapping
Pr = the density of wat er used for calibrating the t ransmi t t er (at 20 C)
P8 = the density of the wat er in the dr um at the operating pressure
Pw = the density of the steam in the dr um at the operating pressure
It should be not ed t hat the differential pressure at t he t ransmi t t er will
be highest when t he separat i on bet ween t he wat er level and t he top
t appi ng is the greatest i.e. at low levels. I n ot her words, a 4 - 2 0 mA trans-
mi t t er will pr oduce 4 mA at the highest level and 20 mA at t he lowest.
This appar ent l y reversed 'sense' must be cor r ect ed by t he DCS before t he
densi t y cal cul at i on is applied.
Since t he densities of t he st eam and wat er in t he dr um will bot h
depend on t he conditions t hat exist in t he dr um, t he pressure needs to be
t aken into account when cal cul at i ng t he act ual level based on the differen-
tial head pr oduced at t he t ransmi t t er. At one time, various pr opr i et ar y
devices were available for performi ng t he requi red calculation, but it is
most economi cal , these days, to perform t he cal cul at i on in t he DCS, based
on t he differential-pressure and pressure signals. I n this case t he above
equat i on will enable t he level to be cal cul at ed for any combi nat i on of
signals from t he two transmitters.
However, an i mpor t ant point to bear in mi nd when using t he DCS to
per f or m t he pressure compensat i on in this way is t hat t he correct ed signal
will be available to t he oper at or only while the D C S is operational. I f a maj or
failure should occur in t he comput er system, it is i mpor t ant t hat t he
oper at or can still be able to moni t or t he dr um level by ot her means, and
these must be compensat ed in a similar way to t he above, so t hat the indi-
cat i on is relatively unaffect ed by pressure.
Figure 6.6 shows a pr opr i et ar y system whi ch generat es a drum-l evel
signal for display locally and in t he cont rol room. This t echnol ogy is based
on t he fact t hat t he conduct i vi t y of wat er is different from t hat of steam,
* Courtesy of Bristol-Babcock Ltd, Kidderminster, Worcestershire, UK
Feed-water control and instrumentation l 15
(a)
Water column
electrodes
Hydrsstep electronics
ountad at drum level
Control room
di spl ay
t r i p signals 4-20mA
signal
(b)
Even i nput board:
power supply,
electrode inputs, Odd i nput board:
discrimination, power supply,
validation, electrode inputs,
fault detection discrimination,
validation
- ~ fauff detection
Power supply B " .
E v e n e l e c t r o d e s =- - ~ ~ L
Even relay board: ~ ~
f ai l safe output ~ ~ ~J~- Odd electrodes
[ L M~ I K,3CI-- Power supply A
Odd relay board:
Display board: I ~ f ai l safe output
display decode, To remote di spl ay
opto fault output,
display drivers
(c)
For opt i mum performance, and t o
minimize density error, Hydrastep
wat er columns should be installed
as shown here.
For columns longer than l m. (3' 3"),
i t may be necessary t o insulate the
column itself.
=nsity error
2 m ( 6 . 5 " ) max,
Figure 6.6 The 'Hydrastep' level-indicating system
$olotron Ltd. Reproduced by permission
and a series of electrodes mount ed in a wat er col umn at t ached to the
dr um uses this fact to det ect the interface bet ween wat er and steam. (The
use of an ext ernal wat er col umn is necessary because of the number of
penet rat i ons t hat woul d ot herwi se be necessary in the boi l er drum. ) The
det ect i on circuit is di vi ded into two groups of ' even' and ' odd' electrodes,
so t hat failure of a single dri ve circuit cannot disable the entire system.
Al t hough such a system will be affected by the difference in the
densities of the wat er in the dr um and the col umn, careful design of the
installation will minimise any errors (Figure 6.6c).
At present, al t hough such devices provi de an excellent i ndi cat i on of
the dr um level, t hey are not suitable for control, because of the t ransi ent
di st urbance t hat occurs as the level moves from the position of one
el ect rode to another. These step changes can pr oduce unpredi ct abl e effects
in the cont rol loop.
6.3.1 Usi ng an external water column
Al t hough the met hod of connect i on shown in Fi gure 6.5 is viable, it has
the di sadvant age of bei ng sensitive to errors duri ng sudden reduct i ons in
the boiler pressure, caused by the condensat e ' flashing o f F- - b o i l i n g as the
t emper at ur e of the fluid suddenl y finds itself above the sat urat i on t empera-
ture.
An ar r angement t hat minimises this pr obl em is shown in Fi gure 6.7,
where an ext ernal wat er col umn is connect ed to the dr um, so t hat the level
of wat er in col umn is (theoretically) the same as the level wi t hi n the dr um.
The col umn uses a vol ume of st ored fluid which is l arger t han the vol ume
of condensat e in the smal l -bore HP leg of Fi gure 6.5, and the system is
t herefore less vul nerabl e to flashing off.
As wi t h the det ect i on col umn shown in Fi gure 6.6, great care must be
t aken to avoi d errors caused by the t emper at ur e of the fluid in the
measuri ng system being very different from t hat of the st eam and wat er in
the drum. Thi s leads to a densi t y error, since the wat er col umn at the
gauge will bal ance wi t h t he level in the dr um, al t hough its l engt h is less
t han the di st ance bet ween the l ower t appi ng and the level of t he wat er
within the dr um.
Consi deri ng the dr um and the col umn of wat er in the col umn to be a
' U' tube, it will be seen t hat bal ance will occur when the wei ght of fluid in
the l eft -hand leg equals the wei ght of fluid in the ri ght -hand leg.
By keepi ng the t emper at ur e condi t i ons wi t hi n the col umn as close as
possible to those in the dr um, the densi t y error will be minimised. This is
done by ar r angi ng the pi pework so t hat fluid flows t hrough the wat er
col umn. Unl i ke the dr um itself and all the ot her pi pework, the wat er
. . . . . . . . . . .
\ ~ I ] Water /
i , / ,,w / |
( ,
Figure 6.7 Drum-level measurement with a column
col umn and the section of pipe connect i ng it to the t appi ng-poi nt
isolation valve (on t he right in Fi gure 6.6) ar e left unl agged so t hat t he
st eam condenses and t he impulse line remai ns full of condensate. A circula-
tion of fluid is t hen established t hr ough the col umn, with st eam leaving t he
dr um at t he upper t appi ng point, condensi ng in t he pipe and wat er
col umn, and r et ur ni ng to t he dr um via the lower t appi ng point. This flow
tends to mai nt ai n t he t emper at ur e wi t hi n t he wat er col umn at a val ue
whi ch is as close as possible to the condi t i on wi t hi n t he dr um. Nevertheless,
some t emper at ur e difference will still exist, and this will have t he effect of
increasing t he density of t he fluid in t he cooler parts of t he system.
6. 3. 2 Statutory requirements
I n ma n y count r i es, t her e is a l egal r e q u i r e me n t to pr ovi de s e pa r a t e
s ys t ems t o ma k e t he boi l er o p e r a t o r a wa r e of t he l evel o f t he wa t e r i n t he
d r u m. I n t he UK, Bri t i sh St a n d a r d BS 1113 : 1998" refers t o t hi s subj ect i n
sect i on 7.3, as follows:
Each st eam boiler, in which a low level, or into which a low flow rate, of wat er
coul d l ead to unsafe condi t i ons shal l have at l east t wo i ndependent and
suitable means of i ndi cat i ng the wat er level or flow. Each i ndi cat i ng device
shall be capabl e of bei ng isolated from the boiler and each device shall be a
wat er level gauge in which the wat er level can be observed except in the follow-
ing cases:
(a) . . . (refers to once-t hrough boilers)
(b) For boilers wi t h any safety valve set at or above 60 bar g, the use of two
i ndependent manomet r i c r emot e wat er level i ndi cat i ons shall be per-
mi t t ed in place of one of the wat er level gauges. In such cases these remot e
wat er level indicators shall have their own i ndependent connections to the
boiler.
(c) For boi l ers of less t han 145 kg/ h evapor at i ve capaci t y, one wat er level
gauge is sufficient.
(d) The use of alternative devices in place of wat er level gauges in which the
wat er level can be observed shall be permi t t ed, subject to agreement be-
tween the manuf act ur er and the Inspect i ng Authority. The design of the
devi ces shall combi ne appr opr i at e desi gn pr i nci pl es such as fai l -safe
modes, r edundancy, di ver si t y and sel f-di agnosi s in or der to pr ovi de
suitable and reliable indication.
The wat er level gauge in whi ch t he wat er level can be obser ved shal l be
mount ed so t hat the lowest wat er level t hat can be observed is at least 50 mm
above the lowest wat er level at which there will be no danger of overheat i ng
any par t of the boiler, when in operat i on at t hat level. Wher e this is not practic-
able, the wat er level gauges shall be sited by agr eement wi t h the I nspect i ng
Aut hori t y in positions t hat have been found by experience to indicate satisfac-
torily t hat the wat er content is sufficient for safety under all service conditions.
At least one wat er level gauge with its isolating valves or cocks shall be con-
nected directly to the boiler, and other t han a drain, no device shall be fitted
to the gauge t hat coul d cause i ncorrect i ndi cat i on of the wat er level in the
gauge.
Fi gur e 6.8 is a n a t t e mp t t o i l l ust r at e s ome o f t he a bove r e qui r e me nt s
pi ct or i al l y. Not e t ha t t he a i m of t hi s d i a g r a m is onl y t o i l l ust r at e t he
r e qui r e me nt s a nd t o pr ovi de ge ne r a l gui dance. I t is i mp o r t a n t t ha t a
* Extracts from BS 1113 : 1998 are reproduced with the permission of BSI under licence no. PD\19981883.
Complete editions of the standards can be obtained from BSI Customer Services, 389 Chiswick High
Road, London,W4 4AL, UK.
A 'water level gauge in which the
level can be observed ...'.
Note that at least two such
gauges ere required on boilers of
145 kg/h (0.04 kg/s) or above.
i
m . m . m , n , B , , ~ , m
i I i I
' For boi l er s wi th any s af et y v al v e set at or a b o v e 60 bar g, t he
use o f t wo i ndependent manomet r i c r emot e wat er l evel
i ndi cat i ons shal l be permi t t ed i n pl ace of one of the water l evel
gauges. In such c as es t hes e r emot e water level i ndi cat or s
shal l hav e t hei r own i ndependent connect i ons t o t he boi l er . '
' . . . . The use of al t emat i ve devi ces i n pl ace of wat er l evel gauges i n whi ch the water l evel can be
obser ved shal l be permi t t ed, subj ect t o agr eement bet ween t he manuf act ur er and t he I nspect i ng
Authori ty. The desi gn of the devices shal l combi ne appr opr i at e desi gn pr i nci pl es such as f ai l - sef e
modes, r edundancy, di ver si t y and sel f - di agnosi s i n or der t o pr ovi de sui t abl e and r el i abl e i ndi cat i on. '
Figure 6.8 Permitted alternatives f or water-level measurement on drum-type boilers and
HRSGs
Based on BS 1113 : 1998
specific installation is designed to meet the standards required at the
actual point of use, since the standards set by authorities in the relevant
country, or by the insurers, may differ substantially from those indicated
by this diagram.
The above Standard also refers to low-level alarms, stating t hat
Every steam boiler shall be provided with a low water level or low water flow
audible alarm device.
Water level alarms, whether of low water or high and low water type, shall be
so fitted that the alarm is actuated while the level is still visible or indicated in
the water level gauges.
With reference to the size of the connecting pipework BS 1113 : 1998
states that:
Where a water level gauge, safety control or alarm device is connected to the
boiler by pipes, the bore of such pipes shall not be less than 25 mm.
(The standard allows the ends of the pipes local to the fittings to be
reduced to not less t han 20 mm bore for water level gauges and to 25 mm
bore for separate safety control and alarm devices.)
In this respect the Standard differs from others, but on one point there
is agreement, there is a need to use comparatively large-bore pipework for
instrument connections. The use of such pipes increases the cost of the
installation, but t hei r use is i mpor t ant from t he vi ewpoi nt of safety, since
small-bore pi pework can become obstructed (for exampl e by sludge) and
an obst ruct ed impulse pipe can lead to an i nst rument provi di ng seriously
i ncorrect i nformat i on on t he par amet er bei ng measured. (The use of ver y
small-bore pipes can also cause a readi ng to be sluggish because of t he t i me
t aken for pressure changes to affect the i nst rument , t hough this is less of a
probl em wi t h today' s i nst rument s whi ch displace very small volumes for
full-scale operation. )
6 . 3 . 3 Di s c r e p a n c i e s b e t we e n d r u m- l e v e l i n d i c a t i o n s
It sometimes happens t hat various i nst rument s connect ed to t he same
boiler dr um display level measurement s t hat are significantly di fferent
from each other. Since it is unlikely t hat t he act ual dr um is anyt hi ng but
hori zont al (except for installations on submari nes duri ng diving opera-
tions) such discrepancies must be due to some error or other. The following
list summari ses the factors t hat can cause errors:
Densi t y errors: Differences bet ween installations can cause one i nst rument
to be mor e affected by density factors t han anot her. (One possibility is
that, i nadvert ent l y, l aggi ng has been applied to one of the condensat i on
reservoirs).
Turbulence: The surface of t he boiling wat er inside t he dr um is anyt hi ng
but still. It has been known for ' st andi ng waves' to exist ar ound t he
downcomers, affecting some measur ement points mor e t han others.
Flashing-off'. Differences in t he geomet r y of t he measuri ng systems can
cause some measur ement s to be mor e affected t han others by flashing-
of f dur i ng pressure changes.
Calibration: It is vital t hat all t ransmi t t ers are carefully and accurat el y
cal i brat ed, and t hat any density compensat i on is correct l y set up.
Installation: As stated earlier, errors or sluggish response can be t he
result of part i al or compl et e plugging of impulse lines, or i mperfect
bl ow-down operations.
6 . 3 . 4 S t e a m e x t r a c t i o n
In bot h t he t wo-el ement and t hree-el ement systems, an assumpt i on made
in t he above examples is t hat the st eam out put by t he boiler event ual l y
ret urns to t he inlet in t he form of water. This is not t rue where significant
losses occur or i f any st eam is abst ract ed for applications such as soot-
blowing. Her e t he st eam is effectively lost, and i f a soot-blower abst ract i on
is made at t he dr um t he amount being used will not be i ncl uded in t he flow
measur ement . This will result in t he dr um level being offset from t he
desired val ue since the flow of feed wat er into t he system should be equal to
t he st eam t aken by the l oad (plant or turbine) plus t he st eam used for soot-
blowing. The effects of such operat i onal factors will be part i cul arl y signifi-
cant where t he design of the cont rol system has been based on the use of a
proport i onal -onl y controller. In such cases it may be necessary to add a
small degree of i nt egral act i on to t he controller, al t hough this should be
rest rai ned since t he addi t i onal i nt egral act i on can affect t he stability of t he
system.
6.4 The me c h a n i s ms us e d f or f e e d- wat e r cont rol
I n this analysis we have looked at t he principles of control, and seen t hat
because of various problems, a vari et y of cont rol met hods has evolved, t he
selection of whi ch depends on a vari et y of engi neeri ng and economi c con-
siderations rel at i ng to each application. I n t he discussion some reference
had to be made to t he mechani sm for cont rol l i ng t he flow and for simpli-
city it was assumed t hat this was by means of a fami l i ar device, a valve.
Now we shall look at t he nat ur e of valves in great er dept h, and t hen we
shall exami ne ot her met hods of cont rol l i ng t he flow.
6.4.1 Valves
What follows is mer el y a pract i cal overvi ew of valve designs in general. It
is not i nt ended to be a deep analysis of what is in i t sel fa specialised subject.
I f mor e det ai l ed i nformat i on is needed, it can be obt ai ned from t he many
textbooks on valve design or from the publications pr oduced by various
valve manufact urers. (See f ur t her r eadi ng section of this book for i nforma-
tion on t hree such books.)
A cont rol valve consists of many component s whi ch may conveni ent l y
be considered as falling into one of two groups: t he valve body and t he
act uat or. The former is t he par t t hr ough whi ch the wat er flows and this
flow is controlled by adjusting the resistance offered to t he water. This is
done by movi ng t he position of a plug in rel at i on to its seat. The position of
the plug is cont rol l ed by an act uat or whi ch acts via t he stem.
Fi gure 6.9 shows a small-bore feed-wat er cont rol valve body wi t h a
cont our ed t ri m (the ' t ri m' being t he par t of t he valve whi ch is in flowing
cont act wi t h the water). The cont our det ermi nes the relationship bet ween
the position of t he plug and t he flow of wat er past it. The t ype of t r i m will
be di ct at ed by t he application, such as t he need to mi ni mi se acoustic noise
or cavitation, t he rangeabi l i t y needed etc. I n addi t i on t he t r i m design will
det er mi ne t he valve characteristic, whi ch is t he curve rel at i ng t he stem
position to t he rat e of flow of wat er t hr ough t he valve. This is an i mpor t ant
feature, since t he charact eri st i c det ermi nes t he gai n of t he valve system,
whi ch forms par t of t he overall loop gain.
Stem
Seat
Figure 6.9
/
Valve body
A typical f eed- wat er control valve body
Hopkinsons Ltd. Reproduced by permission
For a given opening, t he flow t hrough the valve will be det ermi ned by
t he delivery pressure of t he feed pump and the resistance t hat t he boiler
pipework offers to the flow. To simplify the task of selecting the correct
valve size and characteristic, it is necessary to relate everything to a
definable set of conditions. This is achieved by det ermi ni ng what the flow
t hr ough the valve woul d be i f a fixed differential pressure were to be main-
t ai ned across it. This is t ermed t he inherent characteristic of t he valve.
Once the valve is operat i ng on the act ual plant, the position/flow rela-
tionship achieved in practice will not mat ch the i nherent characteristic,
because in the real world the inlet pressure and system resistance will vary,
produci ng a pressure dr op which is different from the value t hat was used
to define the i nherent characteristic. The pressure/flow relationship
achieved in actual operat i on is called the i ns t al l e d characteristic.
As stated earlier, the gain of t he valve is initially defined by the
i nherent characteristic, three types of which are commonl y available, as
shown in Figure 6.10. The operat i on of these different characteristics is
now examined.
0
0
E
I o o i I I I I I I I I i
. . . . -~ . . . . . i--.-.-i . . . . . ~- . . . . i . . . . -i . . . . . i- . . . . i . . . . 4
/ i i i i i i ! ' . _ . . . . ~ . ~ .
I i ~ i i i i _ . , . - + ' ~ . i i ,
- r . . . . t . . . . . ~ - - - - ~ . . . . . i - . . . . i - ~ , - t - . . . . ! - - ~
i i i i i i i i i
. . . . . L . ,
! Q u i c k
! o p e n i n ( .
6 0 !
I
J
. . . . . . J . . . . . . . L . - - / -
i i
/
4 0 . . . . . . i . _ !
! ,
il ~ i . . . .
' ' ~ i i i
I I
i . . . . 4 . - - - - ~ , ' . - . . . .
' ' i i
. . . . . ~ . . . . 1- . . . . i f - l - - r
.. . . . . .L . . . . . . . . .L . . . . . t
I I I
/ i ' ' i i i
L i n e a r . . . . . . . - t . . . . . t -
I _ . . , ~ E q u a l . . -
. . . . . - - ' ~ ' - ' 1 / I T ' " p e r c e n t a g e
J - . ~ , . . . . . , i - . . . . i . . . . - i . . . . .
i " ! ! J i i
i i i i i
~ - - - ~ . . . . . ~ . . . . i . . . . ~ . . . . . ~ -
i i ! ! !
i i i i i
2 0
~o
- . - I . . . . . 4 - . . . . ~ - .
- ' l - - ~ r . . . . i---
i
i
4 0
i i i i i
6 0 8 0 1 0 0
V a l v e o p e n i n g ( % o f r a t e d t r a v e l )
Figure 6.10 Inherent characteristics of valves
6. 4. 2 Qui ck- openi ng
Wi t h a qui ck- openi ng valve, t he fl ow r at e t hr ough t he val ve changes ver y
r api dl y at l ow openi ngs, wi t h a slope t hat is f ai r l y l i near. Once t he val ve
has opened about h a l f wa y t he r at e of change of fl ow di mi ni shes. Thi s t ype
of char act er i st i c is usual l y appl i ed onl y to shut - of f val ves.
6. 4. 3 Li near
Wh e n a val ve has t he l i near char act er i st i c, t he fl ow r at e t hr ough i t at a ny
gi ven openi ng (in t er ms of per cent age of ma x i mu m flow) is di r ect l y equal
to t he val ve st em posi t i on (as a per cent age of its full t ravel ). Wi t h t hi s t ype
of char act er i st i c, t he gai n of t he val ve syst em is cons t ant for all openi ngs.
However , t he fl ow t hr ough t he val ve at a ny gi ven openi ng depends on t he
pr essur e- dr op across it a nd t he l i near char act er i st i c appl i es only i f the
pressure drop across the valve is constant f or each opening, a condi t i on t hat ma y not
exist i n pr act i ce unless special at t ent i on is pai d t o achi evi ng it.
As a general rule, a l i near charact eri st i c is preferred for feed-wat er
cont rol applications, since it simplifies t uni ng of t he loop and enables good
per f or mance to be achi eved over the widest possible range of flows.
Wher e opt i mum per f or mance and efficiency is requi red t he addi t i onal
cost of provi di ng a const ant pressure drop across the valve should be con-
sidered. Such a solution will add cost, but this will be offset by i mproved
cont rol per f or mance and pl ant life and by the savings achi eved by not
r unni ng t he feed pump or pumps at a hi gher speed t han necessary at
r educed loads.
6.4.4 Equal percentage
The t hi rd charact eri st i c is called equal percent age. Here, for all positions
of t he stem, t he flow change achieved by movi ng t he stem by a given
amount is a const ant proport i on of the previous flow. What this means is
t hat a given stem movement will change t he flow by t he same rat i o of the
previous flow, at any poi nt in the valve travel. Therefore, t he l arger the
opening, t he great er will be t he change of flow pr oduced by a given stem
movement .
This is shown by t he curve in Figure 6.10, where movi ng t he stem from
20% of full travel to 30% of full travel changes the flow from 5% to 7.5%,
while movi ng t he stem from 80% of full travel to 90% of full travel
increases t he flow from 50% to 75%. I n both cases, a stem movement of
10% of full travel results in a 50% increase of the previous flow.
Fr om this it is appar ent that, with this t ype of characteristic, the
change of flow for a given stem movement is smallest at low openings and
greatest at high openings. I n ot her words, the gai n is lower at low flows
t han it is at high flows. Equal -percent age valves are used where t he
mechani cal pl ant design is such t hat t here is onl y a small pressure dr op
available for t he valve, or where t he pressure dr op across the valve is likely
to fluctuate over a wide range.
6. 4. 5 The valve s i z i ng coefficient
Feed valves are designed to pass a flow t hat corresponds to t he ma xi mum
flow r equi r ement of the boiler plus a safety margi n, and t he capaci t y of the
valve is rel at ed to a fact or known as the valve sizing coefficient. The fact or
used widely across the world is based on US units and relates the capaci t y
of t he valve to t he pressure dr op by t he following formul a:
O. = C~v/(aP/a)
where
Cv = sizing coefficient
Q = capacity of valve in US gallons per minute (1 US gallon = 3.785 litres)
Ap __-pressure di fferent i al across valve in pounds per square i nch (psi).
(1 psi = 0.069 bar)
G = specific gravity of the flowing fluid
The sizing coefficient is det er mi ned by exper i ment for each t ype and
size of valve using wat er as t he test fluid. It is equal to t he vol ume of wat er
(in US gallons) t hat will flow t hr ough the wi de-open valve in one mi nut e
when t he pressure dr op across t he valve is 1 psi. The Eur opean equi val ent
of t he sizing coefficient is known as Kv. To convert bet ween these units
mul t i pl y t he Kv figure by 1.66 to obt ai n t he Cv.
Gi ven t he pressure, t emper at ur e, flow and line-size characteristics for a
given feed-wat er application, valve manuf act ur er s will be able to provi de
det ai l ed gui dance on t he correct valve size for a given appl i cat i on and
from this will be able to predi ct t he pressure dr op across t he valve.
6.4.6 Fai l -s@ operation
In t he course of designing a feed-wat er cont rol system, anot her mat t er
t hat must be considered is t hat of selecting t he 'fail-safe' position of the
plug, t he state t hat will arise i f t he act uat or fails, or if t he c omma nd signal
to t he valve is lost. The act ual selection will depend on a range of factors,
but in det er mi ni ng t he safest opt i on it is i mpor t ant to consider t he effects
of t he flow on t he valve itself, since the applied pressure may t end to force
t he plug open or closed. Once again, valve manuf act ur er s will be able to
provide advice on this mat t er in relation to t he act ual installation bei ng
designed.
6.4.7 Selecting the valve size
The size of valve to use in a given appl i cat i on will be det er mi ned by
many factors, onl y one of whi ch is t he physical size of t he line in whi ch it is
fitted. Clearly, t he valve must be large enough to pass t he requi red flow
with ease. Oversi zed valves will be unnecessarily expensive and should be
avoided as unabl e to effectively control small flows.
For a valve to mai nt ai n any control over t he process t here must be
some pressure dr op across it, but i f the pressure dr op is too great a number
of undesirable events st art to occur.
Figure 6.11 shows in profile t he pressure al ong a short section of pipe
cont ai ni ng a restriction (such as a valve). Two curves ar e shown, one for a
valve where careful design produces st reaml i ned flow and mi ni mal
pressure loss, and anot her for a valve t hr ough whi ch t he flow is more,
126 P o we r - p l a n t control a n d i ns t r ument at i on
P 2 . -
I
i
,,, Ps
~
* " ' " " ~ ~ " " " ' ' " " - P4
s contracta
Fi gur e 6. 11 Pressure profile al ong a p i p e containing a restriction
produci ng a gr eat er pressure loss. I n each case, t he lowest pressure occurs
a short di st ance downst r eam from t he restriction (the Vena cont ract a) and
t he pressure rises past this poi nt al t hough it obviously never regains the
initial value. I f t he pressure at t he inlet (P~) is such t hat t he pressure along
this curve drops bel ow t he vapour pressure, t he fluid will st art to boil and
bubbles will form in it at t he poi nt where this occurs.
One of two things will occur as t he pressure rises agai n aft er t he restric-
tion. I f it rises above t he vapour pressure, t he bubbles will c ol l a ps e - - a
process known as cavitation, whi ch is accompani ed by acoustic noise, the
degree of whi ch depends on t he scale of t he cavitation. Apar t from being
undesi rabl e on envi r onment al grounds, this noise represents an expendi-
t ure of energy. Fur t her mor e, i f t he bubbles collapse close to met al surfaces,
t he localised ener gy release can damage t he metal. I n some cases this will
become appar ent as severe pitting of t he valve pl ug and cage, or it could
appear as pitting of t he pipe itself i f t he poi nt wher e the vapour pressure
t hreshol d is passed occurs some way down from t he valve. Hi gh recovery
valves are mor e likely to experi ence cavitation, because t he downst r eam
pressure is mor e likely to exceed t he vapour pressure.
If, on t he ot her hand, t he pressure does not rise above t he vapour
pressure t he bubbles will r emai n suspended in t he fluid, to be carri ed
downst r eam of t he restriction. This is known as ' flashing' , and it can result
in erosion damage to t he valve internals at t he poi nt of ma xi mum velocity
(usually at or near t he poi nt where t he pl ug seats against t he ring).
These bubbles reduce t he valve' s ability to pass fluid and event ual l y it
'chokes'. When this occurs no mor e flow can occur irrespective of how
much wat er pressure is applied at t he valve inlet. Fr om these considera-
tions it is appar ent t hat t he product i on of bubbles may cause noise and
damage to t he valve, and possibly t he pipework, and it will be correct l y
surmi sed t hat an opt i mum set of conditions will exist for a given design of
valve. I n ot her words, t he valve will wor k best at one pressure-drop
point.
To summari se, i f t he supply of wat er to a boiler is cont rol l ed by t hrot -
