Jovic 1992 Process Control Systems Principles of Design Operation and Interfacing

Process Control
Systems
To my sons
Drazen, Dinko, Alan and Ozren
To the memory of the distinguished men and teachers

Emerik Jovic and Zlatko Singer
Process Control
Systems
Principles of design, operation and
interfacing
SECOND EDITION
Fran Jovic
Professor of Control Engineering,
University of Zagreb;
Section Manager
Computer Systems and Intelligent Process Control
Engineering,
A TM Zagreb Company,
Yugoslavia
English Language Edition Consultant:

R.M. Henry,
School of Control Engineering,
University of Bradford
Ignl SPRINGER-SCIENCE+BUSINESS MEDIA, BV

First edition 1986
Second edition 1992
© 1986, 1992 Fran Jovic
Originally published by Chapman & Hall in 1992
Softcover reprint of the hardcover 2nd edition 1992
Typeset in 1O/12pt Times by Graphicraft Typesetters Ltd., Hong Kong
ISBN 978-94-010-5380-8 ISBN 978-94-011-3116-2 (eBook)

DOl 10.1007/978-94-011-3116-2
Apart from any fair dealing for the purposes of research or private study,
or criticism or review, as permitted under the UK Copyright Designs and
Patents Act, 1988, this publication may not be reproduced, stored, or
transmitted, in any form or by any means, without the prior permission
in writing of the publishers, or in the case of reprographic reproduction
only in accordance with the terms of the licences issued by the
Copyright Licensing Agency in the UK, or in accordance with the tenns
of licences issued by the appropriate Reproduction Rights Organization
outside the UK. Enquiries concerning reproduction outside the terms
stated here should be sent to the publishers at the London address
printed on this page.
The publisher makes no representation, express or implied, with regard
to the accuracy of the information contained in this book and cannot
accept any legal responsibility or liability for any errors or omissions
that may be made.
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication data
Jovic, Fran.
Process control systems: principles of design operation and
interfacing / Fran Jovic ; English language edition consultant, R.M.
Henry. - 2nd ed.
p. cm.
Includes bibliographical references and indexes.
1. Process control-Data processing. I. Title.
TS 156.8.J68 1991
660'.2815-dc20 91-26001
CIP
Contents
Preface IX
Part One Systems, Processes and the Role of Process Control

Hardware 1
1 Signals, systems and process control

1.1 Introduction 3
1.2 A system approach to process control systems 4
1.3 Signals 16
1.4 Systems 28
1.5 The basic analytical concept of process control systems 37
References 40
2 The basic process unit 41

2.1 Introduction 41
2.2 The basic process unit 41
2.3 Basic process unit data processing 49
2.4 Process hardware for data input 62
2.5 Process hardware for data output 70
2.6 Auxiliary data for process control 73
References 77
3 Stratification of control tasks and data communication 78

3.1 Introduction 78
3.2 Stratification of computer tasks 79
3.3 Control levels and computer input/output hardware 86
3.4 Characteristics of process control computer systems 97
3.5 A survey of process control computer hardware 98
3.6 Communication codes and circuits 101
3.7 Channel capacity 101
3.8 Types of connection and communication hardware 111
vi Contents
3.9 Practical suggestions and recommendations 121

3.10 Open-ended communication in process control 124
3.11 Communication of smart process devices 125
References 130
Part Two The Role of Software in Process Control Systems 133
4 The relative roles of software and hardware 135

4.1 Introduction 135
4.2 Data processing 136
4.3 Software for process control data processing 147
4.4 Software versus hardware 151
References 156
5 System software 157

5.2 Basic concepts of real-time operating systems 158
5.3 Structure and functions of real-time operating systems 162
5.4 Data and symbols for the operating systems 173
5.5 System software 175
5.6 Cost, safety and reliability of operating system software 180
References 181
6 Application programs and databases 183

6.2 Application program tasks 183
6.3 Structure and timing requirement of application programs 189
6.4 Database and basic process software 202
6.5 Distributed database 214
6.6 Extended real-time software 220
References 226
Part Three The Man-Machine Interface 229
7 Reduction and visualization of data and procedures 231

7.2 Operator-system communication 233
7.3 Visualization and data logging 241
7.4 Recognition of process states 252
References 256
Contents vii
8 Process management and control 257

8.2 Process states 257
8.3 Operator/process interaction - actions, commands and timing 264
8.4 Process protection and automatic actions 278
References 286
9 The role of the operator in process control systems 287

9.2 Formalization of the operator's role 289
9.3 Artificial inferencing tools - expert systems for process
supervision and diagnostics 289
9.4 Controlling the operator's work using a process control system 298
9.5 The operator's reaction to process control systems 299
9.6 The limits of the operator's function 299
References 301
Part Four System Design 303
10 The feasibility study 305

10.2 Data volumes and flows 308
10.3 Stratification of process control and basic system design 311
10.4 Cost/benefit analysis 317
10.5 Benefits 319
References 323
11 Computer control system design 324

11.2 Communication design 325
11.3 Data transmission units 330
11.4 Designing control and dispatching centres 341
12 Cost-effective system selection 350

12.2 Buying and testing hardware 352
12.3 Designing and testing software 358
12.4 Human factors in system assembly 372
12.5 Programming teams 381
References 381
viii Contents
13 The integrated approach 385

13.2 Mounting and installation procedure 387
13.3 Testing and reliability 396
13.4 System commissioning 404
13.5 Commissioning and safety 406
13.6 Training personnel 408
13.7 Maintenance 408
References 415
Author index 419
Subject index 421