tling t he flow t hr ough a valve, this can cause erosion and noise and
al t hough cont rol can onl y be effected by mai nt ai ni ng some pressure dr op
across t he valve, this loss represents a loss of ener gy whi ch should be
r educed to t he mi ni mum.
6.5 Pumps
The wat er flowing t hr ough t he feed valve into t he boiler is delivered at
pressure by one or mor e feed pumps. These pr oduce a head of wat er whi ch
is rel at ed to t he flow t hr ough t he pump by a charact eri st i c t hat will be
similar to Figure 6.12, whi ch shows t hat al t hough the di scharge pressure
remai ns relatively const ant as t he flow rises from zero to (in this exampl e)
about 50%, beyond this value t he pressure tends to decay as t he flow
increases. From this it will be appar ent t hat t he feed valve has to pr oduce a
gr eat er pressure dr op at low loads t han at hi gh loads.
Fi gure 6.13 shows t he fl ow/ pressure charact eri st i c of a pump delivering
wat er into a fixed system. I n practice, as t he flow t hr ough t he system
increases t he resistance offered to it also increases, as shown by t he dot t ed
line in t he di agram.
At any flow, in or der to deliver wat er into t he boiler, t he pressure dr op
across t he control val ve will be t he difference bet ween t he pump del i very
pressure and t he system resistance. For much of t he flow range t he pressure
dr op will inevitably be great er or less t han t he opt i mum for t he valve
design. Thi s can be overcome by changi ng t he inlet pressure so t hat t he
| 28 Power-plant control and instrumentation
1 0 0
"6
o~
7 . ~
5 c _
Figure 6.12
! ! ~ - 1 - - ~ - - ! . . . . T . . . . . F
! ! ! ! ! ~ ! !
i i J I ! ! ! ~ , , L i
i i i i i m ' , , j
. . . . . . . . . . . . . , I - - ' - ] . . . . . . T . . . . ! . . . . G - i / - - - ! . . . . - r - - - - ~ - ,
i i I i i i i i
i i i i i i i i
i i i i i i i i
. . . . . . , . . . . . . r - - - - i . . . . . t . . . . i . . . . ! . . . . . . t . . . . i . . . . - t . . . . . r
I I I I
i i i i
k I I I
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - . . . . . i . . . . . . . . . . . . . . . i -
J i i I
I I I I
. . . . . ! . . . . . . . . . . . . . . . i . . . . . . . . . j . . . . . L . . . . . . . . . . . . . . . L
I I
2 0 4 0 6 0
I I
I I
I I
i i
I I
8 0 1 O 0
% o f r a t e d f l o w
Typical speed curve f or a f eed pump
1 2 ~
= =
~ . 1 0 0
o~
7 5
5 o .
, , , " , ~ - ~ - - ! . . . . v . . . . . r
! P u m p c h a r a c t e r i s t i c i I P r e s s u r e ~ i
I , , , I I d r o p a c r o s s ! !
. . . . . i " . . . . . i - ' - ' - i . . . . . i - . . . . i . . . . - ] V a l v e ! - - ' % - ; ' - T
i i i i i ~ _ _ - L . . - . . . ~ " " - J
- - i - - i - - i - - i - - ~ i " - P " " i - i i
i i i i / ~ i i i i ; i
. . . . . . . . . h - - - - ; ~ - - i . . . . 1 . . . . . 1 - . . . . I . . . . . . . . . . . l -
S y s t e m r e s i s t a n c e ' / " ! ! ! ! j !
i = i i i i i i
. . . . . r . . . . ! . . . . . . . . . . F - d . . . . . T . . . . . i . . . . d . . . . . . . . . . ! -
i ! ~ I I i I
i I i i I I I
. . . . . . . . . . L . . . . . . . . . L . . . . J . . . . J . . . . . L . . . . i . . . . . . . . . . L
I I I I I I I
I I I I I J I
I J i i i i I
I i I I I I i
I I I I I I I
2 0 4 0 6 0 8 0 1 0 0
% f l o w
Figure 6.13 Pump delivery into a system resistance
pressure dr op always remai ns at t he opt i mum value. This is achi eved by
t he use of variable-speed pumps.
6.5.1 Variable-speed pumps
Al t hough in many cases t he feed pumps operat e at a fixed speed, at t he
design stage consi derat i on should be given to t he option of using variable-
speed pumps (sometimes known as ' controllable-speed' pumps) because
these will enable t he feed valve to operat e at t he opt i mum pressure condi-
tions for all loads.
The characteristics of a variable-speed pump ar e shown in Figure 6.14.
The four curves show t he pressure/flow characteristics at four di fferent
operat i ng speeds ( A- D) , with A being t he slowest and D t he fastest speed.
From this set of curves it will be appreci at ed t hat by adjusting t he speed of
t he pump it will be possible to mai nt ai n a fixed differential pressure across
t he feed regul at i ng valve at all boiler loads. I n t he exampl e shown t he
pressure dr op across t he valve (6P) achieved by operat i ng t he pump at
speed A at 70% flow is mai nt ai ned at t he same value by operat i ng t he
pump at speed B at 80% flow, speed C at 90% flow etc., as i ndi cat ed by
t he lines wi t h arrows at each end.
Q .
25
. . . . . 1. . . . . ~ . . . . . T . . . . ! . . . . T . . . . . F . . . . ! . . . . 7 . . . . . . , r - - " 3 -
I I I I I I I I I I
! ! ! ! ! ! ~ ! !
. . . . T . . . . . F - n . . . . . . T . . . . 1 . . . . q . . . . T . . . . ! - " ~ . . . . F
I I I I I I ~ I I " ~ . I
I i ' , ; i i i ~ ' I " , , ~ i " ~ O
i i i i i i ~ m ' - - , - , , k ~ ' k , , , ~
. . . . . - r . . . . . ( - - - - - i . . . . . 1 - . . . . i . . . . ~ - ~ . . c
i i i i i i s P ~ ~ ~ i
- - i - - i - - i - - i . - - _ . - i - - - i - - - - ~ - i ~' I ' ~ ~i B
i J i i .,,'~ i i i i i ~ " , i A
. . . . . - ~ . . . . . ~ - - ~ . . . . , , T - - ! . . . . ~ . . . . . . i - . . . . ! . . . . - . i . . . . . . .
S y s t e m r e s i s t a n c e ! ! ! ! ! !
! I I I I I I
. . . . . . L . . . . . . . . . ! . . . . . L . _ . J . . . . . i . . . . J . . . . J . . . . . z . . . . . .
I I I I I I I I
I J I I I I I J
I I I I I 1 1 I
. . . . . z . . . . . . . . . J . . . . . i _ . . . . ! . . . . j . . . . . i . . . . ~ . . . . . . . . . .
I I I I I I I I
i W i I I I I I
I I I I t i I t
I I I I I I I I
I I I I I I I I
2 0 4 0 6 0 8 0
% f l o w
lO0
Figure 6.14 Variable-speed pump operation
The advant ages of using a variable-speed pump include:
I mpr ovement of efficiency because of reduced pressure loss.
Reduct i on of pumpi ng power.
Reduct i on of feed-valve wear due to erosion when operat i ng at low
~]OWS.
I mpr oved cont rol because t he valve operates at its designed pressure-
drop.
I mpr oved cont rol offered by the ability to operat e with const ant loop
gain.
Variable-speed pumps ar e mor e expensive t han fixed-speed ones, but t he
increase in capital cost is offset by t he revenue savings t hat will be gained,
part i cul arl y if t he boiler operat es at r educed t hroughput s for a significant
time over its life. Careful financial analysis will det er mi ne whet her the
savings do justify the addi t i onal capital cost, but these calculations must be
based on assumptions about t he operat i onal regime, and these may change
under t he influence of ext ernal factors whi ch are difficult to predi ct at t he
design stage.
Of course, t he pumps, valves and boiler pi pework ar e onl y par t of t he
overall system. Wher e spray at t emper at or s are used (see Chapt er 7), the
feed pumps must also be capabl e of delivering cool wat er to t he nozzles.
The design of at t emper at or s requires t he wat er to be at del i vered at a
pressure whi ch exceeds t he st eam pressure by a mi ni mum value. Bypassing
the pressure dr op across t he system as shown in Figure 6.15 allows suffi-
cient differential pressure to be mai nt ai ned at t he spray nozzle. But, i f a
variable-speed pump is used and r un down to a low del i very pressure,
t here is a risk t hat the requi red differential may not be available. I n such
situations, t he decision to use fixed- or variable-speed pumps will be
affected by t he need to mai nt ai n an adequat e wat er pressure at t he at t em-
porators.
Boiler
feed
pump
I "
Figure 6.15
Steam
drum
$1x, ml
Maintaining adequate spray-water differential pressure
6.6 De ae r at or cont rol
Strictly speaking, cont rol of t he deaer at or is not a funct i on of feed-wat er
cont rol whi ch is t he subject of this chapt er. However, as t he deaer at or is an
essential link in t he feed-wat er supply system it is appropri at e to consider
its cont rol systems here.
I n Fi gure 2.5 we saw how st eam admi t t ed to t he deaer at or rises
upwards past met al t rays over whi ch t he wat er is si mul t aneousl y cascadi ng
downwards. As t he wat er and st eam mix and become agitated, ent r ai ned
gases are released. The dissolved gases are vent ed to t he at mosphere
because t he vessel is pressurised by t he steam. The deaer at or is situated in
t he wat er circuit bet ween t he di scharge of t he condenser ext ract i on pump
and t he inlet of t he feed pumps, as shown in Figure 6.16.
It will be evi dent t hat t wo cont rol functions are requi red by t he
deaerat or: one to mai nt ai n t he st eam pressure at t he opt i mum value, t he
ot her to keep t he storage vessel r oughl y hal f full of wat er.
6.6.1 Steam pressure control
The pressure of t he st eam ent er i ng t he deaer at or is mai nt ai ned by a
simple cont rol l er whose measured-val ue signal is obt ai ned from a trans-
V
=
Feed-water
control valve
BoUer
pump
pump
Demrator
Condenser
p~q
- /
Make-up wider
control valve
. . . . . . . . . . . . . . . . . . . . . . . . !
Figure 6.16 Principle of deaerator level control system
mi t t er measuri ng the st eam pressure in the deaerat or. The set val ue of
the cont rol l er is nor mal l y fixed.
It has al r eady been expl ai ned t hat the steam suppl y may be obt ai ned
either from the boi l er or from an ext ract i on poi nt on the turbine. I f the
l at t er source is used, special consi derat i on has to be given to ensuring t hat
an event such as a t urbi ne trip does not depri ve the deaer at or of t he st eam
suppl y it needs for its operat i on. Thi s purpose is served by t aki ng a st eam
suppl y from the boi l er and passing it to the deaer at or via a pressure-
reduci ng and desuperheat i ng system ( PRDS) . Thi s st eam suppl y is
referred to as ' peggi ng steam' . Judi ci ous adj ust ment of the PRDS con-
troller setpoint will ensure t hat an adequat e st eam suppl y for the deaer at or
is obt ai ned at all times. However , to ensure rapi d response to a t ur bi ne
trip, a system of interlocks should be provi ded, so t hat the peggi ng st eam
suppl y is br ought into service i mmedi at el y on det ect i on of a trip.
6.6.2 Level control
The storage vessel provides a measure of reserve capaci t y for the plant.
To achieve this funct i on the level of wat er in it must be mai nt ai ned at
roughl y the mi dpoi nt . Thi s is achieved by means of a level cont rol l er
whose measured-val ue signal is obt ai ned from a differential-pressure trans-
mi t t er or from capaci t i ve probes whi ch woul d nor mal l y be connect ed to
tappings of an ext ernal wat er col umn which is in t urn connect ed to the t op
and bot t om of the deaer at or storage vessel. I f there were no losses in the
system, the amount of wat er woul d be const ant and the level in the
deaer at or st orage vessel woul d r emai n at the correct val ue set duri ng com-
missioning. However , losses are inevitable (for exampl e, due to leakages at
pump glands or duri ng soot -bl owi ng or bl owdown operations), and a
suppl y of t reat ed wat er must t herefore be made available. The deaer at or
level cont rol l er out put adjusts the openi ng of a valve t hat admi t s this
make- up wat er to t he condenser, as shown in Figure 6.16.
The make- up suppl y is convent i onal l y fed into the system at the
condenser. Fi gure 6.16 shows t hat i nt eract i on bet ween the level controllers
of the deaer at or and condenser is inevitable. The situation is made mor e
compl ex because the condenser ext ract i on pump has to be provi ded wi t h a
bypass ar r angement to mai nt ai n a mi ni mum flow t hr ough the pump at all
times.
In fact, t he conditions which cause the deaer at or level cont rol l er to call
for mor e wat er to be added to the system will also cause the condenser
level to fall, and so the t wo systems do not act in opposi t e senses.
Nevertheless t hey do interact, and care must be t aken to minimise the
instability t hat is likely to arise.
6. 6. 2. 1 Spi l l control
I n addi t i on to the level control system described above, a system must be
provi ded to dr ai n excess wat er from t he deaer at or storage vessel in t he
event of an oversuppl y of wat er and steam. This funct i on is nor mal l y
achi eved on l arge power-station boilers by using the out put of the level
cont rol l er to operat e t he make- up and spill valves in split range. A typical
ar r angement is shown in Tabl e 6. 1. Smal l er boilers will be limited to
havi ng a spill valve whose openi ng is nor mal l y cont rol l ed on an on/ of f
basis from a high-level al arm.
6.6.3 Integrated level control
The long dead t i me and long t i me const ant of the deaer at or level system
complicates t he design and commissioning of this cont rol loop. This
situation has been worsened by t he commer ci al pressure to operat e power-
station boilers at part i al load and t he i ncreased need for two-shifting
operations. ( The t er m ' two-shifting' refers to operations wher e t he pl ant
runs for two ei ght -hour periods (or shifts) in a worki ng day, and is shut
down for t he r emai ni ng ei ght -hour shift.) Ther e is also a dri ve to reduce
the consumpt i on of cont rol l abl e make- up water, al t hough this necessarily
increases wi t h t he amount of load changi ng to whi ch t he boiler is
subjected. A typical scenari o is t hat if the boiler experiences a large load
reduct i on (say 50%) t he deaer at or level woul d t end to rise because as t he
st eam flow from t he boiler reduces so must t he feed flow. Dur i ng this time,
the condensat e flowing to t he deaer at or woul d not be requi red by the feed
pump, causing t he level in the storage vessel to rise. The level controls
woul d react by dumpi ng t he excess wat er t hat is filling the deaerat or. I f
t here is a l oad increase to the original load, t he deaer at or level woul d dr op
due to t he increased flow from t he feed pump, and t he level controls woul d
t end to add make- up wat er to the condenser. This cycle of maki ng- up and
spilling can occur every t i me t he load changes.
Table 6.1 Operation of deaerator level-control valves
( Scottish Power plc. (Reproduced by
permission)
Controller output Valve duty Valve position
0- 25% Spill valve 100-0%
25-75% Deadband n/ a
75-100% Make-up valve 0-100%
The r equi r ement to reduce the wast ed make- up wat er has resulted in
t he use of advanced cont rol strategies for cont rol l i ng t he make- up and spill
on l arger boilers. For exampl e a fuzzy logic cont rol al gor i t hm embedded
in t he DCS software can be used. The inputs to t he cont rol l er are t he
st eam flow and t he deaer at or level, t he out put s are to t he make- up and
spill valves. The system uses rule-based logic to t ake decisions on maki ng-
up and spilling. The goal of t he scheme is to mi ni mi se make- up and spill
by ret ai ni ng wat er wi t hi n t he condensat e system until it is requi red.
6.7 Summary
Once t he combust i on process has occurred, t he feed wat er has boiled and
t he st eam has been generat ed, t he next r equi r ement is to ensure t hat t he
t emper at ur e of t he st eam t hat is delivered to t he t urbi ne or heat l oad is
mai nt ai ned at t he correct value. I n t he next chapt er we shall look at t he
cont rol and i nst r ument at i on systems t hat are empl oyed for this purpose.
Chapter 7
St eam- t emperat ure control
7.1 Why s t e a m- t e mpe r a t ur e cont rol i s ne e de d
The rat e at whi ch heat is t ransferred to t he fluid in t he tube banks of a
boiler or HRSG will depend on t he rat e of heat i nput from t he fuel or
exhaust from t he gas turbine. Thi s heat will be used to convert wat er to
st eam and t hen to increase t he t emper at ur e of t he st eam in t he superheat
stages. I n a boiler, t he t emper at ur e of t he st eam will also be affected by t he
pat t er n in whi ch t he burners are fired, since some banks of tubes pick up
heat by di rect r adi at i on from t he burners. I n bot h types of pl ant t he t em-
per at ur e of t he st eam will also be affected by t he flow of fluid wi t hi n t he
tubes, and by t he way in whi ch t he hot gases ci rcul at e wi t hi n t he boiler.
As t he st eam flow increases, t he t emper at ur e of the st eam in t he banks
of tubes t hat are di rect l y influenced by t he r adi ant heat of combust i on
starts to decrease as t he increasing flow of fluid takes away mor e of t he
heat t hat falls on t he metal. Ther ef or e t he st eam- t emper at ur e/ st eam- f l ow
profile for this bank of tubes shows a decline as t he st eam flow increases.
On t he ot her hand, t he t emper at ur e of t he st eam in t he banks of tubes
in t he convect i on passes tends to increase because of t he hi gher heat
t ransfer br ought about by t he i ncreased flow of gases, so t hat this t emper a-
t ur e/ f l ow profile shows a rise in t emper at ur e as t he flow increases. By
combi ni ng these two characteristics, t he one rising, the ot her falling, t he
boiler designer will ai m to achieve a fairly flat t emper at ur e/ f l ow char ac-
teristic over a wi de range of st eam flows.
No mat t er how successfully this t arget is at t ai ned, it cannot yield an
absolutely flat t emper at ur e/ f l ow characteristic. Wi t hout any addi t i onal
control, t he t emper at ur e of t he st eam leaving t he final superheat er of t he
boiler or HRSG woul d var y wi t h t he rat e of st eam flow, following what is
known as t he ' nat ur al charact eri st i c' of t he boiler. The shape of this
will depend on the part i cul ar design of plant, but in general, the t emper a-
t ure will rise to a peak as the l oad increases, after whi ch it will fall.
The st eam t urbi ne or the process pl ant t hat is to receive the st eam
usually requires the t emper at ur e to r emai n at a precise val ue over the
entire l oad range, and it is mai nl y for this reason t hat some dedi cat ed
means of regul at i ng the t emper at ur e must be provi ded. Since different
banks of tubes are affected in different ways by the radi at i on from the
burners and the flow of hot gases, an addi t i onal r equi r ement is to provi de
some means of adj ust i ng the t emper at ur e of the st eam within different
part s of the circuit, to prevent any one section from becomi ng over-
heat ed.
In t heory, the design of the pl ant should be t ar get ed on arrangi ng for
the nat ural characteristic to at t ai n the correct st eam t emper at ur e when
the rat e of st eam flow is t hat at which the boiler will nor mal l y operat e. I f
this is possible, it means t hat spray wat er is used onl y while the unit is
bei ng br ought up to l oad or when it operat es at off-design conditions. In
pract i ce this obj ect i ve can be at t ai ned onl y to a limited extent, because the
boiler' s nat ural charact eri st i c changes wi t h t i me due to factors such as
fouling of the met al surfaces, whi ch affects the heat transfer. In general, it
is common to oper at e wi t h cont i nuous spraying, whi ch has the advant age
of allowing the st eam t emper at ur e to be adj ust ed bot h upwar ds and
downwards. I f the requi red t emper at ur e were to be met solely by
empl oyi ng the nat ural charact eri st i c as described, it woul d not be possible
to pr oduce t emper at ur e increases.
Before looking at the types of st eam- t emper at ur e cont rol systems t hat
are applied, it will be useful to exami ne some of the mechani sms whi ch are
empl oyed to regul at e the t emper at ur e accordi ng to the cont rol l er' s
commands. Dependi ng on whet her or not the t emper at ur e of the st eam is
l owered to bel ow the sat urat i on poi nt the cont rol l i ng devices are known as
at t emperat ors or desuperheaters. (Strictly speaking, the correct t erm to
use for a device which reduces the st eam t emper at ur e to a poi nt which is
still above the sat urat i on poi nt is an at t emperat or, while one t hat lowers it
bel ow the sat urat i on poi nt may be referred to either as an at t emper at or or
a desuperheat er. However , in common engi neeri ng usage bot h terms are
appl i ed somewhat indiscriminately.)
7.2 The s pr ay- wat e r a t t e mpe r a t o r
One way of adj ust i ng the t emper at ur e of steam is to pump a fine spray of
compar at i vel y cool wat er dropl et s into the vapour. Wi t h the resulting
i nt ermi xi ng of hot st eam and cold wat er the cool ant event ual l y evaporat es
Steam-temperature control 137
so t hat the final mi xt ur e comprises an increased vol ume of st eam at a t em-
per at ur e whi ch is l ower t han t hat pri or to the wat er injection point. Thi s
cooling funct i on is achi eved in the at t emperat or.
The at t emper at or is an effective means of lowering the t emper at ur e of
the steam, t hough in t her modynami c terms it results in a reduct i on in the
performance of the pl ant because t he st eam t emper at ur e has to be raised
to a hi gher val ue t han is needed, onl y to be br ought down to the correct
val ue later, by injecting the spray wat er.
Al t hough the i nherent design of the at t emper at i on system may, in
theory, per mi t cont rol to be achi eved over a ver y wi de range of st eam
flows, it shoul d be underst ood t hat t he curve of the boiler' s nat ural charac-
teristic will restrict the l oad range over whi ch pract i cal t emper at ur e
cont rol is possible, regardless of the t ype of at t emper at or in use. It is not
unusual for the effective t emper at ur e- cont r ol range of a boi l er to be
bet ween onl y 75% and 100% of the boiler' s maxi mum cont i nuous rat i ng
( MCR) . Thi s limitation is also the result of the spray-wat er flow bei ng a
l arger pr opor t i on of t he st eam flow at low loads.
7.2.1 The mechanically atomised attemperator
Vari ous forms of spray at t emper at or are empl oyed. Fi gure 7.1 shows a
simple design wher e the high-pressure cooling wat er is mechani cal l y
at omi sed into small dropl et s at a nozzle, t her eby maxi mi si ng the ar ea of
cont act bet ween t he st eam and the water. Wi t h this t ype of at t emper at or
the wat er dropl et s leave the nozzle at a high vel oci t y and t herefore travel
for some distance before t hey mix wi t h the st eam and are absorbed. To
avoid stress-inducing i mpi ngement of cold dropl et s on hot pi pework, the
l engt h of straight pi pe in whi ch this t ype of at t emper at or needs to be
installed is qui t e long, t ypi cal l y 6 m or more.
Wi t h spray at t emperat ors, the flow of cooling wat er is rel at ed to the
flow rat e and the t emper at ur e of the steam, and this leads to a furt her lim-
itation of a fixed-nozzle at t emper at or . Successful break-up of the wat er
into at omi sed dropl et s requires the spray wat er to be at a pressure whi ch
exceeds the st eam pressure at the nozzle by a cert ai n amount (typically 4
bar). Because the nozzle presents a fixed-area orifice to t he spray wat er,
the pressure/ fl ow charact eri st i c has a square-l aw shape, resulting in a
restricted range of flows over whi ch it can be used (this is referred to as
limited t ur n- down or rangeability). The t ur n- down of the mechani cal l y
at omi sed t ype of at t emper at or is ar ound 1.5 : 1.
The t emper at ur e of the st eam is adj ust ed by modul at i ng a separ at e
spray-wat er cont rol val ve to admi t more or less cool ant into the steam.
Spray water
t
Nozzl e
Figure 7.1 Mechanically atomised desuperheater
Because of the limitations of the single nozzle, t he accuracy of control
t hat is possible with this type of at t emper at or is no great er t han + 8.5 C.
7. 2. 2 The variable-area attemperator
One way of overcomi ng the limitations of a fixed nozzle in an at t emper at or
is to use an ar r angement which changes t he profile as the t hr oughput of
spray water alters. Figure 7.2 shows the operat i ng principle of a variable
area, multinozzle at t emperat or. This employs a sliding pl ug which is
moved by an actuator, allowing the water to be injected t hrough a great er
or smaller number of nozzles. Wi t h this type of device, the amount of
water injected is regulated by the position of the sliding plug, a separate
spray-water control valve is therefore not needed.
Adequat e performance of this type of at t emperat or depends on the
velocity of t he vapour at t he nozzles being high enough to ensure t hat the
coolant droplets remai n in suspension for long enough to ensure their
absorption by the steam. For this reason, and also to provide t hermal pro-
tection for the pi pework in the vicinity of the nozzles, a t hermal liner is
often included in the pipe ext endi ng from the pl ane of the nozzles to a
poi nt some distance downst ream.
Spray water
Nozzles
F i g u r e 7 . 2 P r i n c i p l e o f a m u l t i n o z z l e desuperheater
The accuracy of control and the t ur ndown range available from a
multi-nozzle at t emperat or is considerably great er t han t hat of a single-
nozzle version, allowing the steam t emperat ure to be controlled to
+ 5.5C over a flow range of 40: 1.
7.2.3 The variable-annulus desuperheater
Anot her way of achieving accurate control of the steam t emper at ur e over
the widest possible dynami c range is provi ded by the variable-annulus
desuperheat er (VAD) (produced by Copes-Vulcan Limited, Road Two,
Winsford Industrial Estate, Winsford, Cheshire, CW7 3QL.). Here, t he
approach cont our of t he VAD head is such t hat when t he inlet steam flows
t hr ough an annul ar ring between the spray head and t he i nner wall of t he
steam pipe its velocity is increased and the pressure slightly reduced. The
cool ant enters at this poi nt and undergoes an instant increase in velocity
and a decrease in pressure, causing it to vapourise into a mi cron-t hi n l ayer
whi ch is stripped of f t he edge of t he spray head and propelled downst ream.
The stripping act i on acts as a bar r i er whi ch prevents the cool ant from
i mpi ngi ng on t he i nner wall of the st eam pipe. The downst r eam port i on of
the VAD head is cont oured, creat i ng a vort ex zone into whi ch any unab-
sorbed cool ant is drawn, exposing it to a zone
of low pressure and high t urbul ence, whi ch t herefore causes addi t i onal
evaporat i on.
Due to the Vent uri principle, t he pressure of t he cooled st eam is
quickly restored downst r eam of t he vena cont r act a point, resulting in a
very low overall loss of pressure.
An advant age of t he VAD is t hat , due to the cool ant injection
occurri ng at a point wher e t he st eam pressure is lowered, the pressure of
t he spray wat er does not have to be significantly hi gher t han t hat of t he
steam.
7.2.4 Other types of attemperator
At least two ot her designs of at t emper at or will be encount er ed in power-
station applications. The vapour-at omi si ng design mixes st eam with the
cooling water, thus ensuri ng mor e effective break-up of the wat er droplets
and shroudi ng the at omi sed droplets in a sheat h of st eam to provide rapi d
at t emperat i on.
Variable-orifice at t emperat ors i ncl ude a freely floating plug whi ch is
positioned above a fixed s e a t - - a design t hat generat es high t urbul ence
and mor e efficient at t emperat i on. The cool ant velocity increases simulta-
neously with the pressure drop, instantly vaporising t he liquid. Because of
the movement of t he plug, the pressure dr op across t he nozzle remai ns
const ant (at about 0.2 bar). The design of this t ype of at t emper at or is so
efficient t hat compl et e mi xi ng of t he cool ant and the steam is provi ded
wi t hi n 3 to 4 m of t he cool ant ent ry point, and the t emper at ur e can be
controlled to __+ 2.5 C, t heoret i cal l y over a t ur ndown range of 100 : 1.
Because t he floating plug moves against gravity, this t ype of at t em-
per at or must be installed in a vertical section of pipe with the st eam
t hr ough it travelling in an upwar d direction. However, because of the
efficient mi xi ng of st eam and coolant, it is permissible to provi de a bend
almost i mmedi at el y aft er t he device. Fi gure 7.3 shows a typical installa-
tion.
Temperature sensor
STEAM OUTLET
Long-radius bend
) k, Nonretum valve
J
Water from
sp~y.2atvaer e
STEAM INLET
Figure 7.3 Variable-orifice attemperator installation
7.2.5 Location of temperature sensors
Because the st eam and wat er do not mix i mmedi at el y at the pl ane of the
nozzle or nozzles, great care must be t aken to l ocat e the t emper at ur e
sensor far enough downst r eam of t he at t emper at or for the measur ement to
accurat el y represent the act ual t emper at ur e of the st eam ent eri ng the next
stage of t ube banks. Di rect i mpi ngement of spray wat er on t he t emper at ur e
sensor will result in t he final st eam t emper at ur e bei ng hi gher t han desired.
Fi gure 7.4 shows a typical installation, in this case for a vari abl e-annul us
desuperheat er.
7.2.6 Control systems f or spray-water attemperators
The simplest possible t ype of cont rol woul d be based on measuri ng t he
t emper at ur e of the st eam leaving the final superheat er, and modul at i ng
1
4
2

P
o
w
e
r
-
p
l
a
n
t

c
o
n
t
r
o
l

a
n
d

i
n
s
t
r
u
m
e
n
t
a
t
i
o
n

.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

~

/
)

0

g

r
~

~
n

~
m

I

%
=

m

~
U

~
n

I
'

P

e
~

!

t he flow of cooling wat er to t he spray at t emper at or so as to keep t he tem-
per at ur e const ant at all flow conditions. Unfort unat el y, because of t he
long t i me constants associated with t he superheat er, this form of cont rol
woul d pr oduce excessive deviations in t emper at ur e, and a mor e compl ex
ar r angement is requi red.
Two t i me constants are associated wi t h the superheat er. One represents
t he t i me t aken for changes in t he firing rat e to affect t he st eam t emper a-
ture, t he ot her is t he t i me t aken for t he st eam and wat er mi xt ur e leaving
the at t emper at or to appear at t he outlet of t he final superheat er. I n t erms
of t emper at ur e cont rol it is t he l at t er effect whi ch predomi nat es because,
al t hough changes in heat i nput will affect t he t emper at ur e of t he steam, a
fast-responding t emper at ur e- cont r ol loop will be able to compensat e for
t he al t erat i ons and keep t he t emper at ur e constant. It is t he react i on t i me
bet ween a change occurri ng in t he spray-wat er flow and t he effects being
observed in t he final t emper at ur e t hat det ermi nes t he ext ent of t he t em-
per at ur e variations t hat will occur.
Anot her probl em wi t h a simple system, as out l i ned above, is t hat it
does not per mi t any moni t or i ng anc~ control of t he t emper at ur e to occur
within t he st eam c i r c ui t - - onl y at t he exit from t he boiler.
These difficulties are addressed by t he appl i cat i on of a cascade cont rol
system as shown in Fi gure 7.5. This shows a simple st eam- t emper at ur e
cont rol system based on t he use of an i nt erst age at t emper at or whi ch is
l ocat ed in t he st eam circuit bet ween the pr i mar y and secondar y banks of
superheat er tubes.
Since it is t he t emper at ur e of t he st eam leaving t he secondar y super-
heat er t hat is i mport ant , this par amet er is measur ed and a correspondi ng
signal fed to a two- or t hr ee- t er m cont rol l er (proport i onal -pl us-i nt egral or
proportional-plus-integral-plus-derivative). I n this cont rol l er t he
measured-val ue signal is compar ed with a fixed desired-value signal and
t he cont rol l er' s out put forms t he desired-value i nput for a secondar y con-
troller. (Because t he out put from one cont rol l er ' cascades' into t he i nput of
anot her, this t ype of cont rol system is commonl y t er med ' cascade
control' .)
The secondar y cont rol l er compares this desired-value signal wi t h a mea-
sur ement represent i ng t he t emper at ur e of t he st eam i mmedi at el y aft er t he
spray-wat er at t emperat or.
It is a mat t er of some debat e as to whet her a t wo- t er m or t hr ee- t er m
cont rol l er should be used in this t ype of application. Because t he steam-
t emper at ur e sensors used are subjected to t he hi gh pressures and t emper a-
tures of t he superheat er, t hey have to be enclosed in substantial steel
pockets. Even wi t h t he best designs, pockets are usually slow-responding,
with t he result t hat any high-speed fluctuations in t he measured-val ue
Steam L
drum - ( ~
Figure 7.5
Primary
superheater
Attemperator
Secondary
superheater
Steam out
m
I i
, . s i
l
<
Water from
feed pumps
Steam-temperature control with a single inters/age spray a/tempera/or
signal will be smoot hed out and t he resul t ant signal will be fairly stable.
The use of a derivative t er m is therefore easier t han in, say, flow measure-
ment applications where small-scale but sudden changes in flow can occur.
When rapi d i nput changes are di fferent i at ed, t he cont rol l er out put
changes by a large amount , and for this reason t uni ng t hr ee- t er m flow con-
trollers for opt i mum response can become difficult. This is not a probl em
with the t emper at ur e controllers described here, and the application of
derivative act i on may be viable i f it is felt t hat this coul d provide i mproved
performance. As usual, it is i mpor t ant t hat the cont rol l er design should be
such that the derivative t erm affects only t he measured-val ue signal (not
the desired-value or error signals), since differential response to operat or-
i nduced setpoint changes is always undesirable.
I n Chapt er 6, reference was made to t he r equi r ement for t he spray
wat er to ent er t he a/ t emper a/ or at a pressure whi ch exceeds t he steam
pressure by a mi ni mum value. It is wort h r emember i ng t he poi nt made
there: that, wher e a variable-speed feed pump is used, car e must be t aken
to ensure t hat adequat e wat er / s t eam differential pressure is available
under all operat i onal conditions.
7. 2. 6. 1 Cont r ol l er s a t u r a t i o n effects
The t ype of cont rol system described above is commonl y encount er ed in
a wide vari et y of applications, and it is subject to an effect whi ch must be
underst ood and adequat el y addressed by t he design of t he controllers used
in the system. The effect is known as ' i nt egral sat urat i on' or ' reset wind-
up', and it is a charact eri st i c of i nt egral -act i on controllers whose out put
commands are fed into t he inputs of cascade or secondary controllers. It
sometimes confuses people when t hey are first i nt roduced to this sat urat i on
effect in st eam- t emper at ur e cont rol applications since the word ' satura-
tion' is also applied to a t her modynami c pr oper t y of steam. It is t herefore
i mpor t ant t hat t he poi nt is clearly underst ood t hat in this cont ext t he wor d
' sat urat i on' refers to a cont rol l er out put reachi ng a limiting val ue and t hen
at t empt i ng to exceed t hat figure.
Whet her t he i mpl ement at i on of a controller is achi eved in har dwar e or
by software its out put must always be const rai ned by some design limit or
other. At first sight, it may appear t hat the exception to this rule is the so-
called pulse or i ncrement al controller, where the out put commands di ct at e
a change in t he position of t he cont rol l ed device. However, i f t he cont rol l er
and act uat or are considered t oget her as a system it will be seen t hat satura-
tion will still occur when the act uat or reaches the limit of its movement .
Wi t h a gai n of x, t he out put of a cont rol l er will be the i nput er r or multi-
plied by x, but what happens if ei t her x or the error is so l arge t hat t he
resulting out put is outside t he range of signals t hat t he cont rol l er can
handl e (i.e. beyond t he limiting value of t he out put range)? I n this case t he
cont rol l er out put will adopt a magni t ude whi ch is fixed at t he l i mi t i ng
value, in whi ch condi t i on the out put is no l onger represent at i ve of t he
i nput error.
The r ange of inputs wi t hi n whi ch t he cont rol l er out put is represent at i ve
of the i nput error is inversely proport i onal to t he gain: with a l arge gain, a
small i nput error may force t he out put signal to the l i mi t i ng value, and
vice versa.
It is for this reason t hat cont rol engineers refer to the ' proport i onal
band' of t he controller, the proport i onal band of a cont rol l er being t he
reci procal of its gain, with a gai n of 1 being equi val ent to a proport i onal
band of 100%. It is wort h expl ori ng the significance of this st at ement wi t h
t he following exampl e.
I n Fi gure 7.6a t he controller has a gai n of 1. The i nput error is defined as
bei ng in the range 0 - 100% and t he out put is also in t he r ange 0 - 100%.
( a )
( b )
i |
Figure 7.6 Controller saturation
I f t he i nput error is slowly swung from one ext reme limit to t he ot her (i.e.
from - 1 0 0 % to +100%) , t he out put also moves from one limit to t he
ot her (in t he above case from - 1 0 0 % to +100%) .
In Figure 7.6b t he cont rol l er gai n has been i ncreased to 2, and this
time the out put reaches its l i mi t i ng val ue of +100% when t he i nput error
is only +50%, and it becomes - 1 0 0 % when t he i nput is - 5 0 %. In ot her
words, t he out put moves from one limit to t he ot her wi t h an i nput swing of
50% of t he ma xi mum range. I f t he i nput is i ncreased beyond t he val ue of
50% in ei t her direction, as shown by t he dot t ed lines, the out put cannot
respond, since t he out put has r eached the limit of its range and t he con-
troller has become sat urat ed.
The significance of t he expression ' proport i onal band' should now be
apparent . It is t he r ange of i nput signals wi t hi n whi ch t he out put is propor-
tional to t he input.
A good way of under st andi ng what is happeni ng is to t hi nk of the
i nput in t erms of a window. Every change t hat occurs wi t hi n t he wi ndow
Steam-temperature cont r ol 147
will be react ed to, but any change t hat occurs when t he i nput is outside
of t he wi ndow is invisible to the controller. The size of t he wi ndow is
inversely proport i onal to t he cont rol l er gain: t he hi gher t he gain, t he
smal l er t he window.
Wi t h a simple loop, t he effects of cont rol l er sat urat i on are not too
severe since t he out put has reached t he limit of movement of t he act uat or
and it starts to react as soon as t he cont rol l er i nput error has re-ent ered t he
pr opor t i onal - band window. In a cascade loop, however, t he presence of
t he two controllers leads to hi ghl y undesirable results.
To illustrate this point, Figure 7.7 shows a cascade loop wi t h pract i cal
values of t emper at ur e and gain added. As is common wi t h this t ype of
system, t he t emper at ur e t ransmi t t er has a suppressed range, wi t h 4 mA
bei ng t ransmi t t ed when t he t emper at ur e is 300C and 20 mA when t he
t emper at ur e is 600 C. When t he st eam t emper at ur e is 400C t he error
Pr i ma r y S e c o n d a r y
s u p e r h e a t e r s upe r he a t e r
- s f
/
Gai n = 2
400oc
3 0 0 - 6 0 0 C
4 mA - 20 mA
,.s I
\
DV = 550C
Gai n = 1
Figure 7.7 Saturation effects in steam-temperature control system
bet ween the measur ed and desired val ue signals is 150 C, whi ch is 50%
of t he 300 C i nput range and with a gain of 1 the cont rol l er out put is also
50%. Since t he secondary cont rol l er has a gain of 2, t he 50% signal from
t he pr i mar y cont rol l er is right at t he edge of its own pr opor t i onal - band
window.
Now let us see what happens when the t emper at ur e is outside t he pro-
por t i onal - band wi ndow when, as stated above, the secondar y cont rol l er
cannot react to any changes in it. I f the t emper at ur e is at, say, 350 C and
rising, the pr i mar y cont rol l er will react to the rise in t emper at ur e and
pr oduce out put commands whi ch should start increasing the flow of
cooling spray-water. However, these commands are initially invisible to
the secondary controller, and the spray valve will not be moved until t he
signal event ual l y enters the proport i onal -band wi ndow of the secondary
controller. By t hen it is far too late, and the situation is exacerbat ed by t he
long time const ant of the final superheat er (up to 1 minutes in some
cases). As a result, t he t emper at ur e continues to rise, and when t he spray
valve finally opens t he result is a severe overcorrect i on. The result is t hat
t he final st eam t emper at ur e and the spray valve openi ng will bot h begin to
oscillate.
This is a classical exampl e of t he effects of cont rol l er saturation, but
the probl em is not always underst ood by DCS vendors who are unfami l i ar
with boiler control systems. On the ot her hand, vendors who do recognise
t he probl em offer a vari et y of solutions. These usually involve t he use of
' t rack' and ' reset' facilities in the controllers (or t he software funct i on
blocks whi ch assume t he dut y of controllers). With such solutions the con-
troller out put is forced to follow a signal whi ch is connect ed to its ' track'
t ermi nal when the reset (integration) funct i on is disabled. I n one config-
urat i on, t he system monitors the out put signal of t he pr i mar y controller,
and when this reaches 100% the cont rol l er out put is forced to t rack the
i nput t emperat ure.
7.2.6.2 Prevention of over-cooling
In st eam- t emper at ur e cont rol applications it is i mpor t ant to prevent t he
t emper at ur e being reduced too far. I f the t emper at ur e at t he inlet of the
secondary superheat er falls to a value approachi ng t he sat urat i on t empera-
ture, wat er dropl et s could form in t he flowstream, raising t he possibility of
t her mal shock to t he pipework, and in addi t i on t he st eam circuit could
become part i al l y plugged. The flow t hr ough the obst ruct ed tubes will t hen
be reduced and t hei r surface t emper at ur e will rise, possibly causing
pr emat ur e t ube failure.
7.2.6.3 Mul t i st age attemperators
Some boilers have several banks of superheat er tubes. In these cases spray
at t emperat ors are normal l y provi ded bet ween the maj or banks, as shown
in Figure 7.8.
It will be seen t hat t he cont rol systems ar ound each superheat er
compri se cascade loops t hat are quite similar to those discussed earlier.
However, t he set-value signal for t he first stage of sprayi ng is deri ved from
the out put of the cont rol l er regul at i ng t he final st eam t emperat ure. I n
fact, the signal may be charact eri sed in some way to accurat el y represent
the relationship bet ween the t emper at ur e of t he st eam leaving t he second
stage of at t emper at i on and t hat at t he exit of t he first stage. When the
system is operat i ng correctly, with the final slave cont rol l er mai nt ai ni ng its
desired-value and measured-val ue signals at t he same value, the effect is to
mai nt ai n a const ant t emperat ure-di fferent i al across the second at t em-
perator. The t emper at ur e dr op across t he at t emper at or is a measure of t he
Prim ary Secondary 2M'steg e Tertiary
superheater superheater ettempemtor superheater
,A A A
I ls t-s ta ~
attemperator
8
11
Boiler sir
flow
14
13
10
Water from
feed pumps
Figure 7.8 Steam-temperature control with two-stage spray attemperation
work bei ng done by it, and by cont rol l i ng this to a known value t he
cooling l oad t hr ough the entire string of tubes can be apport i oned as
requi red.
The funct i on of t he maxi mum- sel ect or uni t (item 9) is to prevent
chilling as described in Section 7.2.6.2. The st eam pressure at t he dr um is
measur ed and its value charact eri sed (8) to pr oduce a signal whi ch repre-
sents t he sat urat i on t emper at ur e. I t em 8 also i ncorporat es a bias to
represent a safe mar gi n of operat i on and t he resul t ant signal is fed to t he
maxi mum-sel ect or unit. I f t he out put of the secondary-superheat er con-
troller should fall to a value t hat is at or below t he safety mar gi n above t he
sat urat i on t emper at ur e, it is i gnored by t he maxi mum-sel ect or unit, whi ch
clamps t he desired val ue signal for t he first stage at t emper at or at this
limit.
Anot her feat ure of t he system shown in this di agr am is t he pr ogr ammi ng
signal whi ch presets t he openi ng of t he first-stage spray-wat er valve
accordi ng to a charact eri st i c t hat the boiler designer has predicted. The
t emper at ur e- cont r ol systems t hen t ri m this openi ng to el i mi nat e any
residual error. These pr ogr ammi ng signals can overcome some of t he
boiler' s time-delays, pr oduci ng a bet t er and faster response to changes in
load.
7.3 Temperat ure cont rol wi t h t i l t i ng burners
As expl ai ned in Chapt er 3, t he burni ng fuel in a corner-fi red boiler forms
a l arge swirling fireball whi ch can be moved to a hi gher or lower level in
t he f ur nace by tilting the burners upwards or downwar ds wi t h respect to a
midposition. The repositioning of t he fireball changes t he pat t er n of heat
transfer to t he various banks of superheat er tubes and this provides an
efficient met hod of cont rol l i ng the st eam t emper at ur e, since it enables t he
use of spray wat er to be reserved for fi ne-t uni ng purposes and for emergen-
cies. I n addition, the tilting process provides a met hod of cont rol l i ng
furnace exit t emperat ures.
Wi t h such boilers, t he st eam t emper at ur e control systems become signif-
i cant l y di fferent from those of boilers with fixed burners. The boiler
designer is able to define the opt i mum angul ar position of t he burners for
all loads, and t he control engi neer can t hen use a funct i on gener at or to set
t he angle of tilt over t he load range to mat ch this characteristic. A t em-
per at ur e cont rol l er trims t he degree of tilt so t hat the correct st eam
t emper at ur e is at t ai ned.
7.4 Co nt r o l l i ng t he t e mp e r a t u r e of r e h e a t e d s t e a m
I n boi l ers wi t h r e he a t stages, c ha nge s i n f i r i ng i ne vi t a bl y affect t he t e m-
p e r a t u r e of b o t h t he r e h e a t e r a n d t he s upe r he a t e r . I f a si ngl e c ont r ol
me c h a n i s m wer e t o be us ed f or b o t h t e mp e r a t u r e s t he r es ul t i ng i nt er -
act i ons wo u l d ma k e c ont r ol - s ys t e m t u n i n g di ffi cul t , i f n o t i mpossi bl e, t o
opt i mi s e. Suc h boi l ers t her ef or e use t wo or mo r e me t h o d s o f cont r ol .
Because of t he l ower o p e r a t i n g pr essur e of r e he a t s t e a m syst ems, t he
t h e r mo d y n a mi c c ondi t i ons ar e si gni f i cant l y di f f er ent f r o m t hose of super -
heat er s, a n d t he i nj ect i on of s pr a y wa t e r i nt o t he r e h e a t e r s ys t em has a n
u n d u e effect on t he effi ci ency of t he pl ant . For t hi s r eason, i t is pr ef er abl e
for t he r e he a t st ages t o be c ont r ol l e d by t i l t i ng b u r n e r s (i f t hese ar e
avai l abl e) or by a p p o r t i o n i n g t he fl ow of hot c o mb u s t i o n gases over t he
var i ous t ube banks . Howe ve r , i f t he s u p e r h e a t t e mp e r a t u r e is c ont r ol l e d by
b u r n e r t i l t i ng, gas a p p o r t i o n i n g or s pr ay a t t e mp e r a t i o n mu s t t h e n be us ed
for t he r e he a t stages.
I n boi l ers wi t h fi xed bur ner s , s t e a m- t e mp e r a t u r e c ont r ol ma y be
a c hi e ve d by a dj us t i ng t he o p e n i n g of d a mp e r s t h a t c ont r ol t he fl ow o f t he
f ur na c e gases across t he var i ous t ube banks. I n s ome cases t wo s e pa r a t e
sets of d a mp e r s ar e pr ovi de d: one r e g u l a t i n g t he fl ow over t he s u p e r h e a t e r
banks , t he o t h e r c ont r ol l i ng t he fl ow over t he r e h e a t e r banks .
Be t we e n t he m, t hese t wo sets o f d a mp e r s deal wi t h t he ent i r e v o l u me
o f c o mb u s t i o n gases pas s i ng f r o m t he f ur na c e t o t he c hi mne y. I f b o t h wer e
t o be cl osed at t he s a me t i me, t h e fl ow of t hese gases wo u l d be sever el y
r est r i ct ed, l e a di ng t o t he possi bi l i t y o f d a ma g e t o t he s t r uc t ur e d u e t o over -
pr es s ur i s at i on. For t hi s r e a s on t he t wo sets ar e c o n t r o l l e d i n a so- cal l ed
' s pl i t - r ange' f ashi on, wi t h one set be i ng al l owed t o cl ose onl y wh e n t he
o t h e r has f ul l y ope ne d.
Th e s e d a mp e r s pr ovi de t he ma i n f or m of cont r ol , b u t t he r es pons e o f
t he s ys t em is ver y slow, pa r t i c ul a r l y wi t h l ar ge boi l ers, wh e r e t he t e mp e r a -
t ur e r es pons e t o c ha nge s i n he a t i n p u t exhi bi t s a s e c ond- or de r l ag of
a l mos t t wo mi n u t e s ' dur a t i on. For t hi s r eason, a n d also t o pr ovi de a me a n s
of r e d u c i n g t he t e mp e r a t u r e of t h e r e he a t s t e a m i n t he e v e n t o f a f ai l ur e i n
t he d a mp e r syst ems, s pr ay a t t e mp e r a t i o n is p r o v i d e d for e me r g e n c y
cool i ng.
T h e s pr a y a t t e mp e r a t o r is s hut unl ess t he t e mp e r a t u r e at t he r e h e a t e r
out l e t r eaches a p r e d e t e r mi n e d h i g h l i mi t . Wh e n t hi s l i mi t is exceeded, t he
s pr a y val ve is ope ne d. I n t hi s c ondi t i on, t he a mo u n t o f wa t e r t h a t is
i nj ect ed is t ypi cal l y c ont r ol l e d i n r e l a t i on t o t he t e mp e r a t u r e at t he
r e h e a t e r i nl et , t o b r i n g t he exi t t e mp e r a t u r e b a c k i nt o t he r egi on whe r e
g a s - a p p o r t i o n i n g or b u r n e r t i l t i ng c a n once a ga i n be effective. T h e
relationship bet ween the cold reheat t emper at ur e and the requi red spray-
wat er flow can be defined by the boiler designer or process engineer.
I f a t ur bi ne trip occurs the reheat flow will collapse. In this situation
the reheat sprays must be shut i mmedi at el y in or der to pr event serious
damage bei ng caused by the admission of cold spray wat er to the turbine.
7.4.1 Spray attemperators f or reheat applications
At first, it may seem t hat reheat spray-wat er at t emper at or systems should
be similar to those of the superheat er. This is unt rue, because reheat at t em-
perat ors have to cope wi t h the l ower st eam pressure in this section of the
boiler, whi ch renders the pressure of the wat er at t he di scharge of the feed
pumps too high for satisfactory operat i on. Al t hough a pressure-reduci ng
val ve coul d be i nt roduced into the spray-wat er line, this woul d be an
expensive solution whose l ong-t erm reliability woul d not be satisfactory
because of t he severe conditions to which such a val ve woul d be subjected.
A bet t er solution woul d be to derive the suppl y from the feed-pump inlet.
In some cases, even this is ineffective, and separat e pump sets have to be
provi ded for the reheat sprays.
7.5 Gas recycl i ng
Wher e boilers are designed for burni ng oil, or oil and coal in combi nat i on,
t hey are frequent l y provi ded wi t h gas-reci rcul at i on systems, where the hot
gases exiting the l at er stages of the boiler are reci rcul at ed to the bot t om
par t of the furnace, close to the burners. This pr ocedur e increases the
mass-flow of gas over the t ube banks, and t herefore increases the heat
transfer to t hem.
Because the gas exiting the furnace is at a low pressure, fans have to be
provi ded to ensure that the gas flows in the correct direction. Cont rol l i ng
the flow of recycl ed gases provides a met hod of regul at i ng the t emper at ur e
of the super heat ed and reheat ed st eam, but interlocks have to be pr ovi ded
to prot ect the fan against hi gh- t emper at ur e gases flowing in a reverse
di rect i on from t he bur ner area i f the fan is st opped or i f it trips.
7.6 Summary
Thi s ends our syst em-by-syst em survey of boi l er and HRS G cont rol and
i nst rument at i on systems, and we will now t urn our at t ent i on to some of the
design aspects rel at i ng to the equi pment t hat is used in i mpl ement i ng these
systems.
Chapter 8
Control equipment practice
On an operat i onal plant, t he cont rol systems t hat have so far been
exami ned ma y be i mpl ement ed in any of a vari et y of ways, rangi ng from
pneumat i cs to advanced comput er-based systems, but in all cases it should
be possible to identify t he various loops wi t hi n t he rel evant configuration.
These days, most cont rol functions are i mpl ement ed by means of a
comput er-based system, so we shall now briefly look at a typical configura-
tion. Aft er that, we shall exami ne some of t he ot her har dwar e used in t he
systems and t hen consider t he envi r onment al factors t hat influence the
selection of cont rol and i nst rument at i on equi pment .
8.1 A typical DCS configuration
DCS stands for ' distributed cont rol system' . The t er m ' distributed' means
t hat several processors are operat i ng together. This is usually achieved by
dedi cat i ng tasks to di fferent machines. It does not necessarily mean t hat
t he separat e comput ers are physically l ocat ed in di fferent areas of t he
plant.
Fi gure 8.1 shows how a typical system may be arranged. The following
notes rel at e to i ndi vi dual part s of t hat system. I n practice, each manuf ac-
t urer will usually offer some vari ant of t he system shown in this di agram,
and t he rel evant description should be consulted, but t he comment s made
here are general ones whi ch may help to identify points whi ch should be
considered and discussed when a new or refurbished system is being consid-
ered.
A I o n 'n / e v e n t O p e r a t o r w o r k s t a t i o n 1 O p e r a t o r w o r k s t a t i o n 2
~ e r i
E n g i n u a r i n g / d i a g n e s t i c s
w o r k s t a t i o n a n d p r i n t e r
D a t a h i g h w a y
I C P U , m e m o r y , i n t e r f a c e s ,
p o w e r s u p p l i e s e t c .
T e r m i n a t i o n a n d
m a r s h a llin g
c u b i c l e
Figure 8.1
A n d o g o e a n d
d i g i t a l I / 0 c a r d s
S i g n a l s t o a c t u a t o r s ,
m o t o r s t a r t e r s e t c .
A typical DCS system configuration
S ig n a ls f r o m
t r a n s m i t t e r s , l i m i t
s w i t c h e s e t c .
8. 1. 1 The cent ral syst em cabinets
Locat ed near the cent re of Fi gure 8.1 are t he cabi net s whi ch house the pro-
cessors t hat execut e t he cont rol functions. These cubicles also cont ai n the
at t endant interface and i nput / out put ( I / O) cards and the necessary
power suppl y units (PSUs). The l at t er will usual l y be dupl i cat ed or tripli-
cat ed, wi t h aut omat i c changeover from one to anot her in the event of the
first failing. (This aut omat i c changeover is often referred to as ' di ode auc-
t i oneeri ng' because silicon diodes are used to feed power from each uni t
ont o a common bus-mai n. In the event of the operat i onal power - suppl y
uni t failing, its di ode prevent s a power reversal while the back- up power
uni t takes over.) At this t i me it is i mpor t ant t hat the system shoul d raise an
al ar m to war n t hat a PSU failure has occurred. Ot herwi se the DCS will
cont i nue to oper at e wi t h a diminished power-suppl y reserve and any
furt her failure coul d have serious consequences.
Clearly, the DCS cannot oper at e cont i nuousl y from bat t eri es alone. A
reliable and stable source of power will t herefore need to be avai l abl e
(usually l l 0 V or 240V AC). I f the DCS includes i nt ernal back- up
Control equipment practice 155
batteries it will cont i nue to operate if the AC feed is lost, but such
batteries are normal l y sized to retain essential dat a in the memor y and to
provide a limited amount of functionality. They may also allow l i mi t ed
control to be performed, but all this will funct i on for only a short peri od
(typically 30 min) and it is therefore usual to provi de an external uninter-
ruptable power supply (UPS) system which can allow the pl ant to be
operat ed for a longer time. The durat i on of this period warrant s very
careful consideration. Long periods require large and expensive batteries
and charger systems, and this expense can rarely be justified (especially
since such a maj or power loss will probabl y have disabled all pumps,
motors etc.). Instead, it is common to provide a bat t ery capacity t hat will