Preface
This book reflects the considerable current industrial interest and investment
in process control systems. The use of computer systems in process control
can provide great benefits, and it is estimated that efficiency can be increased
by up to 30%. It is not surprising, therefore, that there have been considerable
efforts by system designers and users to introduce and use such systems.
Process hardware is integrated into a complete production system through
data processing. It is for this purpose that technical specialists (e.g. electrical,
mechanical, electronics, communication and process engineers and program-
mers) are involved in data processing.
The scope of this book is therefore to assist in the selection of computer
hardware and software that match the functional specification of the data
processing component of a particular system. The principal points covered in
this book are set out below.
Part One: Production process hardware for a standard process is outlined
and the information processing hardware is described. Large
mechanical process hardware and process information devices
(e.g. sensors and control elements involved in the process)
create a coherent production unit, or system, which can be the
control unit (i.e. the basic process unit). The hardware processes
are described and the mathematics explained. This enables the
application of control laws in order to linearize the process about
its working point, as well as a stratification of process control
tasks. Linearization allows process control constants to be re-
duced and therefore the information capabilities of the computer
and communication hardware are also presented and analysed.
Part Two: Although the majority of process control tasks can be run using
hardware functions, the natural trend towards software process
control as more cost-effective is presented. The software tasks
in a process control system are dealt with in this part and, in
particular, the specific role of software in data processing. The
design of the information system based on the software instaIred
in the computer is discussed. These systems may be small, large,
x Preface
free-standing or networked throughout the plant. Additionally,

data and functions of system software are presented and are
based on an operating system structure. Application programs,
their tasks, structure, formation of databases, as well as extended
real-time programs, are also described.
Part Three: Willingly or unwillingly, operators must interact with the pro-
cess that is running. True process states are seldom apparent to
the operator. By using the approach of complete state descrip-
tion and a reduced presentation to the operator, the only
meaningful approach to operator-system process communication
is achieved. Some aspects of human behaviour and work practice
are taken into account because they can influence how the system
is used.
Part Four: By using methods described in previous parts, a sound practice
and a good approach to process control systems can be obtained.
However, without a systematic approach to control system de-
sign there is only a vague guarantee of successful system appli-
cation. Specific features (e.g. software and hardware design, team
organization and maintenance procedures) have to be incorpor-
ated into a process control system and special attention must be
given to control hierarchy and open-endedness of the system.
The development and assembly of computer-aided process control
presented in this part is based on a feasibility study that con-
centrates on data volumes, basic system design and cost/benefit
analysis.
The numerous standards, recommendations and suggestions cited in this book
are based on the work of national and international committees that have been
set up to create and codify standards in the field of industrial research and
practice. Introduction and utilization of process control systems involve many
types of engineer, as well as programmers, and it is essential that they have
a good understanding of computer control systems; it is for this audience that
this book has been written. However, above all, this book should, by the use
of. good design practice, provide clear guidelines in determining the most
suitable. process control systems for specific processes. Therefore, important
practices and procedures are reviewed.
I should like to thank Jadranka PetraSek, Jasminka Konja, Vesna HurCak,
Duraa Lovasic and Mira Halar for their help in preparing the manuscript and
Predrag Vranic, Manager of Information Systems, Rade Koncar Electrical
Industries and Engineering, for reading the manuscript and making valuable
comments and corrections.
Fran Jovic
Zagreb
Acknowledgements
The following tables and figures have been reproduced here

with the kind permission of the Institute of Electrical
and Electronic Engineers:
Figures 3.2, 3.3 and 3.5, and Tables 3.1 and 3.2 © 1983;
Figure 6.9 © 1980; and Figure 6.17 © 1983.
Part One
Systems, Processes and
the Role of Process
Control Hardware
1
Signals, systems and process control
1.1 INTRODUCTION
A process control system is made up of a process involved in a controlled

system and its control equipment or hardware and operators. A typical con-
trolled system is a chemical reactor or a machine tool head, and its respective
control equipment is the process controller in the chemical reactor or the
positioning system of the machine tool head. Processes are usually analysed
on the basis of utilization, since this reflects their proper cost-effectiveness.
Process control hardware encompasses process instrumentation devices, final
control devices, process controllers (based on computer units) and communi-
cation devices for system interconnection and integration. The role of process
control hardware is to support and implement the different process control
functions such as regulation, on/off control, process protection and process
supervision. The design of this hardware is based on a functional analysis of
the entire process. The entire process can be broken down into basic process
units and controllers which are designed to implement the functions of each
basic process unit. Controllers are implemented for each different functional
system for specific process control functions and communication hardware is
installed to carry process and regulatory data between controllers.
The general rules of signal generation, data acquisition, data processing,
data communication, command execution and process actions are governed
by information processing. The performance of the process control system
therefore is dependent on the effectiveness of the information processing in
the process devices and controllers and on their interconnections, input/output
devices and coordination.
Part One of this book gives analytical and practical tools for the design of
process control hardware from a functional and informational standpoint. Thus
a simple analysis of process control hardware can be achieved by studying
process control devices and controller design.
Chapter 1 describes the functions of process control systems and develops
some analytical tools for the design of process control systems. A typical
process control system includes a process, a process controller and a process
3
operator. Since signals carry useful information between these parts, a de-
scription of signals and their information content is also presented in this
chapter. The hardware systems are also described based on the response function
concept and each system part is considered as a system in itself and then the
parts are integrated into the complete system.
Since the basic functions of such a system can be modelled, implemented
or optimized using computers, a brief description of automata is also given.
Processes are classified according to their utilization criteria. Analytical
relations are given for some linear and nonlinear processes in order to support
their design into bigger system parts. The rules for connecting process and
process controllers are presented and analysed according to the change of the
response function and their static and dynamic behaviour.
1.2 A SYSTEM APPROACH TO PROCESS CONTROL SYSTEMS
1.2.1 INTRODUCTION
This part describes a computer system designed for production control and
the analysis and design of such a system. It also explains what is meant by a
production process and a production plant. A production control system is
developed, installed, used and dismantled by man, therefore the roles of the
designer, engineer, operator, etc. are also considered.
Before a production system can be developed and implemented, it must be
decided whether such a system is really necessary. Therefore, a preliminary
(or feasibility) study should be carried out. The essential make-up of any
process is the hardware, which consists of elements such as pipes, reactors,
vessels, wires, valves, motors, etc. The quality, connection and interrelation
of process elements determine the operational performance of the process
since it is these elements which ensure materials and energy storage, processing,
exchange and recycling. Examples include reservoirs, condensers, heat ex-
changers, heaters, pumps, motors transformers, generators, vessels, reactors,
etc. Integrated into large units, or plants, it is these units that enable produc-
tion - with associated energy and materials transfer - to be carried out. Such
units include power plants, liquefied gas storage, machine tools, rolling mills
and pipeline systems. The integration of units in a production plant is comple-
mented principally by hardware elements such as pipes, valves, power cables,
pulleys, conveyors, gears and clutches.
Individual units in a plant are often fitted with sensing and control elements
- typically, temperature indicators, level indicators, pressure indicators, valve
motor controllers, electrohydraulic flow controllers, pump motor controllers,
etc. Measuring, sensing and controlling the process variables (e.g. tempera-
ture, voltage and flow and control of interconnections and control elements)
can be used to control each particular production unit. An integrated group of
A system approach to process control systems 5
production units controlled in an appropriate way constitutes a production