allow the pl ant to be safely shut down in t he event of power failure. The
det ermi nat i on of t he t i me required for such an operat i on is a mat t er of dis-
cussion with t he process design engineers and t he pl ant management .
I n addi t i on to supplying t he comput er system, the power-supply
system will usually also have to provide DC supplies for 4- 90 mA trans-
mitters and for limit-switch contacts. (The voltage connect ed to a cont act
and thence to the DCS i nput channel is often referred to as the ' wet t i ng
voltage'.) Transmi t t ers operat i ng on the 4 - 2 0 mA range which are
powered from t he DCS are sometimes called 'passive'. I n comparison,
those t hat operat e from local power supplies are called 'active'.
The I / O cards consist of analogue and digital i nput and out put
channels. Anal ogue inputs convert the i ncomi ng 4 - 2 0 mA signals to a
form which can be read by t he system. The printed-circuit cards for
anal ogue inputs may or may not provide ' galvanic isolation'. Wi t h a galva-
nically isolated device t he signal circuit is electrically isolated from others,
from the system eart h and from the power-supply common rail. Galvanic
isolation simplifies circuit design since it prevents i nadvert ent short-
circuiting, but consideration should be given to the possible build-up of
static charges on compl et el y ungr ounded circuits, which could cause
damage to i nput devices (which are usually rat ed for not more t han a few
tens or hundreds of volts). Thi s is normal l y an i mpor t ant consideration
only in areas of very low humi di t y or where there is a strong presence of
charged particles.
The commi ssi oni ng process, and the task of identifying and correcting
faults, are operations which are considerably assisted by the provision of
l i ght -emi t t i ng diode status indicators (LEDs) on the digital out put cards.
Some systems provide switches on the digital i nput cards, which can be
of assistance with commi ssi oni ng and fault-finding. However, i nadvert ent
or deliberate mal operat i on of such switches can have serious consequences,
since the DCS is t hen provi ded with incorrect plant-status i nformat i on and
it may take i nappropri at e action. (The use of logic probes, which inject
si gnal s i nt o a s ys t em t o check its ope r a t i on, is also t o be de pr e c a t e d, for
s i mi l ar reasons. )
An a l o g u e a n d di gi t al I / O c ha nne l s ar e n o r ma l l y g r o u p e d i nt o 8, 16, 32
or 64 c ha nne l s p e r pr i nt e d- c i r c ui t car d. 8 or 16 a n a l o g u e i n p u t ( AI )
c ha nne l s ar e c o mmo n l y a c c o mmo d a t e d on a car d, but a n a l o g u e o u t p u t
( AO) c ha nne l s consi st of c u r r e n t ge ne r a t or s a n d so oc c upy mo r e s pace a n d
ar e mo r e expensi ve t h a n AI channel s , whi c h ar e based on s mal l ope r a -
t i ona l - a mpl i f i e r devi ces ( op- amps ) . Di gi t al i n p u t ( DI ) c ha nne l s ar e ver y
s i mpl e a n d c h e a p a n d ma y be g r o u p e d i nt o 16 or even 32 i nput s t o a si ngl e
car d. Di gi t al o u t p u t ( DO) c ha nne l s dr i vi ng l owe r - powe r devi ces ar e also
s i mpl e a n d cheap, a n d ma y also c ompr i s e 16 or 32 i nput s t o a si ngl e car d,
b u t DOs f or hi ghe r - powe r devi ces (such as s ol enoi d valves) us ual l y r equi r e
t he pr ovi s i on of relays. Th e s e ma y be i nc l ude d on t he c a r d or t h e y ma y be
s epar at e.
Wh e n c ons i de r i ng t he pr ovi s i on of s par e I / O channel s , car ef ul t h o u g h t
mu s t be gi ven t o t he g r o u p i n g of channel s . I f a s ys t em has 256 a n a l o g u e
i n p u t c ha nne l s avai l abl e, of whi c h onl y 230 ar e act i vel y used, i t ma y be
sai d t o ha ve 11% s par e c a pa c i t y i n t hi s ar ea. Howe ve r , t he g r o u p i n g of
f unc t i ona l ar eas i nt o car ds wi l l i nevi t abl y r esul t i n t he oc c ur r e nc e of mo r e
s pa r e c ha nne l s i n one a r e a t h a n i n a not he r . I t is possi bl e, t her ef or e, t o have
t he r e q u i r e d a mo u n t of s pa r e I / O c a pa c i t y avai l abl e i n t e r ms of t he
over al l syst em, b u t t o be una bl e t o modi f y or e xt e nd a p a r t i c u l a r p a r t o f
t he s ys t em safely, becaus e no s par e c ha nne l s ha ve been p r o v i d e d i n t he
r e qui r e d ar ea.
Spa r e c a pa c i t y s houl d be p r o v i d e d bot h i n t he f or m of ' p o p u l a t e d '
c ha nne l s (i.e. s par e i nput s a n d o u t p u t s on i ndi vi dua l cards) a n d ' u n p o p u -
l a t e d' s pace (i.e. spaces for a ddi t i ona l cards). To avoi d a s paghet t i - l i ke
t a ngl e of cr oss- connect i ons, t he s par e spaces s houl d be sensi bl y di s t r i but e d
t h r o u g h t he syst em.
8. 1. 2 Termi nat i on and marshal l i ng
I t is i mp o r t a n t t o u n d e r s t a n d t ha t t he g r o u p i n g o f i nput s a n d o u t p u t s on
t he I / O c a r ds does not al ways c or r e s pond wi t h t he g r o u p i n g of si gnal s i nt o
mu l t i p a i r cabl es, whi c h is di c t a t e d by t he phys i cal a r r a n g e me n t o f
e q u i p me n t on t he pl a nt . Whi l e i t is sensi bl e t o avoi d mi x i n g di f f er ent
c ont r ol syst ems (e.g. f eed wa t e r c ont r ol a n d c o mb u s t i o n cont r ol ) ont o a
si ngl e car d, t he si gnal s associ at ed wi t h a si ngl e syst em will not necessar i l y
all be c a r r i e d i n t he s a me cabl e. T h e r esul t is t ha t a c e r t a i n de gr e e of cross-
c o n n e c t i o n or ' ma r s h a l l i n g ' is al ways r e qui r e d.
Wel l - des i gned syst ems wi l l pr ovi de a d e q u a t e facilities for ne a t l y ma r -
s hal l i ng t he si gnal connect i ons , b u t this i nevi t abl y r equi r es t h a t t he
identification of signal connections and t hei r location in t he cable system
is known at an earl y stage of t he cont ract . The l at er this probl em is
resolved, the mor e compl ex and unt i dy the system will become.
Compl exi t y and untidiness can be dangerous because it can l ead to
mistakes occur r i ng dur i ng commissioning or afterwards.
8. 1. 3 Operat or wor ks t at i ons
The operat or workstations consist of screens on whi ch pl ant i nformat i on
can be observed, plus keyboards, trackballs or ' mouse' devices allowing t he
oper at or to send commands to t he system. They also comprise printers for
operat i onal records, logging of events (such as st art -up of a pump), or
alarms. Some systems also provide plotters (one use of plotters is to det ect
the possible stalling of an axial-flow fan, as described in Chapt er 3).
The screens can be or di nar y cat hode- r ay tube types as used with
personal comput ers, or t hey may be large-screen pl asma displays or pro-
j ect i on systems. The selection of the t ype of screen depends on t he
operat i onal requi rement s, but will ul t i mat el y be det er mi ned by t he
available budget. Critical ergonomi c factors affect t he opt i mum design of
t he workstations, and great care must be exercised to ensure t hat t he pl ant
can be operat ed safely under all conceivable modes of failure, and t hat no
computer-assisted errors can occur due to t he operat or being confused by
the i nformat i on present ed to hi m or her.
An i mpor t ant consi derat i on is t he screen updat e time. This is t he t i me
bet ween the occurrence of an event and its appear ance on the screen. As
system l oadi ng is increased, this time can become ext ended, but t he
oper at or will need to be made aware of each event as soon as possible aft er
it occurs, so t hat corrective act i on can be taken. An updat e time of 1 s is
barel y adequat e to deal wi t h fast-moving events, but it can be qui t e
difficult to achieve.
8.2 I n t e r c o n n e c t i o n s b e t we e n t he s y s t e ms
The considerations appl yi ng to field cabling are deal t with in Section 8.8.
However, special t hought needs to be given to the dat a highway. This is a
high-speed link over whi ch a great deal of i nformat i on is t ransmi t t ed. The
cable empl oyed for this purpose is very specialised, and great care has to
be t aken in its installation. Physical damage, severe bendi ng or i ncorrect
t ermi nat i on can cause mal operat i on. I f a fibre-optic cable is used, t he con-
siderations t hat appl y to this t ype of cabling must be met i cul ousl y
followed.
The integrity of the data highway is crucial to the safety of the plant
and therefore it is usually duplicated. However, the provision of a sophisti-
cated dual -redundant highway with full error-checking and correction has
on occasion been completely negated by the cable installer runni ng both
cables on the same tray, or over the same route. An incident that damages
one cable will in all probability also damage the second one, with severe
consequences.
8.3 Eq u i p me n t s e l e c t i on and e n v i r o n me n t
Although modern gas-fired plant naturally tends to be clean in comparison
with its coal-fired equivalents, any power-station environment still
presents a severe test for electronic systems. The control-system designer
has to deal with the problems of operating low-voltage, potentially inter-
ference-prone, electronic equipment in close proximity to electrical plant
operating at 11 kV and above, with all its attendant switchgear and trans-
formers. The situation is exacerbated when considerations of safe
operation in hazardous environments are brought into the picture. It
becomes even worse when considering the dust, dirt and vibration that are
significant factors in practical power-plant environments. Naturally, the
latter problems (dust and dirt) become particularly acute in coal-fired
plant.
The success of a control system depends on the designer understanding
and addressing these factors. To assist in this process the following chapter
provides an outline of good equipment design and installation practices.
Because the subject covers so many different disciplines, the chapter is
divided into three sections:
Mechanical factors: the ground rules for providing good facilities for
control and instrumentation equipment.
Electromagnetic compatibility: guidelines for minimising the risk of
maloperation caused by interference.
Physical environmental considerations: dealing with dust, dirt,
vibration and hazardous atmospheres.
These matters must be understood and judiciously applied when an
installation is being planned, but doing this involves considerable interplay
with the civil and mechanical-engineering disciplines, and appropriate
action must therefore be taken at a very early stage in the design and con-
struction phases of the plant. I n a new plant, given diligence and
understanding on the part of all the disciplines involved, one can hope to
achieve this goal. But in the case of a refurbishment project the task
be c ome s mu c h mo r e di ffi cul t , be c a us e her e one is de a l i ng wi t h a p l a n t
whose c o n s t r u c t i o n is a l r e a dy c ompl e t e . I n t hi s case t he c ont r ol - s ys t e m
des i gner mu s t wor k wi t h wh a t a l r e a d y exists. I n t he e nd i t ma y c o me t o a
ma t t e r of f i ght i ng a di t c h- by- di t c h bat t l e, e ve nt ua l l y r e t r e a t i n g t o t he l ast
p r i n c i p l e - - t h e one t h a t mu s t ne ve r be s a c r i f i c e d - - wh i c h is t o o b t a i n a n
i ns t a l l a t i on t h a t is safe t o o p e r a t e a n d ma i n t a i n .
8.4 Me c hani c al f ac t or s and e r g o n o mi c s
I n t hi s s ect i on we shal l cons i der t he me c h a n i c a l i ns t a l l a t i on o f el ect r oni c
c ont r ol e q u i p me n t . Thi s is necessar i l y a s u mma r y , a n d l i ke mos t aspect s o f
t e c h n o l o g y i t is af f ect ed by c h a n g i n g r e q u i r e me n t s a n d t echnol ogi es.
Because r e qui r e me nt s , t echnol ogi es a n d t he avai l abi l i t y o f ma t e r i a l s ar e
al ways c ha ngi ng, s ome f o r m o f g u i d a n c e o n u p - t o - d a t e pr a c t i c e s houl d be
s ought a n d i t is t e mp t i n g t o t h i n k t h a t t he sel ect ed s ys t em v e n d o r will be
abl e t o p r o v i d e this.
Re p u t a b l e ve ndor s s houl d be pl eas ed t o p r o v i d e g u i d a n c e on i ns t a l l a t i on
pr act i ces t o be e mp l o y e d wi t h t he i r e q u i p me n t a n d syst ems. Thi s is p a r t l y
becaus e by p r o v i d i n g s uch i n f o r ma t i o n t h e y d e mo n s t r a t e t h a t t he y ar e
e xpe r i e nc e d i n powe r - s t a t i on wor k, a n d ar e abl e a n d wi l l i ng t o he l p wi t h
s uch ma t t e r s . I t is al so i n t he i r o wn i nt er es t t o do e v e r y t h i n g possi bl e t o
ens ur e t h a t t hei r syst ems wi l l not be expos ed t o me c h a n i c a l or e l e c t r oma g-
net i c c ondi t i ons t h a t c o u l d j e o p a r d i s e t he i r pe r f or ma nc e .
Howe ve r , i t wi l l be unwi s e t o wa i t unt i l a specific v e n d o r has be e n
sel ect ed becaus e i n mo s t cases t hi s a c t i on occur s wh e n t he basi c c onc r e t e
a n d st eel c ons t r uc t i on is a l mos t c ompl e t e , by whi c h t i me i t wi l l be t oo l at e
t o ma k e a n y changes . I t is useful t o obt a i n g u i d a n c e o n c u r r e n t pr a c t i c e
f r om a r a n g e o f s ys t em vendor s .
8. 4. 1 Si t e considerations
T h e el ect r oni c assembl i es t h a t c ompr i s e a c ont r ol s ys t em wi l l ge ne r a l l y
be l oc a t e d i n t hr e e areas:
T h e field: whe r e t r a ns mi t t e r s , sensors, de t e c t or s a n d a c t ua t or s ar e
si t ed.
T h e e q u i p me n t r oom( s ) : a c c o mmo d a t i n g t he c ont r ol cubi cl es,
pr ocessor s, I / O faci l i t i es a n d power - s uppl i es .
T h e c ont r ol r oom: h o u s i n g t he o p e r a t o r faci l i t i es (screens a n d
ke yboa r ds ) , pl us t he s ys t em pr i nt e r s etc.
Th e s e ar eas r e pr e s e nt ver y di f f er ent e n v i r o n me n t s for t he e q u i p me n t t he y
c ont a i n, r a n g i n g f r om t he severe c ondi t i ons of dust , di r t , h u mi d i t y , heat ,
vi br a t i on a n d h a z a r d o u s ar eas t h a t ar e t o be f ound on t he pl a nt , t o t he
c o mp a r a t i v e l y qui e t a n d c l e a n c ondi t i ons t ha t s houl d be f o u n d i n t he
c ont r ol r oom.
8. 4. 1. 1 Fi el d equipment
Ever y c ont r ol syst em d e p e n d s for its o p e r a t i o n on a c c ur a t e i n f o r ma t i o n
on t he p l a n t be i ng c ont r ol l e d, ( whi ch is t he d u t y of t he pr ocess t r a ns mi t -
ters), a n d t he abi l i t y t o a p p l y t he r es ul t i ng c o mma n d s t o t he p l a n t ( whi ch
is d o n e by act uat or s ) . Th e s e e l e c t r o me c h a n i c a l t r a ns duc e r s ar e vi t al t o t he
p r o p e r o p e r a t i o n of t he syst em. T h e y must ope r a t e ef f i ci ent l y a n d rel i abl y.
Un f o r t u n a t e l y , i t ems o f e q u i p me n t i n t hese cr i t i cal ar eas ar e subj ect t o
pa r t i c ul a r l y severe di ffi cul t i es. For a st ar t , t he des i gn of e l e c t r o me c h a n i c a l
t r a ns duc e r s r equi r es a bl e nd of g o o d el ect r oni c e n g i n e e r i n g a n d me c h-
ani cal engi neer i ng. A t h o r o u g h u n d e r s t a n d i n g of me t a l l u r g y a n d engi -
ne e r i ng c he mi s t r y is of t en also r e qui r e d. Des i gni ng e q u i p me n t whe r e
sever al di sci pl i nes ar e i nvol ve d is mu c h mo r e di ffi cul t t h a n wo r k i n g i n onl y
one di sci pl i ne. I f t hi s h u r d l e is o v e r c o me successful l y a g o o d devi ce will
resul t , b u t it will t h e n be i ns t al l ed on t he p l a n t whe r e it will be expos ed to
t he severe e n v i r o n me n t s t h a t of t en exist t her e. T h e e q u i p me n t ' s o p e r a t i o n
wi l l t h e n d e p e n d heavi l y on t he a p p l i c a t i o n of t he best possi bl e i ns t al l at i on
a n d ma i n t e n a n c e pr act i ces.
A successful c ont r ol syst em r equi r es det ai l ed def i ni t i on of each
c o mp o n e n t pa r t , a n d i n t he case of t he t r a ns mi t t e r s t hi s is a c hi e ve d by
me t i c ul ous speci f i cat i on o f t he t r a n s mi t t e r itself, a n d by car ef ul def i ni t i on
of h o w it will be i nst al l ed.
Fi gur e 8.2 shows one st ep i n t he l a t t e r process, a so- cal l ed ' hook- up'
d i a g r a m for a t r a n s mi t t e r me a s u r i n g hi gh pressures. T h e h o o k - u p defi nes
h o w t he t r a n s mi t t e r is p i p e d u p t o t he process. Two t ypes of h o o k - u p
d i a g r a m ar e i l l us t r at ed her e, one s howi ng all t he me c h a n i c a l i t ems t h a t
wi l l be ne e de d t o c o mp l e t e t he assembl y ( such as el bows, t ees a n d uni ons ) ,
t he ot he r out l i ni ng i n s c he ma t i c f or m h o w t he syst em oper at es . Nor ma l l y,
onl y one o f t hese t ypes will be used for a gi ven c ont r a c t . Suc h d i a g r a ms
def i ne t he c onne c t i ons a n d as such ar e a n essent i al pr e r e qui s i t e for i nst al l a-
t i on of an i n s t r u me n t . T h e det ai l ed ver s i on of t he h o o k - u p is useful for
c o s t i n g / e s t i ma t i n g pur pos es , a l t h o u g h t he s ame i n f o r ma t i o n ma y be
p r o v i d e d i n a s i mpl e r f or m.
A few poi nt s a b o u t t hi s d i a g r a m wa r r a n t f ur t he r di scussi on. T h e
t a p p i n g - p o i n t i sol at i ng val ve is usual l y p r o v i d e d by whoe ve r i nst al l s t he
ma i n hi gh- pr es s ur e p l a n t pi pewor k. I t is c onne c t e d t o t he t r a ns mi t t e r
subsyst em by a s mal l - bor e line, k n o wn as t he ' i mpul s e' pi pe, a n d t he
C o n n ectio n to
p ro c e s s
T a p p in g -p o in t
is o la tin g v a lv e
[ ]
P re s s u re
tra n s m itte r
T w o -v a lv e
m a n ifo ld
Im p u ls e
p ip e
J
=Master"=Martyr" t assemblyvalveBlwdwn
P r o c e ~ T j 1
r ~ , ~ . ~ , ~ Vent
I Is o la tio n
E
\
\
B low dow n
Figure 8.2 Pressure transmitter'hook.up' ( HP)
selection of t he correct type and size of pipe will have a considerable
bearing on t he accuracy, reliability and mai nt ai nabi l i t y of t he installation.
The st andards t hat are applied change from time to time and var y
bet ween countries and users. It was usual for HP pressure and differential-
pressure t ransmi t t ers to be connect ed to t he process via 15mm OD
stainless-steel impulse pipes and 20 mm nominal-size valves, while dr um-
level t ransmi t t ers were connect ed via 16mm OD pipes and 3 2 mm
nominal-size valves. Commer ci al pressures have nowadays led to a
situation wher e some purchasers merel y stipulate t hat t he whol e system
should operat e satisfactorily (sometimes wi t hi n defined paramet ers) for a
given time (say, t went y years). Little or not hi ng is said about mat t ers
such as mai nt ai nabi l i t y of ease of access. Usi ng t he above connect i on sizes
will lead to an installation t hat is workable, reliable and easy to
mai nt ai n.
The sizes of t he connect i on pipes and valves form only one par t of t he
picture. Ot her mat t ers concer n t he lengths, lagging, and slope (rise or fall)
of t he connect i ng pipes (which will allow t hem to be vent ed or drai ned),
and so on.
The two-valve i nst rument mani fol d shown in Fi gure 8.2 is a st andar d
subassembly whi ch may ei t her be i nt egral wi t h t he t r ansmi t t er or provi ded
as a separat e i t em as shown. It allows the t ransmi t t er to be connect ed, cali-
brat ed and vent ed before removal .
The blowdown-valve assembly enables t he pi pework to be flushed
t hr ough to remove ent r ai ned gases, deposits etc., to a suitable dr ai n or
vessel. I n this exampl e it comprises two valves, a ' mast er' and ' ma r t y r ' - -
an ar r angement t hat enhances l ong-t erm mai nt ai nabi l i t y in a high-
pressure application. I f a single bl owdown valve were to be provi ded in
such an application, the differential pressures and velocities to whi ch it
woul d be subjected each t i me it is opened or closed woul d quickly erode
t he internals. The ' mast er and mar t yr ' assembly operates as follows to
avoid this probl em. Pri or to initial commi ssi oni ng of the system, bot h
valves are closed. When pressure has been applied, the mast er valve is
opened first, followed by t he mar t yr . When any debris or undesi red gas or
vapour has been ejected, t he mar t yr is shut off first, followed by t he
master. When t he system is to be shut off agai n aft er use, t he mast er is
opened while t he ma r t yr remai ns closed, and t hen t he mar t yr is opened.
By this means t he onerous dut y of openi ng or closing of t he pressurised
system to t he at mosphere is always handl ed by t he mar t yr , wi t h t he mast er
merel y openi ng or closing wi t hout changi ng t he flow. When the ma r t yr
event ual l y succumbs to these harsh conditions of use it can be quickly and
easily repl aced while the mast er is closed wi t hout havi ng to isolate t he
t ransmi t t er at t he t appi ng poi nt (a process t hat ma y necessitate shut t i ng
down t he plant). Since t he mast er valve is never subjected to harsh opera-
tional conditions it should survive for an indefinite period.
Figure 8.3 shows t he act ual installation of such a t ransmi t t er on a
combi ned-cycl e power pl ant (the isolating and bl owdown valves are not
visible in t he picture, and onl y par t of t he impulse pi pework can be seen).
Similar levels of detail must be defined for all types of t ransmi t t ers and
gauges. Each of these has its own peculiarities, and neglect of a simple
r equi r ement can r ender a vital measur ement inoperative, i naccur at e or
unreliable. Again, reput abl e manuf act ur er s will be able to provi de
Figure 8.3 Installation of a pressure transmitter
Photo taken by permission of National Power plc
det ai l ed gui dance for each device. But bewar e of those who cl ai m t hat
their i nst rument s are so simple t hat no such gui dance is needed!
8. 4. 2 Act uat ors
In the chapt ers of this book deal i ng wi t h vari ous cont rol loops, reference
has been made to t he cont rol l i ng devices (such as valves and dampers)
whi ch t ransl at e t he cont rol system' s demands into changes of flow and
pressure. The modul at i on of these devices is the dut y of an act uat or, and in
the next few par agr aphs we shall briefly survey some of the act uat or t ypes
to be found in power plant.
Pneumat i c act uat ors are well established and cost-effective. In addi t i on
to bei ng reliable, accur at e and capabl e of fast response, t hey are simple to
use and mai nt ai n. The y are t herefore found in a great many installations,
and wher e t hey are used the cont rol system' s commands have to be
processed by an el ect ro-pneumat i c positioner (discussed later) or trans-
- - I Diaphragm
Position
indicator
Coupling to
valve stem
Figure 8.4 A diaphragm acuator
lated from electronic form / general l y 4 - 2 0 mA) to pneumat i c form
(usually 0.2 to 1 bar g) by an el ect r opneumat i c convert er ( I / P converter).
The simplest pneumat i c act uat or consists of a di aphr agm coupl ed to
the st em of a val ve (Figure 8.4). In the exampl e shown, air pressure
appl i ed to t he t op of the di aphr agm will cause it to be deflected and the
at t ached stem will move downwards. When the pressure is reduced the
spri ng will act to restore the stem position.
In many cases, the pressure avai l abl e from the I / P convert er will not
exert sufficient force on the di aphr agm to move it against the reactive
force appl i ed to the val ve stem by the process fluid. Unless an ext raordi na-
rily large di aphr agm is used, the 1 bar g out put signal pr oduced by
commonl y used I / P convert ers is unlikely to generat e stem forces great er
t han 20 kN. Also, I / P convert ers are designed to feed into small vol umes so
that, even if t hey coul d generat e enough pressure to move the valve
t hr ough its full travel, t hey will be unabl e to do so quickly, since it will t ake
t i me for t hem to build up sufficient pressure in the large vol ume above the
di aphr agm.
Two solutions are avai l abl e to over come this probl em: boosters and posi-
tioners. A boost er is a pneumat i c relay t hat converts small pressures to
large ones. A posi t i oner is essentially a power cont rol l er whose funct i on is
to t ransl at e the pneumat i c cont rol signals into mechani cal movement of
I P o w e r a i r
( s a y 3 b a r g )
C o m m a n d s i g n a l
( 0 . 2 t o 1 b a r g )
i L _ F e e d b a c k
C o m p a r a t o r , l i n k a g e
Figure 8. 5 Application of a positioner to a pneumatic valve acutator
the val ve stem. A posi t i oner (Fi gure 8.5) applies a high-pressure air
signal to one side or ot her of the di aphr agm, adj ust i ng t he relative
pressures until the st em position corresponds wi t h the demand signal. The
mechani cal design of the posi t i oner may i ncorporat e a cam which can
enabl e t he demand/ pos i t i on rel at i onshi p to be shaped to a linear, square-
law or ot her nonl i near characteristic.
Al t hough the use of a posi t i oner is often essential, it is not al ways
necessary. A posi t i oner cont ai ns a feedback mechani sm and difficulties can
arise because of the inclusion of an addi t i onal integral t erm into the overall
cont rol system. Thi s is part i cul arl y i mpor t ant when critical t uni ng is
requi red, since t he cont rol par amet er s of the posi t i oner are not adjust-
able.
A poi nt which can also raise difficulties is wher e mechani cal stops are
i nt r oduced to limit t he range of stem movement . In such cases it can be
ext remel y difficult to adj ust t he way in whi ch t he posi t i oner operat es at the
ext r eme limits of t he val ve travel. When the valve reaches the limit of
movement set by t he mechani cal stop this will not necessarily correspond
wi t h t he c omma nd signal. The positioner will t herefore at t empt to
el i mi nat e t he appar ent error by appl yi ng addi t i onal pressure to t he
di aphr agm. I f t he c omma nd signal holds t he valve against t he mechani cal
stop for an ext ended peri od t he vol ume of air on one side or ot her of t he
di aphr agm will build up to the full pressure of t he ai r supply. Once this has
happened any subsequent c omma nd to move t he valve of f t he end stop
will be followed by a del ay as t he air pressure wi t hi n t he di aphr agm vents
off. As this vent i ng can occur onl y t hr ough t he pipes and valves of t he
system, and as these are often constricted, t he vent period can easily
become ext ended. I n time-critical applications the results can be unaccep-
table, or even dangerous.
Wher e such mechani cal stops are fitted it is also i mpor t ant t hat t he posi-
t i oner is set up to correspond with t he act ual range of free movement of the
valve. I f one stop is set at a position t hat is, say, 10% from t he fully closed
position of t he valve and t he ot her is set at 90%, t he positioner should be
set up so t hat t he stem is moved to the 10% position when t he command
signal is at t he mi ni mum of its range (e.g. 4 mA), and to t he 90% position
when it is at t he maxi mum (e.g. 20 mA). I f this is not done and the posi-