plant.
Several production processes or plants may be integrated into a larger system
known as a production system, electrical or natural-gas network, which is
monitored and controlled via communication and computer networks.
Control system development starts with a feasibility study, followed by a
design study which includes the design and realization of computer hardware
and software. Finally, to complete the control system development cycle, the
system is put into operation. Two modes of work on control system development
may be distinguished as:
l. Problem-oriented work: this includes the functional specification, speci-
fication of data and information flow analysis. It requires a multidisciplin-
ary approach. (It is important that the customer is consulted at this stage
of the development cycle.)
2. Data-oriented work: this includes the design study, design of computer
hardware and software, implementation and operation changes. It also
involves matching the external functions of the control system with the
chosen hardware and software. Basically, this work is the responsibility
of hardware and software designers. The development cycle of a control
system is presented in Figure 1.1.
1.2.2 THE COMPUTER CONTROL SYSTEM
A typical computer control system for a power plant is illustrated in Figure

l.2. The power plant is divided into power-generating blocks that are moni-
tored from the operator's console. Here all necessary measurements and re-
cordings are made, and all report and alarm warnings are issued. Process
control is performed using an algorithm based on 'block model' software that
controls both set-points and functional groups. A separate protection system
operates issuing shutdown commands to the block as required; signal condi-
tioning, drive controls, interconnections, signal transducers and drives are called
process hardware components. The functions of data processing for supervi-
sion, measurement, recording, issuing a protocol, protection, block and group
control are all performed by a few functional systems. Nevertheless, all these
systems use data from the same set of plant data and therefore the basic
functions of a process control system are formed by functional systems that
are subsystems of the control system itself. In simple process control applica-
tions, the functional system is used as a single process control device (e.g. an
alarm unit of a distillation column, or the current controller of the power
supply in rural areas).
The system is one of several functional systems developed for supervision,
control and data processing in a production process, or production plant, by
preliminary, feasibility
modelling study
Phas:..... _---- ___ ~_ ...... ,
.
/ ... -~-.-.-.-.-.,.,
I analysis and \ , development and
I design of control I i
.
preparation of
process and plant
: system I
j I
hardware
! (
:l
~ realization of
i realization of process
I control system I . and plant hardware
\
',------~--~')
,
r--J-·-·-·j
'. I
implementation
realization
commissioning work
phase
maintenance
working change in production

phase environment
Figure 1.1 Development cycle of a control system: - - - -, control system

development; - . - . - " process hardware development.
means of the collection, storage, processing, dissemination, interpretation and

communication of information.
1. The information is a collection of data about events or other data derived
from the data processing which are then interpreted or supplemented.
2. Data are information elements based on signals.
3. A signal is a common term for any physical quantity by which any system,
or system element, influences other systems or system elements. The
existence of information, as well as its processing, interpretation and
manipulation, is made possible by an information carrier.
4. Information carriers might take the form of the human operator, material
and energy processes, data processing equipment, etc. The human operator
is an essential information carrier for any production process or plant.
crt displayed
data recording
I Ul!J CJ ...
1""I11III
....
-eo @ ~ ~ protocols ' plant-'
supervision
-= I
..i. ~. ,
r.
1 -iii " ...r block control
'3 ..... l J .
~
- 1 1
; ~
~
"
;.
I
~
1 set-point functional protection data
' lid
~ group processing
,g ........ . controls system
...
C:: . controls and alarms control
ai '
. systems
1 ..1! .
~ 10 1 .1 "..t..
signal conditioners/ a _< . ~
analog digital drive controls ':;:'

~ ...
.
L - " .. .'
".... i1
J
J1
~ . i ~. -*. III
-.
~ '" II-'
~
"
.~~
~
, tL .1I
l ;; - :.~
iDterf~iDg~~~t
..."... . . . -. · ~
J in terconnection
l J
1
signal 0
transducers ~ drives
l l ~~ oj p~wer plant block
Figure 1.2 A typical computer control system of a power plant.