t i oner is r anged over t he t heoret i cal stroke of t he valve stem (not bet ween
t he stops), sat urat i on will agai n occur as t he positioner at t empt s to move
t he valve past one or ot her of t he stops. However, even wi t h t he correct
settings, it will be almost impossible to avoid t he probl em of sat urat i on
since, even i f the difference bet ween t he position of the mechani cal stop
and t he correspondi ng signal is microscopically small, over an ext ended
t i me the i nt egral -act i on effect of t he positioner will cause t he air pressure
in t he di aphr agm to build up (or vent), with all t he implications al r eady
described.
I n such cases (and, in fact, wher ever a fast-responding loop al r eady
exists ar ound the valve), it may be advant ageous to abandon the positioner.
Thi s ai m will be easiest to achieve i f sufficient force can be exert ed by t he
signal-air pressure al one but, i f this is done, t he l engt h of pneumat i c
pi pework bet ween the I / P conver t er and the di aphr agm should also be
mi ni mi sed in or der to reduce t he delays t hat can otherwise be caused.
I f t he pressure available from t he I / P conver t er is i nadequat e, or i f it is
impossible to obt ai n a short pipe r un bet ween t he conver t er and t he
di aphr agm, consi derat i on should be given to using booster relays to
provi de adequat e air pressure to the di aphr agm wi t hout t he use of a posi-
tioner.
However it is done, t he poi nt is t hat , in ma ny loops, t he use of a posi-
t i oner should be a last resort r at her t han an aut omat i cal l y applied
solution.
Figure 8.6 A proprietary I / P converter
Watson Smith Ltd. Reproduced by permission
8.4.3 The I / P converter
As ment i oned above, an I / P conver t er is requi red to convert t he electronic
commands from t he DCS to a form t hat can be used by a pneumat i c
act uat or. This funct i onal i t y ma y be i ncorporat ed in t he positioner itself
(an el ect r opneumat i c positioner) or a discrete conver t er may be used
(Figure 8.6). The t ype of conver t er shown in this di agr am offers a 'fail in
position' function, meani ng t hat t he out put pressure is mai nt ai ned at t he
last good val ue when t he cont rol signal is switched of f or i nt errupt ed.
I nher ent to any I / P conver t er is t he mechani cal assembly whi ch
convert s t he electronic signal to t he pneumat i c out put . As this will consist
of some movi ng mass it will i nevi t abl y be affected by vi brat i on to a gr eat er
or lesser extent. As t he valves and damper s are necessarily mount ed on
pi pework and duct i ng whi ch is mechani cal l y coupl ed to movi ng machi nes
such as fans or pumps, t hey will be prone to vi brat e and this fact or must be
consi dered in deci di ng whet her to use an el ect r opneumat i c positioner or a
separat e convert er. The former will be exposed to t he full effects of t he
vi brat i on whi l e t he l at t er can be mount ed on a near by wall or stanchion,
offering some degree of insulation from t he worst of t he vibration. On t he
ot her hand, it is always less costly to install one device (a single I / P posi-
t i oner mount ed on t he valve) r at her t han two (an I / P convert er dri vi ng a
separat e pneumat i c positioner on t he valve). Also, t he di st ance/ vel oci t y
lags t hat affect all pneumat i c systems will be di mi ni shed i f the conver t er is
an i nt egral par t of t he act uat or.
8. 5 E l e c t r i c a c t u a t o r s
Al t hough pneumat i c act uat ors are inexpensive, reliable and fast-
operat i ng, t hei r use necessitates t he provision of compressed-ai r supplies.
The air must be cl ean and dry, entailing t he use of filters and driers. It is
t herefore at t ract i ve to consi der devices t hat do not requi re such expensive
anci l l ary plant. In addi t i on, t he compressibility of air makes it difficult to
provi de a ' dead-beat ' response when deal i ng wi t h large masses.
Wi t h t he evol ut i on of reliable solid-state position controllers for electric
motors, t he scope has opened up for avoi di ng the use of pneumat i c
operat ors by t he use of electric actuators. These are self-contained, and
only requi re an or di nar y source of electric power. The y can provide dead-
beat response and also have t he advant age of provi di ng i nher ent 'fail-fix'
oper at i on since on loss of power t hey lock in position. On the ot her hand,
maki ng an electric act uat or fail to the open or closed position on loss of
power is not so simple.
When specifying an electric
r equi r ed failure mode as well as
fully closed to fully open).
act uat or it is i mpor t ant to state the
t he operat i ng speed (time to travel from
8. 6 H y d r a u l i c a c t u a t o r s
Hydr aul i c act uat ors offer anot her way of dispensing with air compressors
and t hei r anci l l ary equi pment . This t ype of act uat or is powerful, fast and
a c c ur a t e , a n d c a n be p r o v i d e d i n f ai l - open, fai l -cl osed or fail-fix c onf i gur a -
t i ons. T h e poi nt s t o c ons i de r all c e nt r e o n t he n a t u r e o f t he hydr a ul i c
me d i u m e mp l o y e d . Is i t f l a mma b l e or cor r osi ve, wh a t pr ovi s i on is ma d e t o
g u a r d a ga i ns t l e a ka ge etc. Al so, i f for r easons o f e c o n o my a c e nt r a l i s e d
h y d r a u l i c r es er voi r is s h a r e d b e t we e n sever al a c t ua t or s , car ef ul cons i der a-
t i on mu s t be gi ve n t o e n s u r i n g t h a t no f ai l ur e c a n di s abl e ma j o r por t i ons o f
t he pl a nt .
8.7 Cabling
T h e cabl es l i nki ng t r a n s mi t t e r s a n d a c t ua t or s t o t he c ont r ol s ys t em will
be i ns t al l ed i n ar eas wh e r e t he y ma y be expos ed t o i mp a c t f r o m pas s i ng
vehi cl es or f al l i ng obj ect s. I n a ddi t i on, t he y ma y be s ubj ect t o mo v e me n t of
t he s t r uct ur e. For t hese r easons, cabl es s houl d be a d e q u a t e l y p r o t e c t e d a n d
wel l s u p p o r t e d . I t is c o mmo n pr a c t i c e i n s ome c ount r i e s t o use st eel - wi r e
a r mo u r e d cabl e t o p r o v i d e pr ot e c t i on, b u t i f a d e q u a t e me c h a n i c a l s u p p o r t
a n d p r o t e c t i o n is p r o v i d e d by o t h e r me a n s ( such as cabl e t rays) t he r e is a n
a r g u me n t i n f a vour o f us i ng c he a pe r , u n a r mo u r e d cabl e, even i n t he mo s t
sever e p l a n t e n v i r o n me n t s .
Of t e n, a boi l er is a c t ua l l y s u s p e n d e d f r om a steel f r a me t o al l ow i t t o
e x p a n d a n d c o n t r a c t as it heat s. T h e mo v e me n t b e t we e n obj ect s on t he
boi l er f r ont a n d a f i xed r ef er ence p o i n t c a n be qui t e c ons i de r a bl e a n d ,
unl ess t hi s effect is c ons i de r e d, t he cabl e c a n be d a ma g e d .
8.8 Electromagnetic compatibility
T h e h i g h vol t ages, h e a v y c u r r e n t s a n d l ar ge ma g n e t i c fields as s oci at ed
wi t h powe r - s t a t i on e q u i p me n t gi ve rise t o t he ri sk of i nt e r f e r e nc e wi t h el ec-
t r oni c syst ems. I t is a n i mp o r t a n t r e q u i r e me n t t h a t t he s ys t em de s i gne r
r ecogni ses t hi s f act a n d pays car ef ul a t t e n t i o n t o de a l i ng wi t h t he risk.
Gu i d a n c e on de s i gn a n d i ns t a l l a t i on pr a c t i c e is avai l abl e f r om sever al
s our ces [1] a n d it is n o wa d a y s ge ne r a l l y a ma n d a t o r y r e q u i r e me n t t ha t
syst ems c o mp l y wi t h e l e c t r o ma g n e t i c c o mp a t i b i l i t y ( emc) r ul es def i ned by
t he c o u n t r y i n wh i c h t he p l a n t is t o be o p e r a t e d [2].
I n ge ne r a l , a s ys t em de s i gne d t o be i mmu n e t o i nt e r f e r e nc e s houl d
e mp l o y o p t o c o u p l e d i nput s a n d o u t p u t s for di gi t al si gnal s, a n d its
a n a l o g u e i nput s s houl d i nc l ude l owpas s filters t o pr ovi de a hi gh l evel o f
a t t e n u a t i o n t o f r equenci es a bove say 20 Hz. Such filters p r o v i d e di s cr i mi -
n a t i o n agai ns t 50 or 6 0 Hz p i c k u p f r o m ma i n s - o p e r a t e d devi ces a n d
a ga i ns t a n y h i g h - f r e q u e n c y di s t ur ba nc e s t h a t ma y be g e n e r a t e d by
switchgear, variable-speed devices and t he like. I f t he roll-off charact eri s-
tics of these filters ar e adjustable, t he i nput channel s of t he system can be
made to recognise l egi t i mat e variations in t he measur ed signals (deal i ng
wi t h rapi d pressure changes or slow-changing t emperat ures, for exampl e)
while effectively i gnori ng i nt erference-i nduced signals (such as 50Hz
pickup).
8.8.1 Eart h connections
Good ear t hi ng pract i ce demands t he use of a st ar-poi nt connect i on for all
screens. Fi gure 8.7a shows how cur r ent flowing t hr ough a common
i mpedance affects ot her circuits connect ed to it. I n this exampl e, t he
voltage appear i ng at t he i nput of a device (with reference to eart h), is
given by
Vi = (il zl) + (i3 z3).
It must be recognised t hat a maj or fault in an electrical machi ne is a
t ransi ent phenomenon whi ch can result in a ver y l arge cur r ent flowing to
ear t h in a ver y short time. I n addition, l i ght ni ng strikes on st ruct ural
steelwork, cables and machi ne frames can cause current s of hundr eds of
kiloamps to flow to ground. Because of t he hi gh-frequency nat ur e of t he
cur r ent in all such cases, t he compl ex i mpedance of t he ear t h connect i on
becomes domi nant (i.e. its resistance and its i nduct ance).
I f t he device represent ed by Zl is an anal ogue i nput channel of the
pl ant DCS, it woul d nor mal l y be handl i ng 4 - 2 0 mA signals and its i nput
i mpedance woul d be, say, 250~. The voltage across this i nput channel
at full-scale woul d t herefore be expect ed to be no mor e t han 5 V
(0.02 A x 250 V).
Now, assume t hat t he common ear t h i mpedance is 10 f~ and t hat a
t ransi ent fault cur r ent (i3) of 100A flows t hr ough it. (In practice, fault
levels can be much hi gher even t han this.) The voltage occurri ng at this
i nput channel of t he DCS under these conditions woul d t herefore be
V3 = (0.02 x 250) + (100 x 10) = 1 005 V.
Thi s exampl e shows t hat i nput circuits of the DCS can be subjected to
voltages several hundr ed or t housand times hi gher t han expected, due to
such fault currents.
Fi gure 8.7b shows how this effect can be mi ni mi sed by segregating
signal ear t h connect i ons from t he ear t h connect i ons of machi nes and t hei r
switchgear. The current s flowing to t he i nst r ument ear t h will all be of t he
or der of milliamps, and i f t he connect i on to ear t h has a low i mpedance t he
ma xi mum voltage appear i ng at t he common poi nt will be no mor e t han a
few hundr ed millivolts.
. . . . . . . . . . . . , ; i - t z , t l ,
V l _ _ ~ . . . . . . . . . i "V3" -
(a)
Signal screens and instrum ent e a rth connections
,1 1
i 2
i3 i 4 l
rT
i 5
(b)
Heavy-current
e a r t h c o n n e c t i o n s
'P I,
7 7 7
Figure 8 . 7
Signal
screens and
instrument
earth
connections
'Clean earth' s t r i p in
equipment cul~cle
(insulated from cubicle
frame}
(c)
Good earthing practice
a Current flow through common impedance
b Segregation of instrument and electrical (safety) earth connections
c A practical arrangement for an instrument earth connection
Fi gure 8.7c shows how such segregat i on can be achi eved in practice. A
l ow-i mpedance busbar, i nsul at ed from t he met al wor k of t he cable itself, is
provi ded in each equi pment cubicle and connect ed to a ' clean eart h' poi nt
by a l ow-i mpedance cable. All i nst rument ear t h connect i ons (including
cable screens) wi t hi n t he cubicle are connect ed to this busbar. The
met al wor k of t he cubicle is connect ed to a separat e safety ear t h poi nt
(often t hr ough t he mount i ng bolts).
8 . 8 . 2 Cabl es: ar mour i ng, screening a n d gl ands
Reference has been made above to t he use of steel-wire ar mour i ng to
provi de mechani cal prot ect i on for i nt erconnect i ng cable, and it is
sometimes ar gued t hat ar mour i ng provides i mmuni t y to interference. This
is not totally correct. Al t hough steel-wire ar mour i ng does provi de some
degree of prot ect i on against magnet i c fields, its per f or mance as an electro-
static screen is poor. For this reason it is essential to use cable with a
br ai ded (or foil) screen, wi t h or wi t hout overall steel-wire ar mour , for all
signal connections.
Fi gure 8.8 shows how an ar mour ed, screened cable is used to connect
bet ween t he component part s of a system. It also shows how the various
conduct ors should be connect ed to eart h.
Because it is ver y difficult to spot missed, dupl i cat ed or badl y made
ear t h connect i ons once a cable installation has been compl et ed, it is vital
t hat wor k is pr oper l y supervised and very careful l y checked dur i ng t he
installation of a system. In this respect, useful assistance is provi ded in
st andards such as BS 6739:1986 ' Code of Pract i ce for i nst r ument at i on in
process cont rol systems: installation, design and practice' .
8. 9 Re l i a b i l i t y o f s y s t e ms
Because of the l arge number s of electronic component s t hat ar e manuf ac-
t ured, and because component manuf act ur er s keep good records of failure
rates etc., it is fairly easy to obt ai n statistical i nformat i on on reliability t hat
will provide a good i ndi cat i on of t he predi ct ed reliability for any given
system. I n pract i cal terms, what really mat t ers is t he l engt h of time for
whi ch a system will be capabl e of r emai ni ng in operat i on over t he course
of a year or over its operat i onal lifetime. This is governed by bot h t he relia-
bility of t he equi pment and t he speed with whi ch repairs can be effected.
For exampl e, it woul d be t heoret i cal l y possible to const ruct a very
reliable system by ar r angi ng for all functions to be performed by a few
ver y large-scale i nt egr at ed (VLSI) circuits connect ed t oget her wi t hout the
t Overall screen L - _ _
(braid o r tape) I < X
] Steel-wire armour ~ . - - - ~
Uninsulated gland
m a y b e used here
J Clean earth busbar in cubicle
(insulated from cubicle frame)
Drain wire ]
This connection i s needed o n l y where
glet'ld d o e s n o t bOnd armour t o earth
_Safety electrical e a r t h ]
: r m e d i a t e j u n c t i o n b o x I
Link incoming a n d
outgoing drain
connections
F i e l d i n s t r u m e n t ]
1
H O u t e r i n s u l a t i o n ( P V C ) t
t
S t e e l - w i r e a r m o u r
S c r e e n ( c o p p e r b r a i d ) I
I r l u e r i n s u l a t i o n (XLPE~
K E Y
Figure 8.8 Screening and earthing of cables
use of pl ug- and- socket connect or s. Because VLSI devi ces are i nher ent l y
rel i abl e a nd because connect or s ar e a source of fai l ure such a syst em woul d
offer a ver y hi gh level of rel i abi l i t y. Unf or t unat el y, it woul d be ver y
di ffi cul t t o r epai r i f it di d fail.
The reliability of any electronic system can be predi ct ed with a high
level of confidence by referri ng to statistical dat a pr oduced by manuf ac-
turers, i ndependent test laboratories or bodies such as t he defence or
nucl ear authorities*. Such dat a can be used to cal cul at e t he predi ct ed
failure rate, or mean t i me bet ween failure (MTBF), of t he system and by
using ultra-reliable component s and el i mi nat i ng all less reliable devices, it
should be possible to achieve MTBF rates of perhaps one failure in a
million hours of operat i on (i.e. one fault in j ust a over a cent ur y of
operation!). However, i f a failure di d occur in such a system, l ocat i ng its
source and repai ri ng t he fault woul d be ext r emel y t i me-consumi ng. Here,
anot her statistical cal cul at i on is used: t he mean t i me to repai r ( MTTR) .
Thi s figure is based on factors such as t he diagnostic tools available to
locate t he source of a fault, t he availability of spare parts, t he work
involved in r emovi ng t he faul t y component and t hen repl aci ng it.
A useful way of looking at t he pract i cal aspects of reliability is to
combi ne t he two factors. This leads to anot her statistic, t he system ' avail-
ability' , whi ch is a combi nat i on of t he MTBF and MT T R
availability = (MTBF x 100)/(MTBF + MTTR).
Usi ng this f or mul a shows t hat t he availability of a system wi t h a MTBF
of 80 000 hours and an ei ght -hour MT T R is 99.99%. Achi evi ng t he mean
time to repai r of ei ght hours is reasonable. This is t he time from t he fault
occurri ng, t hr ough t he process of l ocat i ng mai nt enance st aff to car r y out a
repair, t hr ough t he fault-finding process, to l ocat i ng a r epl acement
component , to installing it and rest art i ng t he system. I f t he diagnostic
tools are very powerful, enabl i ng t he l ocat i on of a fault to be quickly and
easily pin-pointed, and i f spare pri nt ed-ci rcui t cards are mount ed near by
in t he system cabinets, al r eady powered (and t herefore war med up), t hen
it may be possible to reduce t he MT T R and i f this is cut to say four hours,
t he same system will now offer an availability of 99.995%.
When eval uat i ng t he likely reliability of a system, all t hree of t he
above factors should be exami ned, together, because it may be t hat a high
level of availability is based on a less reliable confi gurat i on but an impos-
sibly short MTTR.
At first gl ance an availability of 99.98% may appear to be ver y good,
but i f this is based on a four-hour MT T R it implies t hat t he MTBF is
20 000 hours. Thi s means t hat t he system is likely to suffer failures on
about nine occasions over an operat i onal lifespan of 20 years. A system
*For example the Systems Reliability Service Data Bank of the United Kingdom Atomic Energy
Authority, AEA Technology, Thompson House, Birchwood Technology Centre, Risley, Warrington,
CheshireWA3 6AT, UK.
with t he same availability, but wi t h a mor e realistic ei ght -hour MT T R
woul d have a 40 000-hour MTBF, meani ng t hat over t he same lifetime,
t he system coul d be expect ed to fail on about four occasions.
It must be r emember ed t hat availability, MTBF and MT T R are all
statistical predictions. Not hi ng in t hem will guar ant ee t hat a system will
operat e wi t hout fault for a defined time. (In fact t he system may still go
wr ong on day 1, t hough t he likelihood t hen is t hat it should not go wr ong
agai n for a very long time, al t hough t hat may seem to be poor consolation
at t he time.)
It must also be appreci at ed t hat it is not realistically possible for these
statistics to be confi rmed by measur ement . At best, a so-called ' reliability
r un' ma y ext end for a few weeks, but this represents onl y a few hundr ed
hours of operat i on, whi ch is a small fract i on of a typical MTBF predi ct i on
(which is usually in t he or der of tens of t housands of hours). A reliability
r un will onl y show up probl ems wher e t he reliability is seriously deficient.
To realistically eval uat e a supplier' s predictions, t he best t hat can be
done is to obt ai n t he dat a on whi ch t he calculations have been based and
compar e one system wi t h anot her, while at t he same t i me asking whet her
any assumptions t hat have been made are reasonable. Beyond that, t he
desi gner should look at what is likely to happen when t he chips are down.
8.9.1 Anal ys i ng the effects o f f a i l u r e
I n t he course of designing a cont rol loop careful t hought must be applied
to t he effects of failure of any component . I f any risk can be posed by such
a failure, precaut i ons must be t aken to limit its effects. Such considerations
must be applied to t ransmi t t ers, process switches and actuators, as well as
to t he DCS itself. It will usually be necessary to have t he design confi rmed
by some form of risk-assessment pr ocedur e such as a HAZOP (hazards
and operabi l i t y study) [3].
The HAZOP pr ocedur e has t radi t i onal l y been applied by considering
t he results of failure of each and every i t em on t he plant. One of t he
approaches t hat is adopt ed is for a t eam from each discipline to look at
each i t em and ask a series of questions such as, for a valve: what happens i f
it opens, shuts or locks in position? Ot herwi se, t he questions may be ai med
at assessing t he effects of mor e or less pressure or t emper at ur e on t he
device in question. The HAZOP pr ocedur e is very specialised, and t he
audi t of t he pl ant is usually conduct ed dur i ng t he design stages of t he
proj ect by a t eam of process engineers, cont rol engineers and others, t he
whol e bei ng co-ordi nat ed by a specialist organisation.
The emer gence of pr ogr ammabl e systems has raised several questions
as to t he val i di t y of this t ype of study. For exampl e, a t radi t i onal HAZOP
may l ead to t he conclusion t hat i f one valve fails to open the si t uat i on
may be deal t wi t h by the openi ng of anot her valve or by the t ri ppi ng of a
pump (either act i on bei ng initiated by the human oper at or or by a safety
interlock system). However , wi t h a pr ogr ammabl e system it may be
necessary to consi der the possibility of a failure in the DCS causing
mul t i pl e failures to occur at the same instant, while at the same time any
correct i ve act i on t hat the oper at or may wish to take, and the prot ect i ve
systems themselves, are di sabl ed or seriously i mpai red. Such questions
have recent l y been addressed, and t he mat t er must be consi dered in t he
light of the new guidelines [4].
The following provi des an overvi ew of some of the sat~ty-related
mat t ers t hat will need to be consi dered dur i ng the design procedure.
8. 9. 1. 1 Fai l ure modes in electronic systems
Oversi mpl i st i c appr oaches are somet i mes adopt ed t owards the analysis of
failure, and the field of industrial control-systems design is littered wi t h ill-
consi dered ideas. One of the most not abl e of these is t hat ' an electronic
signal will al ways fall t o zero under a fault condition' . Worse still is the
t heor y t hat this is the reason for the use of the ' live zero' (e.g. 4 mA) in
signals such as t he well established 4 - 20 mA range.
Though an electronic signal may well fall to zero under cert ai n condi -
tions (e.g. br eakage of the connect i on bet ween the t ransmi t t er and the
receiver), it is as likely to rise to 20 mA (or above), to lock at an inter-
medi at e val ue (even t hough the measur ed par amet er has changed), or to
slowly drift away from the correct value. I f the signal source is pr ovi ded
wi t h self-diagnostic facilities (as descri bed in the following section), the
out put can be confi gured to rapi dl y change to a high or low value, but this
is the onl y condi t i on where such an act i on can be assured.
As for the reason behi nd the use of a live zero, this has little or not hi ng
to do wi t h failure. The real reason is t hat the live zero provides a mi ni mum
cur r ent for the signal source, enabl i ng the devi ce to be power ed from the
receiver. Thi s is known as ' t wo-wi re transmission' since it eliminates the
need to provi de separat e conduct ors to power the transmitter. Fi gure 8.9a
shows a t r ansmi t t er worki ng on a ' dead-zero' range such as the old 0 -
10 mA st andard. Thi s requires the provision of power-suppl y connect i ons
to the device, so t hat four conduct ors are needed for each transmitter.
Wi t h a live zero (Figure 8.9b) the t ransmi t t er electronics can be power ed
from the receiver, and onl y t wo conduct ors are needed.
Gi ven t hat a process t ransmi t t er is capabl e of det ect i ng probl ems
wi t hi n itself and war ni ng of this, so t hat the DCS can t ake the appr opr i at e
action, what can be done about the signals t ransmi t t ed by the DCS to
" - - ' [ . . . . . . . . . I
( a )
Control equipment practice 17 7
[ . . . .
i i
I I
i i
i T '
I 0 - 1 0 m A i
I I
I I
I T r a n s m i t t e r I
I I
I . . . . . . . . . . . . . . . . . i
I . . . . . . . . . . . . . . . . . . . . . . . . . -
I o
i
i 0
i I
I
I
I
i
I
P o w e r s u p p l y
(A C o r D C )
I
I
I
I
I
i
I R e c e i v e r
i
i
R L
I . . . . . . . . . . . . . . . . . . . . . . .
4 -2 o m A T f
T r a n s m i t t e r
i I
I I
I I
I I
I
I R L
I I
I I
i i
1 ( 9 ~ i
i i
/
I
I 2 4 v D C
i o o w e r suD olv J I
i i
I R e c e i v e r I
i i
( b )
Figure 8.9 The advantages of two-wire transmission
a Tr ansmi ssi on wi t hout ' l i ve zer o'
b Two- wi r e t ransmi ssi on (wi t h ' l i ve zer o' /
178 Power~Olant control and instrumentation
act uat ors? A moder n DCS will i ncor por at e powerful diagnostics facilities
and wat chdogs t hat will war n of incipient or act ual problems. But the
act i on t hat can be t aken in response to these warni ngs is mor e difficult to
det ermi ne. I f the entire DCS is failing, t hen it may be necessary to shut
down the pl ant . On the ot her hand, i f onl y one out put channel has failed it
may be possible t o overri de the commands or oper at e emergency devices
t hat bypass the fault.
V t = 10
. L
<
I
< V a = 0 . 0 9 9 V > 1
I I
R s = 1 0
' ~ R E = 1 k . Q V L = 9 . 9 0 1 V ~
/
I 9 . 9 rnA
( a )
Transmitter
<
I t = 12 rnA
RL
Figure 8.10 Current vs voltage transmission
a Voltage transmission
b Current transmission
( b )
8.9.1.2 Voltage and current transmission
The use of cur r ent r at her t han voltage signals for t ransmi t t ers ma y
requi re expl anat i on. I f a t r ansmi t t er generat es a voltage signal (Figure
8.10a), t he voltage t hat appears at t he receiver will not be t he same as t hat
at t he t ransmi t t er because of t he voltage loss across t he series resistance of
t he connect i ng cables. I n t he exampl e shown, t he recei ver i nput resistance
is 1 000 ~, and t he total cable resistance is 10 ~, so t he cur r ent t hat circu-
lates when t he t r ansmi t t ed voltage is 10V is 10/1 010A, or 9. 9mA. The
voltage dr op across t he 10 ~ series resistance causes a voltage dr op of
99 mV, so t he voltage appear i ng at t he receiver is 9.901 V. I n this exampl e
t he signal appear i ng at t he receiver is roughl y 10% below t he t ransmi t t er
out put .
The situation is made worse because t he cable resistance is not fixed. It
will var y wi t h t emper at ur e and will also be affected by factors such as t he
cont act resistance of t he t ermi nal s and any switches in t he circuit.
The effects can be mi ni mi sed by using a receiver with a hi gh i nput resis-
tance, but however hi gh t he resistance is made, some small cur r ent flow
will always exist, and t herefore some errors will be i nt roduced.
Wi t h a cur r ent system (Figure &10b), t he cur r ent sent out by t he trans-
mi t t er is by definition identical to t he cur r ent flowing t hr ough t he recei ver
terminals. The only possible source of error woul d be t he format i on of a
parallel pat h across t he receiver, whi ch woul d divert some of t he cur r ent
flow, but this is very easy to avoid.
Fi gure 8.9 compares t he circuit of a voltage-based signalling system
with a current -based equivalent. It will be seen t hat t he voltage-based
system requires addi t i onal connect i ng wires for t he power supply,
i mpl yi ng t he need for four wires, whereas a current -based system requires
only two. (In figure 8.9b t he power supply is connect ed bet ween t he
negat i ve end of t he load and t he positive t ermi nal of t he t ransmi t t er.
However, t he power unit coul d equal l y be moved to t he ot her side of t he
circuit, whi ch is preferable wi t h some types of circuit because of t he
gr oundi ng ar r angement s of t he various devices. Each system must be
carefully exami ned in order to verify t he opt i mum met hod of connec-
tion.)
A f ur t her benefit of a cur r ent transmission signal is t hat its sensitivity
to i nt erference is less t han t hat of a voltage system. I n a cur r ent system, i f
i nt erference causes t he cur r ent to increase, t he t r ansmi t t er will quickly
self-regulate and restore t he cur r ent to t he correct value. Wi t h a voltage
system, a voltage i nduced by i nt erference and t hat gener at ed by the trans-
mi t t er will be summed at t he recei ver i nput terminals. The t r ansmi t t er will
not be aware of this, and so cannot correct for it.
8.9.1.3 Transmitters
Modern transmitters are frequently able to run self-diagnostics routines
that provide a high level of confidence that a failure will be detected. In
general these routines are arranged to drive the transmitted signal to one
extreme or other (usually outside the normal operating range of the instru-
ment). The direction in which the signal is driven must be defined when
the instrument is first specified.
Thus a 4- 20mA transmitter can be specified to drive the signal to
under 4 mA or above 20 mA when a fault is detected. The DCS must then
be contigured to raise an alarm and take the necessary action to protect
the plant when such an event is detected. Depending on the hazards which
may arise, the required action may be to 'freeze' the relevant system
output command at the value it had held prior to the fault being detected,
or it may be to open or close an isolating valve, or it may be to operate a
separate emergency shut-down system.
The fault-detection process will drive the signal to the selected point
virtually instantaneously, and if the DCS is configured to respond quickly
to such an event the necessary action will usually be taken before any
hazard can arise.
If the nature of the process or its instrumentation is such that it is not
possible to provide adequate protection by the means outlined above, some
other form of protection will be needed. This may take the form of a mea-
surement and control system that monitors the process by an entirely
separate set of instruments and takes action if a dangerous discrepancy
arises between the two sets of measurements, or it may be a sophisticated
two-out-of-three (2oo3) voting system. Various guidelines are available for
making such decisions [3,5], and these should be consulted to justify the
cost of providing adequate levels of supervision and back-up. It is unlikely
that any authority would be able to supervise each step in the design, con-
struction and installation of a power-station DCS. But once the plant has
commenced operation, if an incident does occur and results in damage to
the plant and/ or injury to personnel, the entire design and construction
process will be very closely examined. The consequences will be severe for
all involved if at that stage it is not possible to demonstrate compliance
with the various guidelines and standards.
8.10 Summary
Armed with an understanding of the plant and of its control and instru-
mentation systems and equipment, we can now move on to look at how the
operational requirements are defined on a project, and some aspects of
how equipment on a plant is identified.
1 BULL, J. H: 'Code of practice for the avoidance of electrical interference
in electronic instrumentation systems'. ERA 75-31, ERA Technology Ltd.,
Leatherhead, Surrey, UK
2 'Electromagnetic compatibility for industrial process measurement and
control equipment'. IEC Publication 801-I. International Electro-technical
Commission, Geneva, 1987
3 'A guide to hazard and operability studies' Chemical Industry Safety and
Health Council of the Chemical Industries Association, London, 1992.
4 MINISTRY OF DEFENCE DIRECTORATE OF STANDARD-
IZATION: Draft Standard 00-58: a guide for HAZOP studies on systems
which include a programmable electronic system'. (Ministry of Defence,
Glasgow, UK, 1995).
5 HEALTH AND SAFETY EXECUTIVE: 'Programmable electronic
systems in safety-related applications'. (ISBN 0 11 883913 6 ISBN 0 11
883906 3) (HMSO, London, 1987)
Chapter 9
Requi rement s definition and
equi pment nomencl ature
9.1 Overvi ew
The provision of a C&I system for a power station is a compl ex mat t er
whi ch requi res careful and comprehensi ve admi ni st rat i on. The task is
demandi ng: t he design of t he equi pment must be correct , systems must be
desi gned on t i me, equi pment has to be careful l y specified and purchased,
and ever yt hi ng has to be del i vered to site, installed and commi ssi oned to a
t i ght pr ogr amme whi ch i nt erweaves t he C&I system wi t h t he ma ny ot her
activities t hat will i nevi t abl y be t aki ng pl ace on site at t he same time.
Whe n compl et e t he system shoul d be fully support ed by comprehensi ve
document at i on whi ch enables mai nt enance st aff and users to deal wi t h it.
The act ual process of designing t he C&I systems forms onl y one par t
of t he many activities t hat go t oget her in t he task of engi neeri ng a
compl et e cont ract . Al t hough some of t he ot her operat i ons ma y seem
munda ne and trivial, t hey are real l y anyt hi ng but that. The y are as
essential to t he cont r act as t he t echni cal design work.
9.2 Defining the requirements
The out come of a maj or engi neeri ng proj ect such as t he design, installation
and commi ssi oni ng of t he C&I systems of a power station will to a l arge
ext ent depend on specifying t he r equi r ement s at t he outset so t hat what is
provi ded fits wi t hi n t he budget and is fit for t he i nt ended purpose. Equal l y
exact document at i on is r equi r ed at each stage as t he system met amor -
phosises f r om t he ori gi nal concept to t he final, funct i onal form.
Thi s pr ocedur e requires a considerable amount of definition, and the
following outline lists the document s t hat mi ght be requi red over t he
lifetime of a typical project, listed in the or der in which t hey may be
expect ed to be generat ed. This does not pret end to be an absolute defini-
tion t hat must be rigorously followed on every installation. It is a pract i cal
system t hat has produced good results when tollowed on several projects.
Ot her document at i on systems offered by a system vendor may be perfectly
acceptable, provi ded t hat t he same degree of definition is achi eved at each
stage.
9.2.1 The f unct i onal specification
A system can be defined in several different ways, but an essential require-
ment is t he definition of what t he system should do in relation to the
process it is moni t ori ng a nd/ or controlling. This r equi r ement is met by the
Functional Specification (FS). This is a process-related definition of t he
functions t hat t he system will be requi red to perform. It does not provide
det ai l ed descriptions of the system har dwar e and software, such as
response times, power supplies, envi r onment etc., except where these are
critical to meet i ng the functional objectives of t he installation. Because it
defines t he requi rement s, t he FS should be one of t he document s against
whi ch the vendor is invited to submit a commerci al bid for a project.
A typical FS will describe t he pl ant as a whole, and then discuss the
cont rol loops with the requi red accuracy, response times and dynami c
range of each loop.
9. 2. 2 The technical specification
Havi ng said what t he system should do, it is t hen appropri at e to define, in
some detail, how the functions should be achieved. This purpose is served
by a Techni cal Specification (TS) whi ch describes t he system confi gurat i on
in t erms of t he electronics and comput er technologies to be employed.
Because it defines t he t echnol ogy and facilities requi red, this document
should be requested from t he vendor when he submits a proposal for
execut i ng a project.
The TS should include t he tbllowing definitions:
t he size, t ype and number of oper at or displays to be provided;
t he overall confi gurat i on (heirarchical, distributed etc.);
met hod of pr ogr ammi ng (block-structured, Boolean, ladder-logic etc.);
t he physical envi r onment in which t he equi pment will be expect ed to
operat e (defined in t erms of t emper at ur e, humi di t y, vibration, shock
etc. as well as dust levels to be encount ered, hazar dous- ar ea require-
ment s etc.);
Requirements definition and equipment nomenclature 185
the per f or mance requi red (speed, response time, availability, memor y
etc.);
power supplies available (including voltage and frequency excursions);
safety requi rement s (backup, r edundancy, fault t ol erance etc.).
The TS will also lay down t he requi rement s for testing, such as:
Fact ory-accept ance: where the system is set up in t he supplier' s
premises, connect ed to a simulator, or to switches and signal sources for
inputs and to met ers and l amps (or LEDs) to show outputs, and t hen
put t hr ough a series of routines to show t hat it performs as requi red.
Si t e-accept ance tests: performed after the system has been installed
and commi ssi oned on site, when it is subjected to changes in desired-
value settings, si mul at ed equi pment trips and so on, to prove t hat it
reacts correct l y and in good time to such events.
Reliability run: wher e t he equi pment is left in full control of t he
process, to demonst r at e t hat is is capabl e of operat i ng correct l y for a
defined period, with no malfunctions. The document should state t he
dur at i on of the reliability run, t he conditions under whi ch this test will
be expect ed to operat e and what should happen if a failure occurs (e.g.
start agai n and r epeat the test).
Any commer ci al requi rement s rel at i ng to guarant ees, per f or mance
bonds etc., should be defined separately, al t hough these will i nt errel at e
with t he TS and should t herefore be referred to.
Note t hat t he various accept ance test procedures will at this stage be
defined onl y in ver y general terms. A full definition is provi ded by test spe-
cifications (as described later).
9.2.2.1 Maki ng provision f or site tests
It is usual to ret ai n a sum (typically 5% of t he overall cont r act value)
whi ch is pai d onl y when the vendor has proved t hat t he system is capabl e
of performi ng satisfactorily. However, it should be recognised t hat a
control-system supplier can demonst r at e t hat his equi pment and systems
are capabl e of funct i oni ng as requi red only i f t he pl ant is made available to
hi m for testing t he system' s per f or mance on t he operat i ng plant. It is
unreasonabl e to ret ai n what may in fact be a substantial sum of money
wi t hout giving t he supplier a reasonable oppor t uni t y to prove t hat his
equi pment is as accurat e, fast-responding and reliable as required. Yet, it is
quite common for t he pl ant owner to procrast i nat e over performi ng such
tests. The reason for this is t hat a power pl ant represents a maj or invest-
ment , and st art i ng t he recovery of t hat i nvest ment must nat ural l y
commence as quickly as possible. As soon as the pl ant has been compl et ed
to t he poi nt wher e it can st art ear ni ng its keep, st rong commer ci al
pressure comes into play, requi ri ng t he pl ant to operat e at ma xi mum
out put . The pressure will be to st art ear ni ng revenue as quickly as possible
and for as l ong as possible, to maxi mi se earnings. Reduci ng t he out put of
t he pl ant for t he purpose of car r yi ng out tests is t herefore unpal at abl e.
It is in everybody' s interests t hat this di l emma is recognised and a
suitable form of words devel oped to cover t he situation, yet this is not oft en
done. One solution is to rel at e t he financial ret ent i on to t he test
pr ogr amme, defi ni ng when t he tests will be per f or med and addi ng words
to t he effect t hat i f t he tests ar e not car r i ed out wi t hi n a defined t i me aft er
commi ssi oni ng has been compl et ed, per f or mance retentions will be
released provi ded t hat t he del ay is for commer ci al reasons outside t he
cont rol of t he C&I supplier.
9. 3 Th e KKS e q u i p me n t i de nt i f i c a t i o n s y s t e m
Each i t em of equi pment on a power-pl ant site has to be identified by a
met hod whi ch enables it to be uni quel y defined, specified, purchased,
installed, commi ssi oned and mai nt ai ned, and this requires some logical
way of number i ng equi pment .
Al t hough several systems of nomencl at ur e can be identified, two
met hods are in wi despread use: one Amer i can t he ot her Eur opean. The
l at t er is knows as t he KKS ( Kr af t wer k Kennzei chen system t ransl at ed as
power station desi gnat i on system), whi ch was devel oped by a consort i um
of l arge manuf act ur er s and users under t he auspices of t he Ge r ma n VGB
Techni cal Commi t t ee on t echni cal classification systems [1].
KKS nomencl at ur e is ext r emel y comprehensi ve, and once it has been
underst ood t he system provides a ver y useful met hod of identifying any
piece of equi pment and its operat i onal role in a plant. The system has its
weaknesses, but it is so wi del y used, bot h in Eur ope and wher ever
Eur opean influences are felt, t hat it is ver y i mpor t ant for t he cont rol
system desi gner to have at least a r udi ment ar y under st andi ng of its
operat i on.
KKS defines ever yt hi ng on a plant, from the smallest electronic
component to t he largest t ur bo- al t er nat or and even covers t he buildings
t hat cont ai n it all.
Unf or t unat el y much of t he Engl i sh-l anguage expl anat or y document a-
tion t hat is available has been t ransl at ed from t he original Ge r ma n in such
a way t hat t he meani ng is sometimes obscure, and fine levels of classifica-
tion have oft en become bl urred. (For exampl e, t he distinction bet ween
' anci l l i ary systems' and ' auxi l i ary systems' ma y not be i mmedi at el y
obvious.) The system also uses al phabet i c charact ers whi ch do not readi l y
rel at e to Engl i sh-l anguage names of t he equi pment or function. For
exampl e, valves ar e prefixed AA, and t he desi gnat i on l et t er C is used to
identify t he purpose of an i nst rument . Thi s can add to t he confusion, since
ma ny engineers rel at e C to a controller. Fur t her confusion can be caused
because KKS uses t he letters DP to i ndi cat e a pressure i nst r ument used in
conj unct i on wi t h a cont rol system or DCS, when DP is usually t aken as
referri ng to differential pressure.
Those who are used to t he Amer i can system of nomencl at ur e ma y find
it confusing t hat t he KKS codes for all equi pment ar e det er mi ned by t he
funct i onal ar ea of t he pl ant to which the equipment relates. For exampl e,
in a st eam- t emper at ur e loop wher e t he measur ed vari abl e is t he t emper a-
t ure of t he st eam at t he final superheat er outlet, t he t r ansmi t t er will be
have a KKS code begi nni ng wi t h LBA, but t he spray-wat er cont rol valve,
as par t of t he feed-wat er system, will be given a code begi nni ng wi t h LAA.
This is ver y di fferent from t he Amer i can system, whi ch allocates a con-
zsistent ar ea code to the whole loop, and this can be confusing. It is
part i cul arl y i mpor t ant t hat this distinction should be cl earl y underst ood
by pl ant personnel, since wi t h t he KKS nomencl at ur e t he t ag number
al l ocat ed to t he valve mer el y indicates t hat it is handl i ng wat er, and gives
no i ndi cat i on t hat t he valve is cont rol l i ng t he st eam t emperat ure. Critics of
t he KKS system use this as evi dence against t he system, stating t hat when
t he compl et ed installation is in place, t he nat ur e of t he medi um bei ng
handl ed by t he valve is fairly obvious (because of t he size, nat ur e or
col our-codi ng of t he pipework), whereas t he effect of t he valve is not. Wi t h
t he ol der system, t he fact t hat it is a spray-wat er valve is appar ent from t he
t ag number.
The differences are illustrated in Figure 9.1, whi ch shows how t he t wo
systems are applied to t he same t ype of cont rol loop. It will be appreci at ed
t hat , on t he act ual plant, t he t ag number al l ocat ed to this valve will
mer el y identify it as bei ng par t of t he feed-wat er system, and it will not be
appar ent t hat its cont rol l i ng effect is on t he st eam t emper at ur e.
9.3.1 The importance o f agreement and co-ordination
The KKS system allows di fferent users to adopt di fferent approaches,
stressing t hat all parties involved on t he proj ect should agree on t he
selected pat h to be followed. Al t hough this does show a degree of flexibility,
on a given proj ect it becomes confusing to those who were not par t y to t he
decisions t hat were originally made. I f onl y for this reason, it is i mpor t ant
t hat an agr eement on t he number i ng met hodol ogy is r eached as soon as
work commences on a project. It is equal l y i mpor t ant t hat a single individ-
ISA system
KKS
Figure 9.1
T CV - 2 LAA30AA004
Comparison of KKS and ISA numbering systems applied to a steam-tempera-
ture control loop
ual is appointed to control and co-ordinate the allocation of numbers.
This will avoid duplication and different symbols being applied to instru-
ments that serve the same sort of function.
It is also i mport ant to understand that the KKS code for an item is
almost invariably dictated by the functional area of the plant on which it is
used (an except i on is the DCS itself, since this general l y serves al most all
areas of the plant). Therefore, before a number can be appl i ed to, say, a
t emper at ur e t ransmi t t er, it is necessary to know the KKS code for t he pi pe
to which it is fitted. Thi s dependence renews the pressure for the pl ant
pi pi ng and i nst rument at i on di agr ams ( P&I Ds) to be defined at an earl y
stage in a proj ect and for t hem to be as accur at e and compl et e as possible
before work begins on specifying t he i nst rument at i on and systems. Thi s
poi nt cannot be overemphasi sed. Probl ems will i nevi t abl y arise as soon as
any at t empt is made to allocate KKS numbers to i nst rument at i on before
the P&I D stage has been compl et ed.
It must be recognised t hat t he KKS system offers a comprehensi ve
met hod of housekeepi ng which can great l y enhance the task of al l ocat i ng
nomencl at ur e to all equi pment on a plant. Thi s is a routine, but surpris-
ingly costly, facet of the admi ni st rat i on of a project. Properl y used, KKS
can yield consi derabl e economi es and efficiency: poorl y admi ni st rat ed it
becomes a ni ght mare.
9. 3. 2 Revi ew o f K K S
The following is an at t empt to expl ai n how KKS identifies the component s
t hat go to make up the C&I systems (and only the C&I systems) of a power
plant. The st ruct ure of this book does not allow for a comprehensi ve gui de
to be provi ded, but the following bri ef out l i ne shoul d at least enabl e the
r eader to grasp the principles.
9.3.2.1 The levels o f coding
The number al l ocat ed by the KKS system to a piece of equi pment is
broken down into a number of sections or ' levels' (see Figures 9.2 and 9.3)
and wi t hi n each of these levels is a field or set of fields, each field being
occupi ed by a l et t er or number. Each l et t er or number is al l ocat ed a field
identification: G for the plant, F for the function, A for the equi pment and
B for the component .
The Level 0 code is used where there are, for exampl e, t wo power
pl ant s on the same site. These are usually desi gnat ed as 'A' and 'B', but
the system allows numeri c charact ers to be used if desired. I f onl y one
pl ant exists on the site the first char act er is omi t t ed altogether.
The first digit of t he Level 1 code identifies the boi l er / t ur bi ne uni t on
whi ch the rel evant piece of equi pment is fitted. Thi s is al ways a numeri c
char act er and is usual l y 1 for Boiler 1 etc. Wher e the equi pment is
common to all areas of the pl ant (e.g. a cool i ng tower) the number 0 is
al l ocat ed to it.
I I I
' i i
I
~ i I . m ~ 1: FUNCTION i L e w l 2 EQUIPMENT i L ~ d 3 : COMPONENT
F ' ~ ' ~ " ~ . i F . , . , . v. ~. j , ~ ~ . . . . . * . j ~ e , , ' - B . - ' - . . . . .
Vl'l'lcl l"l'/l' l'lll'll 1'1
I P T / /
~---r ~ ' v
Cubicle No. S e q u e n t i a l
ID No.
I Control system
, w o r s , , . m I
S h e l f No.
. . A I C K N I 2 D A 0 5 I
Figure 9.2 Application of the KKS numbering system to DCS equipment
The r emai ni ng par t of t he Level 1 code defines t he funct i on of t he
equi pment (e.g. a feed-wat er pump system, compri si ng t he pump itself,
its mot or and all equi pment associated wi t h it, such as l eak-off valve,
st art er etc.), t oget her wi t h a uni que sequential number for t he system
(e.g. pump system no 2). Thi s group of digits (the first t hree al phabet i c,
t he next two numeri c) is combi ned wi t h t he uni t -i dent i fyi ng prefix, into
what is called t he ' funct i on code' .
The Level 2 code (which ma y compri se five or mor e digits) defines t he
par t i cul ar pi ece of equi pment (such as a pressure t ransmi t t er) and its
uni que sequential number (e.g. 101).
The Level 3 ( component or signal) code identifies t he component
itself, and in t he case of a device t hat generat es some form of electrical
signal, it also defines t he nat ur e of t hat signal.
! ! ! ! ! !
I I I I I L e v e l 3 I
J q " J i L e v e l 1 i L e v e l 2 i i
, ! ~ . I ~d i F U N C T I O N i E Q U I P M E N T U N I T C O M P O N E N T o r i
,. ~ i A =,.! A ~ i A S I G N A L ~ i
! = - ! : ~ ! - " " ! " " ! " " i
i l l ' '
~ . , , . I , l o l , l o o ] , l , l o l , l x i l o l , I
PoMUon code G F. ~ ~ ~ Fm r T A~ ~ Am Am ~ I k I b IIm Bw
Typeof z N* A k A N N A k N N N A A . N N
. _ . ~- _ . . _ , : _ . . . . ~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
A = a l p h a b e t i c c h a r a c t e r
N = n u m e r i c c h a r a c t e r
* = o p t i o n a l ( m a y b e o m i t t e d )
B n ~ k d o w n o f t h e c o d e
- - I P ,.,,,.o . I
/ - I o " " " u m ~ " I
/ / - i I
/ - I
" - ] ' J g i ' J l l P ] ' ] ] ' I x J Q l o J ' j
\ v /
\ /
S e q u e n t i a l n u m b e r s
C o m p o n e n t o r
s i g n a l c o d e
Figure 9.3 Overall KKS numbering philosophy
It is i mpor t ant to under st and t hat i f there are, for exampl e, 200
pressure gauges on a pl ant , the sequential numbers for these do not st art at
001 and cont i nue t hr ough to 200. The number s relate to the funct i onal
area of the pl ant on whi ch t hey are used. Therefore, t hey start afresh each
t i me t he precedi ng code changes. For exampl e the first pressure gauge on
the HP st eam pi pi ng system of a pl ant (where the pi pi ng system is
number ed LBA10) will be al l ocat ed a sequential number LBAIOCP001,
t he next will be LBA10CP002, and so on. The number i ng starts again on
anot her system: the first pressure gauge on the hot reheat st eam piping
system ( number ed LBB10) being al l ocat ed the sequent i al number
LBB10CP001, t he next bei ng LBB10CP002 and so on.
The st ruct ure and appl i cat i on of this codi ng system is illustrated in
Figures 9.2 and 9.3.
9.3.2.2 Explanation o f the main coding principles
Ther e are several hundr ed categories within the KKS codes but the
following is a bri ef s ummar y which will provi de some assistance in deal i ng
wi t h boiler C&I systems. Wher e it is felt t hat the wor di ng is ambi guous,
some exampl es have been provi ded. However , it should be r emember ed
t hat these i nt erpret at i ons are not universally i mpl ement ed, and the i mpor-
t ance of project-specific agr eement and co-ordi nat i on (as ment i oned in
Sect i on 9.3.1) cannot be t oo st rongl y emphasised.
I,evel
A
B
C, D
E
F
G
H
J
K
L
M
N
P
q
R
S
U
1: Function
= grid and distribution systems
= power transmission and auxiliary power supply
= i nst rument at i on and control equi pment
= conventional fuels supply and residue disposal
=- handling of nuclear equi pment
= wat er supply and disposal
= conventional (i.e. nonnuclear) heat generation
= nuclear heat generation
= reactor auxiliary systems
= steam, wat er and gas cycles
= mai n machi ne sets
= process energy (e.g. district heating)
= cooling-water systems
= auxiliary systems (e.g. air compressors)
= gas generation and t reat ment
= ancillary systems (e.g. heating and ventilation)
= structures
W = renewable energy plants
X = heavy machi ner y
Z = workshop and office equipment
Level 2: Equi pment unit
A = mechanical equi pment (e.g. valves, dampers, fans, including
actuators)
B = mechani cal equi pment (e.g. storage tanks)
C = direct measuri ng circuits (e.g. local indicators)
D = closed-loop control circuits
E = anal ogue and digital signal conditioning
F = indirect measuri ng circuits (e.g. sensors feeding remot e indicators)
G = electrical equi pment (e.g. cubicles, j unct i on boxes, generators,
inverters, batteries, lightning-protection system)
= subassemblies of mai n and heavy machi nery (e.g. bearings)
= nuclear assemblies (e.g. absorbers, moderators, shielding equipment)
H
J
9.3.3 An example o f how the codes are used
Fi gure 9.4 shows how t he KKS codes may be applied to a par t of t he
feed-wat er pumpi ng system on a plant. In this exampl e, t he pump relates
to Boiler 1 and t he pressure i ndi cat or at its di scharge is t herefore number ed
1LAB25CP001, t he C being t he Level 2 code for a di rect -measuri ng device
(C) i ndi cat i ng pressure (P). The feed-flow t ransmi t t er forms par t of t he
t hree-el ement feed-wat er cont rol system and its Level 2 code t herefore uses
D for a closed-loop cont rol circuit, and F for flow.
When it comes to t ransmi t t ers whose signals do not form par t of a
closed-loop cont rol system, di fferent i nt erpret at i ons are applied by various
users to t he Level 2 codes. I n some cases t he classification l et t er D is
applied to any t ransmi t t er which feeds t he DCS, irrespective of whet her or
not t he signal is used in a closed-loop cont rol system. Ot her users ext end
t he meani ng of F (indirect measur i ng circuits) to include this t ype of mea-
surement . In Fi gure 9.4, t he former of these i nt erpret at i ons is used, and t he
t emper at ur e t ransmi t t er that is used to provi de a signal for t he oper at or
display, is t herefore al l ocat ed t he l et t er D as t he first char act er of its Level
2 code (the second char act er is T, for t emperat ure).
This poi nt illustrates t he i mpor t ance of co-ordi nat i on and agr eement
bet ween all parties ment i oned at t he begi nni ng of Section 9.3.1.
When t he Level 3 component code is considered, the KKS classification
serves two functions. One identifies the component itself, for exampl e
OT01 for a t hermowel l , while the ot her designates the signal. The signal
codes ar e broken down into classifications whi ch identify whet her t hey are
1
9
4