Due to specific limitations on speed and operable information capacity when

involved in complex control of production process plants, the abilities of the
human operator may be improved or replaced by digital electronic devices -
mostly computers. This is particularly so in the role of information process-
ing. Figure 1.3 shows comparative data on human and computer process control
speed and operable information capacities.
Personnel involved in computer control system development and application
usually work in teams organized by the customer and the manufacturer. Since
process control systems vary in their function, content and magnitude, it is
not recommended that any specific development model be adopted. Nonetheless,
there are jobs which have to be done by the customer and the manufacturer,
and these are illustrated in Figure 1.4 in which activities (1), (2), (3) and (7)
are executed mainly by the customer, whereas activities (4), (5) and (6) are
predominantly carried out by the manufacturer.
Data in Figure 1.4 do not include hardware production and assembly
carried out by the manufacturer. Selecting, buying and installing process
hardware are predominantly the customer's responsibility. The cost of hard-
ware is generally estimated not to exceed 25% of the cost of the total system
and, of the total estimated work on the system, 59% is usually done by the
manufacturer.
Some specific difficulties encountered by people working on the develop-
ment of a new system are set out below.
1. Communication with specialists in other technical disciplines.
2. Delays in system development.
3. System costs underestimated by a factor of 2-3.
4. Poor starting, inadequate briefing of requirements.
5. Missing or incomplete, or inadequate, or unintelligible documentation on
system analysis and design.
6. Software maintenance requiring more and more resources, originally planned
for the development of a new system.
It would seem that the main way to overcome these difficulties is by teamwork.
Careful control of a team's objective, strength and work should be performed
in an environment where each member may take the role of a specialist. A
typical allocation of resources for a process control system development is
also outlined in Figure 1.4.
1.2.3 TYPES OF PROCESS
A process is defined as a set of operations that perform physical or chemical

transformation, or a series of transformations. Such processes might include a
production process, plant, transportation of matter and energy, transmission
and processing of information used for production and processing of material
human "-
"-
./
"
/
/
/ memory
up to lOIS (bit) ,
\
I
/ operational
\
\
/ memory speed \
/ 100 (bit/s)
\
I \
motives
I \
I
I
I
I
•
\ reflex paths I
\ '4 I
\ muscles ------ ... "- receptors I
/
\ actions ./
"- stimuli /
~ external world /
/
.-
computer control system
" ,,
I \
\
\
I
memory \
I I
108 - 10 10 (bit) \
main memory
I
lOs - 10 7 (bit)
I operation speed \
\
10 6 - 10 7 (bit/s) \
\
I
I
up to 106
,,
\
\
I process I
\ transducers /
'--------_.
\ /
,'--------~~ ./
process and interface

devices
Figure 1.3 Comparative data on speed and operational memory for human and
computer control systems.
System deliverer Customer
lOO% 50% 50% 100%
I I I
requirement definition
~s I 14% I and analysis (1)
1 22 17% nary design (2)
%18 A l I
,l 19% I main projec
21%
I coding and debu
~ factory testing (4
~s {/\ I
l 5% I hardware installation U
1 13% I 8 12% installation and work (6)
II site installation;
I
e maintenance one
I 12% I ,er handover (7)
I 59% 41% I total efl
Figure 1.4 Typical allocation of the activities carried out during process control
system development: c::::=J, data according to Wolverton (1974); data
0,
according to Koch and Hoffmann (1978); 6., data according to Siemens (1990).
production
process
Figure 1.5 Production environment aspects.
goods, information and energy. The process is a controlled system and forms
part of a set of production or processing functions executed in and by means
of process hardware. The features of a process are usually measured by process
quantities, or conditions, termed process variables. The control of process
variables is achieved by the control equipment (e.g. electromagnetic valves,
transformer tap positioners). The processes are situated in the production
environment and they are affected by human, time-space, materials/raw
materials, energetic, socioeconomic, informational and environmental aspects
(see Figure 1.5). These aspects are interrelated in a complicated way. Table
1.1 shows the main process descriptors according to the given production
environment aspects.
1. Human aspects mostly determine the mode of process control. The given
number of operators, their work, educational background and qualifica-
tions, and process control responsibilities are dependent on the particular
Table 1.1 Main process descriptors
Human aspects Time-space Materials and Energetic aspects Socioeconomic Informational Environmental
aspects raw materials aspects aspects aspects
aspects
Manually Fast/slow process Fluids Energy-generating New process Process control Environmentally
controlled process process information safe process
Manually/ Concentrated/ Raw pieces Energy- Capacity additive Supervisory Process with
automatically distributed transforming process process removable
controlled process process process information contaminator
Automatically Continuous/ Single parts Process with Capacity Auxiliary process Process with
controlled process discrete process energy supply replacement information unremovable
process contaminator
Automatically Aperiodic/cyclic Sheets and
controlled process process textiles
with learning
Automatically Deterministic/
controlled process stochastic
with learning and process
prediction features
process type and the control system applied. Human aspects seem to be
the most influential factors in proper system functioning and overall pro-
cess productivity.
2. Time-space aspects determine the character of the process type, showing
five main process characteristics: speed (slow, fast), spacing (concentrated,
distributed), continuity (continuous, discrete), periodicity (cyclic, aperiodic)
and determinacy (deterministic, stochastic). Time-space aspects influence
the complexity, automaticity and sensitivity of a particular process.
3. Materials/raw materials aspects show the four types of materials handling
processes: fluid processes, raw pieces, single parts and processes for sheets
and textiles. Materials/raw materials aspects influence the type of process
variables, process sensitivity and the type and extent of process control.
4. Energetic aspects determine the main process categories: energy-generating
process (e.g. a chemical reactor), energy-transfonning process (e.g. a
hydroelectric power plant) and process where the energy is supplied for
materials processing only (e.g. a machine tool). Energetic aspects are
prominent in the calculation of process production efficiency and affect
the type and extent of process control applied.
5. Socioeconomic aspects determine the adaptation to the motives and inter-
ests of people engaged in the production process. Consequently, the main
types of process development that can occur are purchase of new equipment
for a new process, enlargement of existing production capacity or simply
replacement of existing production capacity. All estimates are based on
the analysis of future expenses and on expected benefits.
6. Informational aspects determine the selection of process instrumentation
and data processing equipment.
7. Environmental aspects determine the environment protection counter-
measures.
According to the main descriptors given in Table 1.1 there exist 8100 different
process types. All the processes have four features in common:
process economy;
process capacity;
process speed;
process safety.
Different stages of these process features have led to different stages of process
complexity and these are shown in Table 1.2.
To obtain certain qualitative and even quantitative relations between process
features, and some production environmental aspects, a simplified model is
shown in Figure 1.6. For instance the amount of material on the market de-
creases with the increase of product and investment costs. The extra demand
for goods on the material market promotes production, which in tum increases
the size of the labour force needed to make the goods and thus increases
Table 1.2 Main classes of process complexity
Control source Type of process response Process complexity level Process capacity
Human Variable response Handtool Increasing