P
o
w
e
r
-
p
l
a
n
t

c
o
n
t
r
o
l

a
n
d

i
n
s
t
r
u
m
e
n
t
a
t
i
o
n

O
0
0
~
W
I

I
.

~

o

I
.

I
.
O
~
-
J
T
.
B
~
r
l
L

1

~

o

~
_

_
_
_

~

~

"
~
.

{

I

[

e
~

Table 9.1 Some KKS Level 3 codes f or DCS inputs and outputs
Function Field identification Class of signal Si gnal
B1 B2 BN
B 11 Binary-MCC Alarm
91 Binary-MCC Other
31 Process switch LOW limit switch 1
32 Process switch LOW limit switch 2
Input to DCS X G 33 Process switch LOW limit switch 3
41 Process switch HI GH limit switch 1
91 Process switch Other
01 Analogue input Resistance thermometer
element
Q 11 Analogue input Transmitter
91 Analogue input Other
B 01-05 Binary-MCC Vari ous START commands
09 Binary-MCC STOP
G 01 Binary output Sol enoi d valve
Output from Y 91 Binary output
DCS
11 Analogue output Speed control
Q 21 Analogue output Position control
91 Analogue output Other
t r ansmi t t ed to or from t he DCS. The y can also show t he funct i on of t he
signal. Tabl e 9.1 gives some exampl es to show how this is achieved.
I n general , an X in t he B1 field indicates an i nput to t he DCS and Y
shows an out put from it. I n t he B2 field, B indicates a bi nar y signal asso-
ci at ed wi t h mot or controllers; G is for bi nar y signals such as limit switches,
pressure switches, solenoid valves and so on; Oi s for anal ogue signals.
By careful use of these identifications it is possible to economise on t he
work of defi ni ng t he functions of various signals, since each funct i on will
fall into a defined group, for whi ch t he actions and connect i ons are al r eady
defined.
9.4 Summary
This al most compl et es this st udy of boiler C&I systems. However,
updat i ng and refurbi shment needs to be discussed. Thi s is a subject t hat
has become i mpor t ant as the pace of electronics devel opment has acceler-
ated. As t here is little sign of this t rend levelling out, we must look at its
implications for power - pl ant management , users and mai nt enance staff.
DIN 40719 ' Circuit diagrams, circuit
classification (diagrams, charts, tables, item
classification examples, 1978
representations, definitions,
designations) - - alphabetical
Chapter 10
Upgradi ng and ref urbi shment s ys t ems
Even the newest and most advanced of cont rol systems event ual l y
becomes dat ed. Up to about t he mi ddl e of the t went i et h cent ury, C&I
systems were mai nl y pneumat i c, the pace of devel opment was slow and it
was general l y possible to keep a system in good funct i onal order, however
old it may have been, si mpl y by repai ri ng the equi pment or, i f the worst
came to the worst, buyi ng r epl acement component s from the original man-
ufacturer. (In those heady days manuf act ur er s coul d be relied on to
r emai n in business for mor e t han a few years, and to be abl e and willing to
provi de r epl acement component s at reasonabl e cost for as l ong as the pl ant
itself r emai ned in operat i on!)
Thi s happy si t uat i on persisted for a while after the first (analogue) elec-
tronic systems arrived, but the advent of comput er - based systems
event ual l y br ought about a sea-change in t he mai nt enance of equi pment
and systems.
For t he hapless end-user, it often seemed t hat as soon as a new syst em
was installed it had become obsolete. To make mat t ers worse, the repl ace-
ment was not onl y bet t er, faster and cheaper but it was so wi del y used t hat
it qui ckl y began to be difficult to find mai nt enance st aff who were able t o
under st and and wor k wi t h t he ' old fashioned' equi pment t hat had so
recent l y been t he owner' s pri de and j oy. Worse still, as t i me went by the
r epl acement part s became steadily mor e expensive and difficult to
obt ai n.
The ul t i mat e disaster woul d be when the manuf act ur er of the system
was t aken over or went out of business entirely. In the t akeover si t uat i on
the new pr opr i et or woul d somet i mes t ry and persuade the end user to
repl ace the old system wi t h one from his own stable. I f the pressures were
resisted, mai nt enance of the old system coul d become very expensive
indeed!
It' s not al t oget her a desirable situation, but it's what we have, so in this
chapt er we shall look at t he probl em and provi de some advice on how to
live wi t h it. We st art wi t h a quick look at t he reasons behi nd t he situation,
t he i nt ent i on bei ng to show t hat , r at her t han being due to manuf act ur er s
t ryi ng to bolster t hei r sales revenues, t here are l egi t i mat e reasons for t he
const ant evolution of systems and technologies.
10.1 The r e a s o n s be hi nd t he c ha ng e s
The devel opment of a comput er-based system is a very complex, time-
consumi ng and expensive business. The need for a new system is usually
brought about by a compet i t or offering somet hi ng bet t er and cheaper (or
at least appar ent l y mor e cost-effective), and when t hat happens sales of t he
present gener at i on begin to collapse. Clearly, i f this situation is allowed to
persist it will ul t i mat el y l ead to t he demise of t he company, so a proj ect has
to be l aunched to pr oduce t he next -generat i on system.
At this poi nt t he systems manuf act ur er must look at the latest offerings
from t he manuf act ur er s of t he mi cr opr oces s or s - - t he component s t hat are
at t he hear t of all moder n systems. For such companies, industrial process-
cont rol systems ar e not among t he most val uabl e applications, because
sales into this mar ket ar e small in vol ume yet face t he most st ri ngent
requi rement s for performance, reliability and safety.
The manuf act ur er of a domest i c washi ng machi ne will buy very many
mor e devices and will be far less demandi ng wi t h respect to t he perfor-
mance and reliability requi rement s.
Reliability and consequential loss are maj or issues in t he process-
cont rol market . A washi ng machi ne t hat lasts for five years seems to be
almost t he except i on these days, whereas power stations have to operat e
over a lifetime of 20 years or more. A failed device on a washi ng machi ne
may in t he long t er m result in compensat i on claims, but these are likely to
be small in compari son wi t h those resulting from an i nci dent in a power
station. When t he pl ant is a nucl ear power station, t he costs can become
enormous, and t he implications of failure ar e daunt i ng. (After t he i nci dent
at Uni t 2 of t he Thr ee Mile Isl and nucl ear power pl ant in t he USA on 28
Mar ch 1979, one mi ni comput er manuf act ur er pl aced a compl et e and
worl dwi de embar go on t he use of t hei r product s on any nucl ear plant. This
present ed somet hi ng of a probl em to t he many C&I systems vendors who
were al r eady compl et i ng t he manuf act ur e of systems based on these
products, for installation on nucl ear power stations ar ound t he world.)
Once t he at t ent i on of t he device manuf act ur er has been capt ur ed, t he
C&I system manuf act ur er has to persuade hi m to divulge the devel opment
Upgrading and refurbishment systems 199
plans for his pr oduct range. The reason for this is simple: t he time-scale
involved in devel opi ng and l aunchi ng a new comput er - based system ma y
be two or t hree years. Thi s time-scale is compar abl e wi t h t he ent i re life
expect ancy of a single gener at i on of microprocessors. I n ot her words, i f t he
C& I systems manuf act ur er l aunches a devel opment pr ogr amme based on
a vari et y of mi croprocessor t hat is available today, t he device will be
obsolete by t he t i me t he system is l aunched.
The C&I manuf act ur er t herefore has to st art designing his system and
bui l di ng prot ot ypes using earl y devel opment samples of microprocessors
and so forth. At this stage his designs must be based on a great deal of
predi ct ed dat a, some of whi ch ma y prove to be overoptimistic.
At t he end of all this, t he new system is l aunched and aft er earl y trials
it becomes widely used (or so t he supplier hopes). Because of this success,
t he manuf act ur er of a rival system t hen begins to lose his own mar ket
share, so he too starts looking at a new devel opment . Thi s is event ual l y
l aunched. A compl ex race, an expensive game of ' t echni cal leap-frog' , is
t hen on.
Even wi t hout this compet i t i on, each new gener at i on of mi croprocessor
offers advances in t erms of power and performance, and offers solutions to
probl ems t hat usage of t he previous generat i on has exposed. Worse still,
t he manuf act ur er s of t he electronic component s cannot support endless
varieties of devices, so t he day event ual l y arrives when t he ol der generat i on
of component s is wi t hdr awn. Thi s adds f ur t her compl exi t y to t he life of t he
C&I manuf act ur er who has a responsibility to keep supplying his
cust omers with r epl acement part s over several decades.
10. 2 L i v i n g wi t h change
I f t he above sounds like an excuse, it is, but it is a ver y valid excuse,
because t he pressures are real, and t hey ar e inexorable. The only realistic
appr oach therefore, is to accept t hat change is inevitable and to formul at e
a pl an to deal wi t h it. All t he same, when a system is purchased every
possible precaut i on should be t aken to ensure t hat it provides satisfactory
service over a long period. Some precaut i ons ar e listed below:
10.2.1 Defining the requirements f o r long-term support
As mi ght be expected, much depends on t he actions t hat are t aken at t he
outset, even before t he system has been ordered. At t he definition stage of a
DCS cont r act an i mpor t ant t arget is to wri ng from t he system vendor as
ma ny wr i t t en assurances of l ong-t erm support as possible, and to make as
sure as is possible t hat the vendor is reliable and t hat he will be abl e to
st and by his promises. It is a har d goal to meet, but the time to press for
such promi ses is when an order for several hundr eds of t housands of
pounds, somet i mes millions, is in the offing.
Unf or t unat el y, t he commer ci al clauses in large cont ract s are usual l y
f r amed by peopl e who know little or not hi ng about t he evol ut i on of DCS
or C&I t echnol ogy and t he effects on pl ant mai nt enance. In this si t uat i on
a vendor can respond wi t h weasel words t hat are wort h very little in
reality.
Havi ng obt ai ned the promises, it is vital t hat t hey are exami ned as
careful l y as possible by professionals who really under st and the rel evant
technology. The t i me and cost of this eval uat i on will be r epai d many
times, over the oper at i onal life of the plant.
10.2.2 Keeping track of costs
Once the equi pment has been acqui red, the users must focus on how best
to use it, and t hey must be pr epar ed to i gnore all i nformat i on on new
systems (whi ch are al ways bet t er t han the one t hat has j ust been bought!).
The fact is t hat it has been bought , and now the users must gain the
maxi mum advant age from the investment.
An i mpor t ant pr ocedur e is to keep careful records of all the markers
al ong the r oad t hat stretches from initial acquisition to ul t i mat e repl ace-
ment:
Log each failure t hat occurs, recordi ng the t i me and dat e of the event.
Thi s is useful i nformat i on for the user and the vendor, since it support s
or disproves reliability claims.
Keep a not e of the t i me t hat elapses bet ween pl aci ng an or der and
receiving t he goods. Thi s will hel p reinforce your argument s t hat delays
in shipping spares are increasing, exposing the pl ant to i ncreased levels
of risk, causi ng reduced out put , or i ncurri ng i ncreased oper at i onal
costs.
Moni t or and record how l ong it takes to repl ace the failed item of
equi pment wi t h a spare from stock.
Recor d t he cost of spare part s (pri nt ed-ci rcui t cards, power supplies
etc.).
Not e the del i very t i me for spares.
When recordi ng details of a tMlure, make a not e of any pl ant out age
or reduct i on in out put t hat can be di rect l y at t r i but ed to t he failure.
It is also useful to t ry and relate the failure to any event t hat occur r ed
at or j ust before the t i me at which the failure occurred, such as a severe
t hunder st or m, or t he failure of any ot her electrical pl ant or machines.
Thi s can assist in post - mor t em analyses whi ch may poi nt to a design
defect such as i nadequat e screening or eart hi ng, or poor design of
power supplies.
Tabl e 10.1 shows a simple f or mat t hat can be used for recordi ng this
t ype of i nformat i on.
10.3 Maki ng the deci si on to change
The day will event ual l y dawn when it is wise to exami ne the ar gument s
for and against upgr adi ng the DCS, and to t hen deci de on t he best course
of act i on to take. Reachi ng a decision will be eased i f careful records have
been mai nt ai ned as described above, but even t hen pressures will be
applied to upgr ade t he system.
10.3.1 The impact of change on the operators
System refurbi shment can pr oduce maj or changes to t he oper at or inter-
faces. Such changes are greatest when a system based on the use of a hard-
wi red desk is repl aced wi t h a so-called 'soft desk'.
I n t he former, every act uat or is cont rol l ed via its own discrete
ha nd/ a ut o station (al t ernat i vel y called an a ut o/ ma nua l station). These
enabl ed t he oper at or to compl et el y di sconnect t he electronic cont rol l er
from t he act uat or and to manual l y position the final el ement via a discrete
cont rol circuit. Such a system provi ded a high degree of r edundancy since,
with t he exception of t he wires linking t he control desk to t he act uat or, t he
circuit for provi di ng manual cont rol was totally separat e from t he electro-
nics of t he aut omat i c controller.
Har d desks also provi ded discrete indicators (often moving-coil
meters) showing t he i mpor t ant pl ant paramet ers. Somet i mes these indica-
tors were dri ven by t ransmi t t ers whi ch were compl et el y separat e from
those feeding t he electronic controller, agai n provi di ng a high degree of
r edundancy.
Wi t h t he advent of high-reliability systems ar gument s began to be
raised t hat t he cost of provi di ng such levels of r edundancy was no l onger
justified. But it was onl y wi t h t he advent of sophisticated faul t -det ect i on
systems, coupled wi t h hi ghl y reliable cont rol comput ers (often embodyi ng
t wo-out -of-t hree faul t -det ect i on systems) t hat t he mor e conservative
engineers began to accept t hat it was safe to ent rust t he control of large-
scale hazardous processes to a soft-desk concept.
Wi t h boiler cont rol systems, even i f such a decision is reached, car e
should be t aken to ensure t hat essential i nformat i on is available to t he
oper at or in the event of a serious failure of t he control system or its power-
suppl y system. For exampl e, t here is a need to provi de densi t y compensa-
tion for cer t ai n measur ement s i f a DCS failure occurs. (An exampl e is t he
drum-l evel measur ement , wher e t he requi rement s for compensat i on in the
presence of a DCS failure were poi nt ed out in Section 6.3.)
10. 4 A r e f u r b i s h me n t c a s e s t u d y
Changi ng to a 'soft' desk results in t he oper at or bei ng present ed wi t h a
system of displays and cont rol s whi ch is compl et el y di fferent from t hat
wi t h whi ch he or she was fami l i ar when using t he har d- wi r ed prede-
cessor.
It ma y be useful to briefly exami ne a par t i cul ar exampl e, t he cont rol
desk and console of a l arge coal-fired boiler. The original cont rol -room
facilities had been supplied in t he days of anal ogue electronic systems (c.
1975) and had vert i cal panels provi di ng full i ndi cat i on and cont rol facil-
ities of all ei ght mills, plus a mul t i pl exi ng system t hat enabl ed t he oper at or
to gai n access from t he desk to any two of t he mills at a given time. At t he
t i me of t he ori gi nal design, this present ed some difficulties for ensuri ng
t hat gai ni ng cont rol of a pai r of mills did not i nt r oduce any di st urbance
into t he plant.
Thi s system was repl aced by a di st ri but ed system in 1989 (Figure 10.1)
but this was, in t urn, upgr aded wi t h mor e up-t o-dat e systems in 1998.
Fi gure 10.2 shows t he new ' soft-desk' cont r ol - r oom facilities aft er t he
second refurbi shment .
The use of screen-based cont r ol - r oom facilities enables a weal t h of infor-
mat i on to be provi ded for t he oper at or in a form t hat is easy to assimilate
and use. A 'soft desk' is also i nherent l y mor e flexible, allowing addi t i onal
cont rol and display facilities to be easily added as requi rement s evolve or
new pl ant is added.
10.4.1 The involvement of plant staff
Experi ence has shown t hat t he success of 'soft desk' upgrades can be corre-
l at ed to t he level of ownershi p t aken by t he uni t operators. The cur r ent
st rat egy of ma ny of t he DCS & SCADA i mpl ement ers is to involve t he
operat ors in t he process of designing the mi mi c displays, so t hat t hey are
happy to use t he new screens and thus t ake ownershi p in t he success of t he
scheme.
I n any case, dr amat i c changes such as these necessitate a pr ogr amme
of ret rai ni ng for all pl ant personnel. The operat ors must l earn how to use
t he new display and cont rol facilities and t he mai nt enance st aff must be
Figure 10.1
The 'hard-wired' control console and panel o f a large coal-fired power-plant
before refurbishment
Scottish Power plc. Reproduced by permission
Figure 10.2
A ' soft desk' replacement f o r the "hard-wired" operator f aci l i t i es af t er
refurbishment
Scottish Power plc. Reproduced by permission
t rai ned in the mai nt enance of the new systems. It is of course essential
t hat t he costs of such t rai ni ng are consi dered in the cost / benef i t analysis of
a refurbi shment project.
10.5 Why refurbish?
Pl ant manager s will be const ant l y besieged by DCS sales personnel who
will be onl y t oo happy to pr ovi de pl ent y of argument s to j ust i fy repl ace-
ment of t he existing systems wi t h their own. To count er bal ance this wei ght
of evidence, here are a coupl e of exampl es of reasons why a system should
not be upgraded.
10.5.1 "The performance of the present system has degraded"
An oper at i onal cont rol syst em t hat has at one t i me met its per f or mance
requi rement s is unlikely to st art pr oduci ng inferior results merel y because
of det er i or at i on of t he electronic component s. A component failure may
occasi onal l y result in drift, or it may cause the system to oper at e wrongly,
but t he solution in such cases is to repl ace the component or t he car d in
whi ch it is fitted.
Det er i or at i on of a DCS' s abi l i t y to cont rol the pl ant is far mor e likely
to be the result of poor per f or mance of the interface t ransducers (the trans-
mi t t ers and act uat ors), r at her t han any fact or in t he syst em itself. For
exampl e, i nadequat e mai nt enance of a val ve act uat or will result in hyster-
esis or overshoot . Changi ng the electronic system t hat controls the val ve
will do not hi ng to i mpr ove on cont rol performance.
Equal l y likely is t he possibility t hat the cont rol l ed el ement itself (a
val ve or damper ) requires mai nt enance. In several cases, the appl i cat i on
of grease to a shaft or bear i ng has worked wonders on the per f or mance of
t he loop!
The poi nt is this: i f the per f or mance of an act uat or or a final el ement
has det er i or at ed because of lack of mai nt enance, it is unlikely t hat a maj or
i nvest ment in a new DCS will result in any dr amat i c l ong-t erm i mprove-
ment. It was the at t i t udes of the management or st af f t hat resulted in the
lack of mai nt enance and these will not necessarily change when the new
system arrives. A short -t erm i mpr ovement may well be obt ai ned at the
out set (because the manuf act ur er will have been forced to fi ne-t une the
per f or mance of all act uat ors and final elements in or der to meet perfor-
mance guarant ees, and because any new system engenders some
ent husi asm among mai nt enance staff, if not the pl ant operat ors), but once
t he situation has settled down t he lack of mai nt enance will once agai n
result in an event ual degr adat i on in performance.
10. 5. 2 N e w syst ems are f a s t e r , more p o we r f u l , better
Users will always be aware of a system t hat seems to work bet t er in some
way t han t he one installed on t hei r own plant. That ' s the nat ur e of t he
game.
But t he system installed on any installation has been adj ust ed to work
effectively with t he idiosyncrasies of t he plant. It will have t aken somebody
a long t i me to reach t hat stage. I f the system is changed t hat l earni ng
curve will have to be renegot i at ed when the system is replaced. This under-
lines the i mpor t ance of document at i on, as described in the following
section.
10.6 Document i ng the present s ys t em configuration
It is almost inevitable t hat t he design of a cont rol system will al t er
bet ween the initial concept ual stage, t hr ough handover and dur i ng
prol onged use afterwards. The initial system confi gurat i on will be
carefully defined before commissioning begins. Al t hough t he subsequent
changes may be recorded, all too often t he reasons for i nt roduci ng t hem
may not be logged. Al t hough t he changes may be t he result of errors, it is
equal l y likely t hat t hey are requi red because of some mi sunderst andi ng,
dur i ng t he design phase, of the plant' s characteristics or functions.
The changes are i mpl ement ed in order to correct such errors or misun-
derstandings, so t hat t he system works better. But i f t he reasons for t he
changes are forgotten, t here is every likelihood t hat when t he t i me comes
to refurbish or repl ace t he systems the original mi sunderst andi ngs will be
repeat ed.
When embar ki ng on a r epl acement or refurbi shment project, the task
will be i mmeasur abl y assisted by the availability of comprehensi ve and
accur at e document at i on of t he systems as finally configured, t oget her wi t h
det ai l ed i nformat i on on t he reasons behi nd all changes to the original
designs.
10.7 Summary
We have now r eached t he end of this overview of a wide and compl ex
subject. I hope t hat I have been able to t hrow some light on t he t echnol ogy
and t hat t he expl anat i ons ma y have lifted some of t he veils of myst er y
t hat somet i mes seem to obscure it. The fact remai ns t hat it is a compl ex
mat t er and it is unwise to ent rust t he safety of a pl ant to peopl e who do not
under st and ei t her t he cont rol aspects or t he pl ant operations.
It has been my privilege to have worked wi t h power pl ant t hr oughout
my car eer and I hope t hat some of what I have put down here will be
useful, and t hat it ma y encour age ot hers to t ake up a ver y i nt erest i ng and
i mpor t ant subject.
U
p
g
r
a
d
i
n
g

a
n
d

r
e
f
u
r
b
i
s
h
m
e
n
t

s
y
s
t
e
m
s

2
0
7

o

r
.
-

E

)

0

n
"

o

6

~
z

a
~

0

,
~
.

6

@

o
0

o
~

I
-
.
i

I
~
~

~
i

~

~

~

~
-

e

d

o
~
'
,
-

~
o
o
o

~

~
-
~
.
~

-
~

L
L
I

z
~
o

"
-
d
~

m

.
~
_

F
-

"
-

E

~
d

~
z

I
0

m

1
3
.