Handtool with auxiliary energy supply
Machine tool with manual control
Control mechanism with Fixed response inside a Fixed cycle machine tool Increasing
predetermined pattern of work process Programmed cycle tool or machine
Remotely controlled machine
Machine activated by process
Process and environment Response with a signal Measurement of process parameters Increasing
variables Measurement and signalization of process
parameters with change in control function
Process supervision and control
Response with action Process behaviour changes with measurement Increasing
Limited set of possible Elimination of incorrect process products after
determined actions measurement
Choice of corresponding process cycle
Action after identification of certain states in
process or environment
Broad set of possible Correction of behaviour after the action Increasing
process responses Correction of behaviour during the action
Anticipation of system behaviour according to
process variable states
investment production material

costs market
operating product
labour market
costs
Figure 1.6 A simplified relation model of the production process and some
production environmental aspects: • >. increasing action;~.
decreasing action.
operational and investment costs. A rise in product market also increases the
production. The model in Figure 1.6 may also be used for quantitative pur-
poses by determining the conversion factors Ai,i+l between specific produc-
tion factors for a particular product, and by determining initial quantities and
time relations ..!lti,i+l between the causes and consequences for specific pro-
duction factors.
The following main categories of process control system are recognized
according to the application field.
1. Energy generation, electric power transmission and distribution systems.
2. Chemical and petrochemical industry.
3. Metallurgical industry.
4. Traffic and transportation systems.
5. Cement, paper and pulp processing industry, and food and fermentation
industry.
6. Environmental systems.
7. Metalworking and electrotechnical industry.
8. Mining.
9. Instrumentation and laboratory equipment production.
The export of engineering products for the above industries covered 70% of
the total world exports for engineering products in 1987 (Bulletin of Statistics
on World Trade in Engineering Products, 1989), making a total of approxi-
mately $168 billion.
The amount of control equipment for the engineering products of these
same industries is calculated to be between 1% and 5% of the total.
1.3 SIGNALS
1.3.1 INTRODUCTION
The influence of a process element, a system part or a system upon other

elements, system parts or systems by any kind of physical quantity can be
considered, broadly speaking, as a signal. Thus a signal, besides its physical
quantity and concrete effect on a process or system, also carries information
contained in its signal parameter. Essential to the process and process control
is the identity of the process effect and the signal parameter value. The most
usual signal parameter value is the value of the signal's physical quantity but
it can also be a part of this or the fluctuating value over its mean value. For
example, when the vibrating force is considered, the static pressure of a body
on another body is not considered as a signal. Simply speaking, it has no
informative value. Thus, signals usually describe a certain aspect of an in-
teraction between two system parts.
Besides the informative aspect of the signal its direction is also important.
Signals are always directed values, which means that they have an origin, a
signal source and a destination or a place where they can be interpreted and
used. The signal source is the output signal for one process part that issues the
signal and the destination is the input signal for a second process part which
receives the signal. The different signal sources and destinations in a simple
typical process control are illustrated in Figure 1.7. Thus, a signal can carry
data on what is actually happening during a process or take the form of a
command to control the process.
Signals differ in type, value and physical form to the extent of the variations
of all possible physical values and forms. In order to sacrifice diversity to
technical and descriptive simplicity, there exists practically only a small number
of technically feasible signal forms that are traditionally named, according to
their forms, analog or digital signals. Both forms are further standardized, so
that only a very small number of distinctive signal forms is applied in process
control. For example the 1151 GP pressure gauge transmitter of Fischer
Controls has an output of 4 to 20 rnA DC or 10 to 50 rnA DC for the whole
range of pressures.
Signals 17
---------,
output signals data
display
control
board
[[I input
valves
sensors [[I signals
PROCESS
input signals
00- signal source

[Q]- signal destination
Figure 1_7 Signal sources and destinations in a typical process control system.
Analog signals can be either continuous or quantized but digital signals are
always quantized signals. The most important types of signal for the presen-
tation, analysis and execution of control functions are time-discrete signals,
because they represent the actual data scanned, acquired and processed in
process control systems.
Signals represent the result of actions in process control systems and they are
measured and registered at different points in the process or in the process con-
troller. These signals can be put into an analytical form which allows quick
comparison, analysis and the preparation of data processing algorithms in
process control. Quantized signals are analytically described by the z-transform
and continuous signals by the Laplace transform. These analytical tools are
briefly presented in this chapter, in order to support the further system design.
Signals carry information and the information content of signals can also be
given analytically. Data are formed as based on the information content in the
signal processing part of the process control system. Thus information content,
data acquisition, data processing and process control execution are highly
influenced by signals.
Some process control data are communicated to the process system in a coded
form through communication channels in order that their content is not affect-
ed by noise. The basic analytical features of these channels are also de-
scribed. Details of process control data communication are given in Chapter 3.
1.3.2 TYPES OF SIGNAL