_
I

0

0

m

L
r
~

e
~

N

Further reading
Boi l e r and t ur bi ne de s i g n and c o ns t r uc t i o n
BLOCH, H. E: A pract i cal gui de to st eam t ur bi ne t echnol ogy' (McGraw-Hi l l ,
London, 1996)
BS 1113: 1998: ' Speci f i cat i on for desi gn and ma nuf a c t ur e of wat er - t ube st eam
gener at i ng pl ant (i ncl udi ng superheaters, reheaters and steel t ube economizers' ,
BSI, London, 1998
ELLI OTT, T. C., CHEN, K., and SWANEKAMP, R. C.: ' St andard handbook of
powerpl ant engineering' (McGraw-Hill, London, 1998)
JERVI S, M. W. (Ed.): ' Moder n power station pract i ce ( BEI / Per gamon, Oxford,
1991)
KEARTON, W. J.: ' St eam t ur bi ne t heor y and practice (Isaac Pi t man and Sons,
London, 1960)
KE HL HOF F E R, R. , BACHMANN, R. , NI ELSON, H. , and WARNER, J:
' Combined-cycle gas and steam power plants (Pennwell, Tulsa, OK, USA, 1999)
LEYZEROVI CH, A. : ' La r ge powe r s t eam t ur bi nes : des i gn a nd o p e r a t i o n '
(Pennwell, Tulsa, OK, USA, 1997)
SINGER, J. G. (Ed.): ' Combust i on- - f ossi l power.
(ABB/ Combust i on Engi neeri ng Inc., Connecticut, USA, 1991, 4t h edn.)
Cont r ol and i n s t r u me n t a t i o n
ASTROM, K., and HAGGLUND, T.: ' PI D controllers: theory, design and t uni ng'
(AmericanTechnical Publishers, Hitchin, UK, 1995, 2nd edn.)
BISSELL, C. C. : ' Cont rol engi neeri ng' ( Chapman and Hall, London, 1994)
BOLTON, W.:' Newnes control engi neeri ng pocket book' (Newnes, Oxford, 1998)
CONSIDINE, D. M.: 'Process i nst rument s and control handbook' (McGraw-Hill,
NewYork, 1985)
DUKELOW, S. G. : ' The control of boilers (AmericanTechnical Publishers, Hitchin,
UK, 1991, 2nd edn.)
GI LLUM, D. R.: ' Indust ri al pressure, level and density measurement ' (Ameri can
Technical Publishers, Hitchin, UK, 1995)
HUBER, J. C.: ' Indust ri al fiber optic networks' (ISA, ResearchTriangle Park, NC,
USA, 1995)
I ns t r ument Society of Ameri ca. ' I nst r ument s i n the power i ndust r y' ( Amer i can
Technical Publishers, Hitchin, UK 1990)
I ns t r ument Society of Ame r i c a ' The ISA fieldbus gui de (ISA, ResearchTri angl e
Park, NC, USA, 1997)
JERVI S, M. W. (Ed.): ' Moder n power station practice' (BEI/ Pergamon, Oxford,
1991)
JERVI S, M. W.: ' Power st at i on i nst r ument at i on' (Inst i t ut e of Measur ement and
Control, But t erwort h-Hei nemann, Oxford, 1993)
J ORDAN, J. R.: ' Seri al net worked field i ns t r ument at i on' ( Amer i can Technical
Publishers, Hitchin, UK, 1995)
LEVINE, W. S.: ' Control handbook' ( CRC/ I EEE, Boca Raton, FL, USA, 1996)
LINDSLEY, D. M. :' Boiler control systems' (McGraw-Hill, London, 1991)
MORRI S ON, R. : ' I n s t r u me n t a t i o n f u n d a me n t a l s and a ppl i c a t i ons ' Wiley,
Chichester, UK, 1984)
NOLTINGK, B. E. (Ed. ):' Instrumentation reference book' (Butterworths, London,
1988)
OLSSON, G., and PI ANI , G.: ' Comput e r systems for a ut oma t i on and cont r ol '
(Prentice Hall, London, 1992)
PLATT G.: ' Process cont r ol : a pr i me r for t he non- speci al i st and ne wc ome r '
SPI TZER, D. W. (Ed.): ' Fl ow me a s ur e me nt ' ( Amer i can Techni cal Publ i shers,
Hitchin, UK, 1991)
Human factors (ergonomics) and control-room design
I VERGARD, T. ' Handbook of control r oom design and ergonomics' (Taylor and
Francis, London, 1989)
Mc CORMI CK, E. J. , and SANDERS, M. S. ' Human factors in engi neeri ng and
design' (McGraw-Hill, Dur ham, NC, USA, 1983)
PHEASANT, S. T.: ' Er gonomi cs - s t andar ds and guidelines for designers' (Taylor
and Francis, London, 1986)
SALVENDY, G. (Ed.): ' Handbook of human factors' (Wiley, Chichester, UK)
WAGNER, E.: ' The comput er display designer' s handbook' (St udent Li t erat t ur,
Chartwell-Bratt, Sweden, 1989)
Interference, electromagnetic compati bi l i ty and
shielding
Counc i l Di r ect i ve E MC/ 8 9 / 3 3 6 / E E C ' Th e EMC di r ect i ve" Counci l of t he
Eur opean Communi t i es, Brussels, 1989
Further reading 211
MORRISON, R. and LEWIS, W. H.:'Grounding and shielding in facilities' (Wiley,
Chichester, UK, 1990)
SCOTT, J. S., and VAN ZYL, C.: Introduction to EMC (Newnes, Oxford, 1997)
Safety and rel i abi l i ty
CHEMICAL INDUSTRIES ASSOCATION: A guide to hazard and operability
studies'. Chemical Industry Safety and Health Council of the Chemical Industries
Association, London
COX, S., and TAIT, R.: 'Safety, reliability and risk management: an integrated
approach" (Butterworths, Oxford, 1998)
GOBLE, W. M.: ' Control system safety evaluation and reliability' (American
Technical Publishers, Hitchin, UK, 1998, 2nd edn.)
HEALTH AND SAFETY EXECUTIVE: ' Programmable electronic systems in
safety-related applications. HMSO, London
INSTITUTE OF ELECTRICAL AND ELECTRONICS ENGINEERS: Human
factors in power plants (IEEE, NewYork, 1997)
BS EN 60801-2:1993.'Electromagnetic compatibility for industrial process measure-
ment and control equipment- Part 2: Electrostatic discharge requirements'. British
Standards Institution, London
'Draft Standard 00-58: a guide for HAZOP studies on systems which include a pro-
grammable electronic system'. Ministry of Defence Directorate of Standardization,
Glasgow, 1995
NFPA 8502-95 ' Standard for the prevention of furnace explosions/implosions in
multiple burner boilers" National Fire Protection Association, Quincy, MA, USA
REDMILL, E, and RAJAN, J.: ' Human factors in safety-critical systems' (Butter-
worth-Heinemann, Oxford, 1997)
Tuning of control l oops
ASTROM, K., and HAGGLUND, T.: ' PID controllers: theory, design and tuning'
CORRIPIO, A. B.:' Tuning of industrial control systems' (American Technical
Mc MILLAN, G. K.:'Tuning and control loop performance: a practitioner's guide,
(AmericanTechnical Publishers Ltd., Hitchin, UK, 1994, 3rd edn.)
Val ves
BAUMANN, H. D.: 'Control valve primer' (AmericanTechnical Publishers Ltd.,
Hitchin, UK, 1998, 3rd edn.)
FISHER CONTROLS COMPANY: 'Control valve handbook' (Fisher Controls
Company, Rochester, Kent. 1977)
I NSTRUMENT SOCIETY OF AMERICA: 'ISA handbook of control valves'
General
FRIEDMAN, E G.: ' Economics of control improvement' (American Technical
ISA:'Dictionary of measurement and control' (AmericanTechnical Publishers Ltd.,
Hitchin, UK, 1995, 3rd edn.)
ISA: ' Instrumentation standards and practices (American Technical Publishers,
Hitchin, UK, 1991, 10th edn.)
BATTIKHA, N.E.:' The management of control systems:justification and technical
auditing' (AmericanTechnical Publishers, Hitchin, UK, 1992).
I nde x
4- 20 mA signal range, reasons for using
176
absorbers 30
ac feed to DCS, loss of 154
accuracy
effect of impulse pipework on 161
of steam-temperature control 138 41
acoustic noise, from valves 121
active transmitters 155
actuator 163
electric 168
hydraulic 168
pneumatic 163 -6
pneumatic, quality of air supply for
168
adiabatic expansion 6
adjacent-burner discrimination 95
aeolipile 5
agricultural waste 42
air
composition of 26
distribution in furnace 64
entrained 4, 15
air flow, as load indication for oxygen
controller 76
air heater 15, 27
air supply system tor mills (pulverisers)
37
air-lead system 69
alarm, DCS power-supply failure 154
alternator 15
analogue inputs 155
angle of attack, fan blade 31
armour, cable 169
shielding properties 172
atomising, of fuel oil 38
attemperator
function of 136
maintaining adequate water pressure
for 130
mechanically atomised 137
multi-stage 149
spray-water, corner-fired boiler 26
turndown range 137 -40
variable annulus 139
variable area 138
variable orifice, installation 140 - 1
auto/manual station 201
availability, control system 174
axial-flow fan 30
Babcock ' E' mill 35
back-up, DCS power supplies 154
bag filters 30
balanced-draught boiler, pressure profile
92
basic states of matter 2
batteries, DCS 154
bias signal, use in fuel/air ratio system
67
binary control 95
blade pitch control, axial-flow fan 30
blades, steam turbine 15-16, 30
block valves 41
blowdown
boiler 20
drum-level transmitter impulse lines
113
blowdown valve, transmitter 162
BMS 45
boiler
dynamic response 53
fired, control of heat input 63
boiler-following control
response 54
tuning 57
boiling (definition ot) 2
boiling point 2
booster t~ns 30
booster, pneumatic 164
bottoming cycle 59
Boyle's law 6
Brayton 10
BS 1113:1998 117
BS 6739:1986 172
bubbling fluidised-bed boiler 42
bumpless transfi~.r 90
214 Index
Bunker C (t ype ofoi l ) 39
bur ner light-off, pre-check 98
bur ner - management system 45
burners 26, 64, 95
effect of changi ng firing pat t er n 135
mul t i pl e 81
busbar, segregat i on 170
busbar, ear t h 170
by-pass, condenser ext ract i on 132
bypass valves, HRSG 23
CCGT 11
cables
ar mour ed 169
C&I 169, 172
resistance 179
Cal l endar, H. L. 3
capi t al cost, of feed pl ant 112
car bon monoxi de
analyser, flue-gas 72
use in combust i on cont rol system 72
carbon-monoxi de analysis, i ndependence
from furnace l eakage 72- 3
carbon-i n-ash moni t or 7 6 - 8
Carnot , S. 5
Car not cycle 4
Car not ' s law 5
carry-over, wat er 104
cascade cont rol system 143
t wo-el ement feed-water cont rol 110
cat hode- r ay t ube 157
cavi t at i on
in feed systems 126
in feed valves 121
charact eri st i c
feed pump 127 - 9
val ve 121
chilling, use of dr um pressure to i ndi cat e
sat ur at i on t emper at ur e 150
choked oper at i on of feed valve 127
CHP 59
Class I i gni t er 44
Class 2 i gni t er 44
Class 3 i gni t er 44
Cl i ni cal waste 42
coal
burni ng of 35
calorific val ue of 83
flow measurement 81
coal feeder, speed adj ust ment 82
Code of Pract i ce for i nst rument at i on in
process cont rol systems: i nst al l at i on,
design and pract i ce 172
coefficient, valve sizing 124- 5
co-generat i on 59
cold reheat t emper at ur e 152
combi ned cycle 10
combi ned-cycl e pl ant
gas-t urbi ne 11
HRSG oper at i on 13
combi ned heat and power 59
combustion chamber 26
combust i on cont rol 64
commissioning 155
condenser 9, 16-17
ext ract i on pump 132
level control, i nt eract i on wi t h
deaer at or level cont rol 132
conduct i vi t y probes, drum-l evel
measurement 114-15
connect i ng pipes
l aggi ng of 162
slope of 162
consequential loss 198
const ant -head reservoir 112
cont rol room, envi ronment al conditions
160
cont rol l er
det er mi nat i on of opt i mum, for single-
el ement feed-water cont rol 108
gai n adj ust ment in feed-water cont rol
system 107
steam t emperat ure, opt i mum
charact eri st i cs 143
convection stages 14
conveyor, screw, in wast e-i nci nerat i ng
pl ant 42
cooling, flame scanner 96
cooling tower 17
cooling wat er 17
outfall 17
co-ordi nat ed uni t l oad control, response
of 57
corner-firing 26
corrosion
back-end 25
caused by residual oxygen 19
in wast e-burni ng pl ant 42
cross-connection 156
cross-limited combustion control system,
descri pt i on of 70
current generat ors, electronic 156
current signals, electronic 179
Cv, valve 124- 5
CW outfal117
damper 33
for ai r- and gas-flow cont rol 33
l i neari si ng the control of 33
mill, hot ai r and t emper i ng ai r 87
dat a
hi ghway 157-8
ret ent i on on loss of power 154
da Vinci, Leonar do 10
DCS 20, 154
al arms from wat er-t reat ment pl ant
2O
HAZOP studies on 175
upgr adi ng 201
dead zero 176
deaer at or
cont rol 131-4
function of 19
spill cont rol 133
delay, due to i nadequat e pipework size
at posi t i oner 164
demand, seasonal 51
density error, in drum-l evel measurement
113-14
deposits, in i nst rument connections 162
deri vat i ve t erm 143- 4
desuperheat ers 136-41
det er i or at i on of C & I system performance
203
devel opment phases of comput er-based
system 198- 9
di agnost i c 176, 178
tools 174
use in cross-limited system 72
di aphr agm oper at or 164
diesel oil 39
di gi t al inputs 155- 6
di gi t al out put s 155- 6
di ode auctioneering, of power supplies
154
di r t 160
dispatcher, l oad 50
di spl ay formats 91
di st ance/ vel oci t y lags, in pneumat i c
systems 168
di st ri but ed control system 153- 4
dol omi t e 42
doubl e-bl ock-and-bl eed valves 41
drai ni ng, t ransmi t t er 162
dr um, boiler 13
size, in rel at i on to boiler t hroughput
112
dr um level controller, cont rol terms used
in 107
dr um level measurement 112-20
al t ernat i ve types 118
back-up measurement , need for 114
Index 215
connections to dr um 119
densi t y compensat i on 113-16
discrepancies between measurements
120
dr ai n connections 118
equat i on 114
l aggi ng of impulse pi pework 116, 120
mi ni mum viewable level 118
numbers requi red 118-19
size of connect i ng pipes 119
slow response 120
sources of error 120
dr um wat er level, opt i mum 104
dryness fraction (steam quality) 3
dual redundancy, negat i on of by
incorrect cabl e rout i ng 158
dust 158
ear t h connection
good pract i ce 170
segregation 170
economiser 14
efficiency, t hermal 9
effect of reheat sprays on 151
electric act uat or 168
i mport ance of specifying all failure
modes 168
el ect romagnet i c compat i bi l i t y 100, 158,
169
el ect romagnet i c conditions 158-9, 169
el ect ro-pneumat i c convert er 164, 167
el ect ro-pneumat i c posi t i oner 163, 167
electrostatic screen, cabl e 172 - 3
emc 100, 158, 169
emissions, stack 25
ent rai ned gases, in i nst rument
connections 162
ent r opy 8
envi ronment al considerations, for C&I
equi pment 158
equal -percent age characteristic, feed
valve 123- 4
equi pment room, envi ronment al
conditions 159
ergonomics 157
erosion, in feed-valve systems 21,127
evaporat i on, basic definition 2
event logging 157
excess ai r 64
in f uel / ai r rat i o system 69
vari at i on with l oad 65
exhauster 36- 7, 86
expansi on of boiler, cabl e rout e
considerations 169
216 Index
explosions,
in mills (pulverisers) 87
risk of 40, 69
extraction pump 131
bypass 132
leakages in 19
factory-acceptance test 185
fail-fix operation, actuator 168
fail-safe operation, feed valves 125
failure
detection 69
modes in electronic systems 176
recording of 200
fan
axial [low 30- 1
booster 30
centrifugal 30
FD 15, 25
gas-recycling 152
FAT (factory acceptance test) 185
fault detection 180, 200
FD fan 15
feed pump 15
characteristic 21,127-9
variable-speed 129
feeder, coal
speed-control methods 82
speed feedback 83 - 6
feed-water control, objectives of 103 - 4
feed-water flow, use in two-element fired-
water control system 105-9
feed-water valve, operation at design
pressure drop 130
FGD, draught control in 91
fibre optics 157
field cabling 156- 7
field conditions 160
filters, low pass, tor I / O 169
final superheater, time constant of 148
fireball, in corner-fired boilers 26, 150
flame
discrimination 46
status indication 95
flame monitor
cooling of 96
importance of selection, location and
maintenance 95
siting considerations 46
use in BMS system 95
flame pattern 46
flame scanner
cooling of 96
importance of selection, location and
maintenance 95
siting considerations 46
use in BMS system 95
flame spectra 98
flashing
in drum-level measurement 116, 120
in feed-valve systems 127
flow-per-burner signal 78
flue-gas desulphurization, draught
control with 30, 91
fluidiscd beds 26, 42
flushing of instrument connections 162
forced circulation 13
torced-draught fan 15, 25
frequency support 58
front-firing 26
fuel valve, failure of 71
tuel/air ratio adjustment, in cross-
limited system 71)
fuel/air ratio control 67
fuel-flow transmitter, failure of 72
fuel-handling system, in WTE plant 60
fuel-lead system 69
fuel-rich operation, parallel system 66
fuels, multiple 79
function generator, use in equating heat
input from different fuels 81
functional specification 184
furnace 25
furnace exit temperatures, control via
tilting burners 150
furnace pressure, maintaining 92
furnace purge 45
Fuzzy Logic, use in deaerator control
134
gain 145-8
ti:cd-valve 12 l
use in fuel/air ratio system 67
gain block, use with multi-fuel system 80
galvanic isolation 155
gas analyser, location of 74
gas analysis controller, combined 76
gas burner, use of flow/pressure
characteristic 78
gas firing 40
gas flow apportioning, used for steam-
temperature control 151
gas oil 39
gas recirculatinn (recycling) 152
gas turbine 10-- 12
exhaust heat usage 23
gas-leak detectors 40
gearboxes, vari abl e-speed 34
Gri d, UK Nat i onal 51
gri d-cont rol centre 50
groupi ng of I / O channels 156
groups, burner 78
hand/ aut o station 201
in CO cont rol system 76
use with oxygen cont rol l er 75
hand/ aut o transfer, probl ems with
mul t i pl e sub-loops 88
hazardous environments 158
hazards 95
HAZOP 175
heat loss, in flue gases and cooling wat er
9
heat -recovery steam gener at or 11
Her o of Al exandr i a 5
hook-up di agr am 160
HP stage 16
HRSG 11
cont rol of heat i nput 63
wat er and steam circuits 21
humi di t y, effect on static electricity 155
hydraul i c act uat or 168
fl ammabi l i t y of oper at i ng medi um 169
hydraul i c couplings 34
hydrazi ne 19
Wo
cards 154
channels, spare 156
I / P convert er 163-4, 167
effect of vi brat i on 168
length of pi pewor k to valve 166
operat i on of 164
response time 164
ID fan 27
idle burner, cooling of 64
i gni t er 44
implosion, furnace 93
impulse pi pe 160
i nci nerat ors 42
i nduced-draught fan 27
i nherent characteristic, of feed valve 122
i nput resistance, of receiving circuit 179
i nput / out put 155-6
i nst al l at i on
gui dance provi ded by manufact urers
and suppliers 159
drum-l evel measurement 118
installed characteristic, feed valve 122
i nst rument mani fol d 161- 2
Index 217
i nst rument -ai r supply, use for flame-
scanner cool i ng/ pur gi ng 96
i nt egral sat urat i on, in steam-
t emper at ur e cont rol systems 145
i nt egrat i ng system, boi l er dr um as an
107
i nt eract i on
condenser / deaer at or 132
in fan cont rol systems 93
of sub-systems 49
interference 158
influence on t ransmi t t ed signals 179
interlocks 95, 98
gas-recycling fan 152
I P stage, steam t urbi ne 16
isentropic compression 10
isolating valve, t appi ng- poi nt 162
i sot hermal expansion 6
Joul e' s law 5
keyboard, comput er 157
KKS system of nomencl at ure 186
appl i cat i on exampl e 193
coding levels 189
need to agree and co-ordi nat e 187
l aggi ng of impulse pipework, drum-l evel
measurement 116
l at ent heat 2
league table (meri t order) 52
leakage, ai r-heat er 28
LED indicators, for DO status i ndi cat i on
155
level control
deaer at or 132
dr um 103-12
l i ght ni ng 170
l i near characteristic, of feed valves 123
linearizing, of FW control-system
dynami cs 21
liner, t hermal , in spray-wat er
at t emper at or system 138
live zero 176-80
I,jungstr6m ai r heat er 27
l oad line, of mill (pulveriser) 81- 2
logic, interlock system 95
logic probes, hazards of using 155 6
long-term support , need to define 199
loop gain, in systems with vari abl e-speed
pumps 130
losses
exhaust gas 64
wat er 147
218 Index
low wat er flow al ar m 119
low wat er level al ar m 119
l ow-NOx burners, fl ame charact eri st i cs
98
low-pass filter 169
LP st age 16
mai nt ai nabi l i t y, effect of impulse
pi pework on 162
mai nt enance 203
make- up cont rol 132
make- up water, consumpt i on of 133
mani fol d, i nst rument 162
marshal l i ng, of cables and t ermi nal s 156
mar t yr valve 162
mast er
CHP pl ant 59
demand 49
wast e-t o-energy 60
mast er valve 162
maxi mum selector, use in cross-limited
system 70
maxi mum/ mi ni mum limiter, use wi t h
oxygen cont rol l er 75
Mean Ti me Between Fai l ure 174
Mean Ti me To Repai r 174
melting, definition of 2
memory, provision of back- up power for
155
mercury, use in t her mal cycles 5
Mer i t Or der 52
met ered system (combustion control) 70
microprocessors 198
mill (pulveriser)
cont rol of S1- 91
hammer type, in wast e-i nci nerat i ng
pl ant 42
mul t i pl e, cont rol of 88
performance adj ust ment 77
suction t ype 36
t emper at ur e cont rol 87
t ube t ype 36
vert i cal -spi ndl e bal l t ype 35
mi ni mum limit, use in mill (pulveriser)
cont rol system 84
mi ni mum selector block, use in cross-
l i mi t ed system 70
mouse, comput er 157
movi ng-coi l meters 201
MTBF 174
MT T R 174
mul t i -nozzl e at t emper at or , principles of
138- 9
mul t i pl e burners, effects on ai r and gas
di st ri but i on 78
muni ci pal waste 26
N2H 4 19
N20 26, 42
Na2SO3 19
Nat i onal Fi re Protection Association 44
nat ur al charact eri st i c of boiler 135
nat ur al gas, leak det ect i on 40
NFPA
st andards for mul t i - bur ner pl ant 78
NFPA 8502- 95 44, 98
Ni t rogen Oxi des 26
Nitrous Oxi de 26
NO 26
NO2 26
nomencl at ure 183
NOx 26
nuisance trips 95
offset, in drum-l evel cont rol 109
oil bur ner 3 8 - 9
use of fl ow/ pressure charact eri st i c 78
oil pressure, as i ndi cat i on of oi l flow 88
once-t hrough boiler 13, 118
opaci t y cont rol 76
oper at or displays 91, 157
oper at or workstations 157
opposed firing 26
opt ocoupl i ng of inputs 169
oscillation, in st eam t emper at ur e control
system 148
override, of safety systems 46
oxygen
analyser, flue-gas 72
concent rat i on, feed wat er 19
cont ent of ai r 76
cont rol l er
i nt egral -t i me setting 76
setpoint l oad vari at i on 75
measurement of, used in combustion
cont rol system 72
trim, combi ned with opaci t y 76
oxygen vs l oad charact eri st i c 75
P& I D (pi pi ng and i nst rument di agr am)
189
PA
fan 38
flow, changes due to ext ernal factors
83
paral l el cont rol 65 - 7
passive t ransmi t t ers 155
PES, safety of 99
PF 35
pilot, in BMS system 45
pipework, size of, in drum-l evel
measurement 119
pi t ch control, fan 30, 92
pitting, of feed-valve i nt ernal s 126
pl ant tests, need to define requi rement s
185
pl asma displays 157
PLC, use in wat er - t r eat ment pl ant 20
pl ot t er 157
plug, valve 121
plume, cooling tower 17
pneumat i c act uat or, qual i t y of ai r
suppl y 168
pneumat i c control systems 197
pockets, t emperat ure-sensor 143
Pool, power 52
Pool Selling Price 52
popul at ed I / O channels 156
posi t i oner 164- 5
i mprovi ng response by abandoni ng
166
l i mi t at i ons of use 165
poor response due to i ncorrect
cal i brat i on 166
power supply
DCS 154
t r ansmi t t er 176
PRDS, peggi ng st eam 132
pressure compensation, of gas-flow
signal 81
pressure equalization, in oil-firing
systems 39
pressure j et 39
pressure, fuel, as a flow i ndi cat or,
l i mi t at i ons of 81
pressure-drop, across feed valve 125
pressure-reduci ng valve (PRDS),
spray-wat er 132
pressurised mill (pulveriser) 35- 8
t emper at ur e control 87
l oad-cont rol st rat egy 82
pre-warmi ng, of fuel oil 39
pr i mar y ai r
flow cont rol l er 88
fan 35, 38
damper , control of 88
flow 81 - 2
pri nt ers 159
procedureless, bumpless transfer 90
Pr ogr ammabl e Electronic Systems
safety of 99
Index 219
HAZOP studies on 175
pr ogr ammi ng, of at t emper at or spray
valves 150
proj ect i on systems 157
pr opane igniters, l eakage hazar d 40
pr opor t i onal band 76, 107, 145
pulverised fuel, bur ni ng of 35
pulveriser
cont rol 81 - 91
performance adj ust ment 77
pump
feed-water 15, 21
power consumpt i on, reduct i on of 130
reheat spray 152
vari abl e-speed 129
pump charact eri st i c, effects on feed-
wat er cont rol 11 l
purge air, for flame scanner 96
al ar m on loss of 97
purge, furnace 92
by BMS system 45
quick-opening, characteristic, valve 123
r adi ant superheat er 14
RAE 10
r angeabi l i t y
of at t emper at or s 137- 40
of feed valves 121
Ranki ne, W. 7
Ranki ne cycle 7
RDF 26, 35, 42
reci rcul at i on, of fuel oil 39
recuperat i ve ai r heat er 27
refurbishment, of control and
i nst rument at i on system 158, 202- 3
case st udy 202
refuse-derived fuel 26, 35, 42
regenerat i ve ai r heat er 27
registers, ai r 78
reheat 16
spray system design 151
spray at t emper at i on, effect on
efficiency 151
reheat er 16
rel i abi l i t y 45, 172
effect of impulse pi pework on 161
predi ct i on of 174
rel i abi l i t y run 175, 185
reservoir, hydraul i c 169
reset t ermi nal of cont rol l er 148
220 Index
reset wind-up, in steam-temperature
control systems 145
response
of DCS 157
of electric actuator 168
of fuel and air systems 66
of master systems 57
of pneumatic actuator 168
reversed sense, drum-level measurement
114
roll-off, filter 169- 70
Royal Aircraft Establishment 10
safe operation, under failure conditions
157
safety 46, 70
defining requirements for, in
Technical Specifications 185
saturated steam, basic definition 3
saturation, in positioners 166
saturation temperature
basic definition 3
changes with pressure 104
necessity of avoiding approach of 148
SCADA 202
scanner, flame 45, 96
scavenging 19, 45
screen, braided, of cable 172
sealing glands, leakage at 19
seat, valve 121
secondary air, flow control 93
segregation of earth connections 172
sensible heat 2
sensors, steam temperature 143
separator, magnetic 42
sequences 95
setting leakage 65
shift operation 133
shrinkage 104
shut down, plant, time required for 155
single-element feed-water control 104
site-acceptance test specification 185
size coefficient, valve 124
slagging 42
slave controller, multi-stage system 149
smoke 26, 66, 76
effect on flame detection 46
SMP (system marginal price) 52
sodium sulphite, use in water treatment
19
soft desk 201 --2
solenoid valves 156, 195
soot blower, air ingress through 65
sootblowing, effect on drum level 120
spare parts, cost of 200
speed control, fan 33- 4
spill-back, oil 39
split-range control, gas pass dampers 151
spray attemperator
emergency 151
reheat 151
water pressure, maintaining adequate
130
spray water 136
impingement on temperature sensor
141
optimum pressure 137
spray-water control valve 137
stability, of drum-level control system
121
stall
booster fans 92
in axial-flow fans 31
use of plotter to warn of approach 157
stall line, axial-flow fan 32
start-up burners, in waste-to-energy
plant 42
start-up, mill (pulveriser) 90
static, charge, build-up in floating
circuits 155
steam drum 13, 112
in HRSG 14
steam extracted, effects of 120
steam flow, as load indication fi)r oxygen
controller 75 6
steam pressure control, deaerator 131
steam pressure, drum, measurement
used to prevent chilling 148
steam quality 3
steam tables 3
steam temperature control
accuracy with mechanically-atomised
attemperators 138
accuracy with multiple-nozzle
attemperators 139
accuracy with variable-orifice
attemperators 140
steam turbine 15
steam-temperature/steam-flow profile,
convection & radiant superheaters
135
steel-wire armouring, of cables 172
stoichiometic fuel/air ratio 27, 64
stokers 26
Stolze 10
stops, mechanical, on control valve stem
165-6
storage vessel, deaerator 132
stored energy 54
stratification, in flue-gas streams 73
stripping, in de-aerators 19
suction mill 36 7
load-control strategy 84
temperature control 87
summator, use in multi-fuel system 81
superheated steam, basic definition 3
superheater 14, 135
supplementary firing, in HRSG plant
23, 46, 63
suppressed range, steam temperature
transmitter 147
surge
booster fan 92
centrifugal fan 32
surge limit, centrifugal fan 32
swell, in drum water level 104
swell and shrinkage, scale of'effects 105
switches,
on DI cards 155
limit 155, 195
switchgear 158
emc considerations 169
system configuration, documentation
204
System Marginal Price 52
system resistance 21, 31,127
tapping-point isolating valve 160, 162
Technical Specification 184
television broadcasts, effects on power
demand 51
temperature
reheat 151
superheat 3, 136
temperature compensation, of gas-flow
signal 81
temperature control of mill (pulveriser)
87
temperature sensor, location in
attemperator pipework 141
temperature-dependence, of gas-flow
signal 81
tempering air 38
terminations, cable 156
thermal efficiency, ofHRSG 11
thermal inertia, of boiler 57
thermal liner 138
thermal shock 148, 150
thermodynamics, laws of 5
three-element t~ed water control 109 12
thunderstorm activity, recording 200
tilting burners 26-7, 150
Index 221
time constant
fuel/air/flue-gas system 76
steam-temperature system 143
topping cycle 59
total loss 64
track, terminal, of control block 148
trackbal1157
training, of plant operators 202
tramp air 65
effect on flue-gas analysis 72
transducers 95, 160, 203
transformers 158
transient, electrical, due to fault 170
transmitter
active 155
connecting pipework size 161
effect on performance 120
hook-up diagram 160
installation of 162
manifold, instrument 162
passive 155
power supplies for 155
redundant 201
specification of 160-3
specifying installation of 160- 3
treated water 132
trim, valve 121
trip
burner 70
turbine 152
tube failure 148
tube mill 36
tuning
feed-water control system 124
positioner, influence of 165
interactions between loops 57, 132,
151
steam-temperature loop 144
turbine
dynamic response of 53
steam 15
turbine following system 53, 55
response of 55, 58
turbine throttle-valves 54
turbine trip 152
turbine-driven tieed pumps 21
turbulence, effect on drum-level
measurement t20
two-element feed-water control system
105-9
two-out-of-three voting 180
two-shift operation 133
two-valve instrument manifold 162
two-wire transmission 176-. 80
222 Index
ul t ra-vi ol et r adi at i on, from flames 98
uni nt er r upt abl e power suppl y 155
unpopul at ed I / O channels 156
upgr adi ng, of cont rol and
i nst rument at i on system 158, 203
uplift cost 52
UPS 155
val ve
charact eri st i c 122
feed-regul at i ng 21, 121
t appi ng- poi nt isolation 160, 162
size, det er mi nat i on of 125
vane control, fan 34, 92
vapour-at omi si ng at t emper at or 140
vari abl e-annul us desuper heat er 139
var i abl e ar ea at t emper at or 138
vari abl e-ori fi ce at t emper at or 1 40
vari abl e speed drives, advant ages of 34
vari abl e-speed feed pumps
advant ages of 21, 129- 30
mai nt ai ni ng adequat e wat er pressure
for at t emper at i on 130
vari abl e-speed gearboxes 34
vari abl e-speed mot ors 34
vena cont r act a 126, 140
vent valves 41
venting, of t ransmi t t ers 162
Venturi, flow-measurement, use in feed-
wat er cont rol 112
Venturi pri nci pl e, in vari abl e-annul us
at t emper at or 140
very l arge-scal e i nt egrat ed circuits 172
VGB Techni cal Commi t t ee on Techni cal
classification systems 186
vi brat i on 158
in axial-flow fans due to stall 31
visible r adi at i on, from flames 98
VLSI 172
vol t age signals 179
vort ex zone, in vari abl e-annul us
at t emper at or 140
wat er col umn, drum-l evel measurement
115
wat er / st eam differential pressure,
i mpor t ance of mai nt ai ni ng 145
wat er-vapour, bl anket i ng effect on flame
emission 98
wet t i ng vol t age 155
Whi t t l e, Frank 10
wi ndbox 78, 93
wi ndbox-t o-furnace di fferent i al pressure
93
window effect, in cascade loops 146- 8
zi rconi um probe, use in oxygen
measurement 73

Power Plant Control and Instrumentation The Control of Boilers and HRSG Systems

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Power Plant Control and Instrumentation The Control of Boilers and HRSG Systems

Hochgeladen von

Copyright:

Verfügbare Formate

Power-plant control

Das könnte Ihnen auch gefallen