Signals are categorized by the form of their signal parameters. If the amount
of a signal parameter has any value in a given measurement range, then it is
considered as an analog signal, but if its amount has only quantized values,
then it is regarded as a digital signal. Analog signals fall into two main types.
1. Time-continuous analog signals, such as the pressure gauge signals.
2. Time-discrete analog signals, such as pressure indicator signals.
Digital signals fall into two main types.
1. Binary signals, such as proximity detector signals.
2. Count signals, such as count rate pulses from a liquid flow meter.
These four types of signal are shown in Figure 1.8.
There are some standard signal forms called singular signals which are used
for signal, system and process analyses. These signals can be aperiodic or
periodic, and are given in Figure 1.9 as:
1. The unit step signals, S(t), which has the value one in time instant to and
lasts infinitely long, and before this time instant it equals zero.
2. The unit pulse signal, del(t), commonly named the delta function, which
has the 'infinite' value within a very short time interval to - e, to + e, in
all other time instants it equals zero, and the total signal intensity of the
signal equals one, i.e.
I to+e
to-e
del(t) dt = 1, e ~ 0 (1.1)
3. The square pulse signal which equals one in a given time interval tlo t2
and outside this interval it equals zero.
4. The stationary sinusoidal or cosinusoidal signal, i.e. the signal which
changes its value according to the following equation
u(t) =a sin (COot + CPo) (1.2)
Example
The process liquid tank given in Figure 1.10 should be provided with necessary
signals for its proper functioning. The levels in the tank are:
L 1, level of pump deactivation;
- L2, working level of pump;
- L 3 , level of security output.
Solution:
The following signals should be provided:
Llo digital signal for the low level indication;
L 2, digital signal for level indication and pump on/off control;
L3, digital signal for level indication and pump on/off control;
S 1, digital or analog level indicator and alarm signal for the total shutdown
function.
y - A • time continuous analog
~----~ signal
B . time discrete analog

----..!' signal
C • binary digital signal

~
!
I
I ~ . t
r' D • count digital signal

~------~ ~---------
L-______~'____________~______~______~ ______~L__ _ _ _~t
count signal instant:
Figure 1.8 Signal parameter values for analog and digital signals.
T
del(t)
J'o+e
,o-e del(t) dt =1
Figure 1.9 Singular signals.

Signals 21
• input
81
L3 --- ---------------------------~=== ~ output
L2 -- ------- ----------------- - - -----
LI--- ----------------------------/--===~ forced output

pumpP
Figure 1.10 Signals in a process liquid tank_
1.3.3 TIME-DISCRETE SIGNALS
Any analog or digital signal can be observed using a scanning procedure,

usually on the side of the controller, over short time intervals. The intervals
between scanning are much longer (say at least ten times) than the observa-
tion intervals. By means of scanning, a discrete row of signals is obtained
from any signal (see Figure 1.11). Such signals are called sampled signals.
When the sampling instants are equal, signal u(t) is transformed into a sampled
signal us(t) according to the following relationship
n=oo
us(t) = L u(t) del(t - nT) (1.3)
n=-oo
Sampled signals are better presented by the z-transform, which transforms
equation (1.3) using Table 1.3 into
0=00
u(z) =L u(n)z-O (1.4)

0=0
for u(t) =0 when t < 0_
1.3.4 A DESCRIPTION OF SIGNALS
Signals usually represent complex interactions between process parts or complex

process features, and their proper description and analysis can give valuable
information on the system and the behaviour of the process. Signals are
represented in many standard ways and forms. The most usual descriptive
forms are:
1. Time-function diagrams where signals are depicted according to the time
dependency of their signal parameter; the best feature of this description
, u(t)
a)
b) us(t)
/
/
/
t
us(t)
c)
d)
u(3)z-3
1
_
u(9)z-9
.,t
Figure 1.11 Sampled and discretized signals.
is its plausibility but it lacks accuracy for long time intervals and it is not
representable to computers.
2. Tables where the signal is presented either as compared to other signals
or in given time intervals; signal tables are usually not easily surveyable
by the process operator, but computed signals can be stored in the com-
puter as tables of raw data.
3. A sum of a series of singular functions, as shown in the three examples
in Figure 1.12; such a description can help in the fast analysis of response
signals in process control.
4. The Laplace transform of the signal time function (see Table 1.3); by
Signals 23
Table 1.3 Mathematical descriptions of signals
Signal Laplace transform z-transform Comment
u(t) F(z) = L f(n)z-n f(n) = u,(t);

n=O
-t = 0, 1, 2, 3, ... , n, n + 1
t.t
As AZ( Z - cos root)

A cos root
S2 + ro6 Z2 -2zcosroot+l
del(t - to) Z-k del(t - to) = d(n - k)
S(t) 1 - Z-l
S(t) = S(n)
s
S(t)e at at = an
s-a 1 - az- 1
Z-l
S2 (1 - Z-I)2
t =n
combining the representation in a series of singular functions and the

Laplace transfonn, a rather simple description of the signal behaviour can
be obtained.
5. The z-transfonn of the signal time function (see Table 1.3); the example
of the z-transfonn of a sampled signal has already been shown in equa-
tion 0.4).
1.3.5 THE INFORMATION RATE AND CONTENT OF PROCESS

SIGNALS
Signals represent infonnation only when they carry a message. A message is

composed of signal characters. A character is an element in a mutually agreed
finite set of different infonnation elements. Process signals possess a very
simple infonnation content because they have a specifically simple character
set. The character set of different analog and digital signals is given in Table
1.4. An analog signal has a continuous set of messages but due to the analog-
to-digital conversion, only eight or ten bits of the infonnation content can be
distinguished. A digital signal is a digital message source which has only two
messages, i.e. 0 and 1. The infonnation rate of a signal is the maximum
change of messages in a time unit. The infonnation rate of the analog signal
as shown in Figure 1.13b is expressed in the following relationship
u(t)
u(t)
u(t)!
I
I
I
I
I
I
I
I
I
I
,,
I
Figure 1.12 The discretization of continuous signals.

Signals 25
Table J.4 The character set of analog and digital signals
Signal Character set Example*
Time-continuous Binary numbers 8, 10 or 12 bits 00110110 = 54

analog signal with/without sign B D
Time-discrete analog Binary numbers 8, 10 or 12 bits
signal interrupted by a row of 'out of
range' characters
Binary digital signal Binary numbers 0 or I in a series
Counter state digital Binary or binary coded decimal 0110 1001 0001 = 691
signal numbers; up to ten decades BCD D
Count digital signal Binary states 0 or 1 in a series. 0101 = 1
Change of 0 ~ I represents the S D
count rate event
* B, binary; D, decimal; S, serial; BCD, binary coded decimal
rA = (dA) ~Q = 5(V /s)~Q (char/V) = 5~Q (charfs) (1.5)

dt MAX
supposing an ideally fast analog-to-digital converter. The information rate of
the digital signal from Figure 1.13a is
rD = l/Tpl = 1000 (charfs) (1.6)
When a signal is sent through any kind of information medium, or informa-
tion channel, the transmission of messages from the signal source to the sig-
nal destination has to be considered. The information channel is analysed in
Chapter 3. The information content of a signal, when treated as an informa-
tion channel, is equal to the negative logarithm of the probability p(Xj) of the
appearance of the message event (character) Xi from the whole set of poss-
ible message events, i.e.
(1.7)
The logarithm basis of equation (1.7) is agreed to be equal to two. Thus the
information unit is defined as a binary amount of the information named
Shannon or bit (binary digit).
Example
Determine the information amount which can be obtained from the occur-
rence of one character out of 16 equally probable character events. Accord-
ing to equation (1.7) where ld = dual logarithm (lOg2)
26 Signals. systems and process control
d igi tal signal

Ds{t)
a)
analog signal
LlA = 1 V
As{t)
~~~----------------~
b)
Tp2 = 0.2 s
Figure 1.13 The information rate of digital and analog signals.
I = ld 16 = 4 (bit) (1.8)
The entropy of the signal source is the mean value of the information amount
per character out of n independent characters. It is equal to
(1.9)
The entropy of the analog signal HA (observed over longer time intervals, so
that the measured values are independent) when converted by an eight bit
analog-to-digital converter is equal to, Pi = 1/256,
HA = 256. _1_ .li = IIi = 8 bitl = 8 bit (1.10)

256
Signals 27
where I I are explanatory brackets for determination of Ii' The entropy of the
digital signal is one bit. The mean information rate is
R = r . H (bit/s) (1.11)
where r is the mean number of signal characters in a time unit. The analog
signal with the mean number of characters given in Figure 1.13b as 5i1Q
(charfs) has a rate
RA = 5 . 8 = 40 bitls 0.12)
and the digital signal from Figure 1.13a has a rate
RD = 1000. 1 = 1000 bitls (1.13)
1.3.6 COMMUNICATION OF SIGNALS
For a given signal with a mean information rate r and for a given communication
channel with the capacity C, where r < C, there is always a way of how the
signal can be coded into another message source so that the information
transfer can reduce the amount of errors and be independent of the noise in
the channel. Process signals are unfortunately not coded into other signal
forms (unless processed by a 'smart transmitter'), but the necessary channel
capacity can be calculated according to the relationship
C = B . Id( I + SIN) (1.14)
where B is the width of the frequency band (Hz) of the information channel
and SIN is the signal to noise power ratio.
Example
When B = 20 kHz (an audio channel), S =1W and N = 1 mW, the channel
capacity is
C = 20 . 10 3 Id 1001 = 200 kbitls ( 1.15)
1.4 SYSTEMS
1.4.1 INTRODUCTION
A system is by definition a set of elements or objects which executes a given

function (or a set of functions) according to a given or agreed specification.
Essential for the system is its environment. A process and a process controller
are systems in themselves and they can be analysed and synthesized as sep-
arate systems. Their combined operation as a process control system can also
be treated using system analysis and synthesis. Thus a 'system' approach is a

practical approach to the design of process control systems.
Systems are time variant or time invariant and both types are presented.
Furthermore, systems are either linear (or linearized around the working point)
or nonlinear. Both can be categorized into three types.
1. Causal, meaning that there was no excitation in the system before t < 0
when regarding the system output at t ~ O.
2. Time instant, meaning that the response is an instant answer to the input
amplified by a given factor a, i.e.
y = au a ~0 (1.16)
3. Dynamic, meaning that the system contains certain memory elements of
analog (e.g. condensers, tanks) or digital type (e.g. on/off valves, switches,
bistables).
Furthermore, systems can be time continuous or time discrete (sampled data

systems). Thus, even using this simple classification there appear to be 24
different system types.
Any system can be reduced to its subsystems, elements, etc. Each subsystem
can be analysed and synthesized as the system possessing the response at its
output y(t) as the result of the input action u(t).
The most interesting systems are discrete systems because any process
connected to the computer represents a sampled data system where the process
controller samples process signals at given time instants and issues process
control signals at other time instants. Various system parts are put together in
order to synthesize the system. The rules of producing structural schemes of
systems are discussed here. When these structures form a loop connection, a
drastic change takes place in the behaviour of the system. The type and cause
of such behaviour is also analysed. Some system parts are processes and some
are controllers. Controllers are mostly computers of any type. Computers process
their input/output data according to the rules given for automata. These rules
will also be considered and modelled. By changing these rules different trans-
fer characteristics of computers as process control parts can be obtained and
the whole process control system can be synthesized.
1.4.2 TRANSFER FUNCTION OF TIME-DISCRETE SYSTEMS
Time-discrete systems have time-discrete excitations at their inputs and time-

discrete responses at their outputs (see Figure 1.14).
Unit function d(n - k) is a singular signal used for the time analysis of
time-discrete systems, i.e.
Systems 29
input 0 j'--__Sy_s_t_em_--' 0 output
//
/
I
Figure 1.14 The time-discrete system.
den _ k) = {O
1
for n
for n
*"= kk ( 1.17)
Time functions of a time-discrete system can be written as
fen) = L f(k) den - k) (1.18)

k=-=
The time response to any input can be expressed as the response to the unit
function
hD(n, k) =F den - k) ( 1.19)
i.e. the response in the nth time instant to the unit function which is fed to the
system in the kth time instant, where n > k for causal systems. Any input
signal can be represented as the sum of a modulated unit function, i.e.
uD(n) = L uD(k) den - k) (1.20)

k=-=
The system response is
YD(n) =F L UD den - k) =L UD . F . den - k) (1.21)

k k
YD(n) =L uD(k) hD(n - k) (1.22)

k
i.e. the output signal is a discrete convolution of the input signal and the
system response to the unit function.
Example
The response of a time-invariant discrete system to the unit function is shown
in Figure 1.15. As depicted in Figure 1.I5b, there is a consistent time invariance
of the system response to the unit function. The input function uo(n) is shown
in Figure LISe. The output is calculated as follows:
for n = 0, Yo(O) =0
for n = 1, yo(l) = uo(O)h(1) + uo(1)h(O) + uo(2)h(-I) = 2
for n = 2, yo(2) = uo(1)h(1) + uo(2)h(0) = 7
for n = 3, Yo(3) = uo(1)h(2) + uo(2)h(1) + uD(3)h(O) = 7
for n = 4, yo(4) = uo(1)h(3) + uo(2)h(2) + uD(3)h(1) + uo(4)h(O) = 8
for n = S, yo(S) = uo(4)h(1) + uo(S)h(O) = S
for n = 6, Yo(6) = yo(5)h(1) = 1
The output function is given in Figure LISe.

For a time-variant discrete system the following procedure for the calcula-
tion of the transfer function and system response can be followed:
- because response ho(n,k) depends on the time instant of observation (nth

time instant) and on the time of application of the input unit function (kth
time instant)
wo(n - k) = hD(n, k) (1.23)
- for input function
uD(n) = L uo(k) d(n - k) (1.24)

k
there exists the output function
uD(n) =F L uo(k) d(n - k) =L uo(k) Fd(n - k)

k k
= L uo(k) hD(n, k) (1.2S)

k
- by substituting n - k = m, equation (1.25) for the output function equals
yo(n) = L uo(n - m) ho(n, n - m) = L uo(n - m) w(n, m) (1.26)

k k
where
ho(n, n - k) = wo(n, m) (1.27)
is the weighting function of the time-variant discrete system. (See the
analogous situation for continuous systems in Figure 1.17.)
(a) t d(n)
-...2
hn(n)
1 t I"n n
0 0
+ hn(n)
~. ~: t~~._ _ _ _._
(b) d(n)
n
K K
7 7 8 5
_LI. L1 I
2
2 3 4 5 2 3 4 5 6
Figure 1.15 The response of a time-invariant discrete system.

1.4.3 TIME RESPONSE OF LINEAR CONTINUOUS SYSTEMS
Time-invariant linear continuous systems are also characterized by their response

to singular signals in the time domain. The basic singular function is the unit
function, i.e. the input to the system
u(t) = del(t) (1.28)
and the response to this input is h(t), where
u(t) = del(t) ~ y(t) = h(t) (1.29)
The response to the unit step function is
u(t) = S(t) ~ y(t) = F(t) (1.30)
and as the unit function is the derivative of the step function, the following
relation holds
dF/dt = h(t)
An example of F(t) and h(t) functions is given in Figure 1.16.
Excitation u(t) of any form can be written as
(1.32)
and the system response is given as the continuous convolution of the input
and response function, i.e.
(1.33)
The time response of the time-variant linear system is given as
(1.34)
where
O>(t, 't) = h(t - 't, 't) (1.35)
Figure 1.17 depicts a way of presenting the calculation of output function y(t)
under time-variant conditions.
1.4.4 CLASSIFICATION OF LINEAR SYSTEMS

Both continuous and time-discrete linear systems can be classified as:
1. Causal systems, i.e. systems where the response lags behind the excitation
and input u(t) causes output y(t) but the singular responses are equal to
zero for t < 0 and the relationships hold
Systems 33
Process control systems
F (t)
I
I
I
I t
~~~----~------r---~------T---~T-----"
~
dF(t)
h ( t) =----
dt
Figure 1.16 A response of a linear continuous system to unit function and unit
step excitation.
h(t, 't) = 0 for 't < 0

(1.36)
oo(t, 't) = 0 for 't < 0
2. Instant response systems, i.e. systems where the output follows the input
according to the relationship
y(t) =a u(t), h('t) = oo(t) = a del(t) (1.37)
or for the time-variant case where
y(t) = a(t) u(t), h(t b 't) = a(t,) del(t) (1.38)
oo(t, 't) = a(t) del(t) (1.39)

34 Signals. systems and process control
w(t,'r}
T
Figure 1.17 The calculation of the system output function under time-variant
conditions.
3. Dynamic systems where the response depends on past excitations; these

are systems with memories containing information on past system inputs
and states.
A linear dynamic system is given in Figure 1.18 where S-l depicts an inte-
grator and Z-l is a unit delay (for time-discrete systems).
The following relation can be used for a time-continuous dynamic system
x = Ax + Bu (lAO)
Y = Cx + Du
The transfer function of the system is given as
= C¢B + D
F (lA1)
where ¢ = [s - A]-l is the characteristic matrix of the linear system.

The time-invariant discrete linear dynamic system can be described by the
following relationship
zXz = Ax z + Bu z
(1.42)
Yz = CX z + Du z
where the transfer function is
Fz = A[z - A]-l B + D (1.43)
Systems 35
II j X S-l X
---
~ B
Z-l
~
r--- - C :> y
-=== A I
Figure 1.18 A linear dynamic system.
Example
A delay line with the delay of 10 ms when scanned at a rate of 1 ms has the
following properties as a dynamic system
A = C = 0, B = D = 1 and transfer function
(1.44)
1.4.5 STRUCTURING OF SYSTEMS
Linking subsystems into a system can be done in the simplest way by regarding
the flow chart of signals in a system. The places of subsystems in a system
can be exchanged in linear systems (or linear parts of them). Let us suppose
that the system has the structure given in Figure 1.19a. When the paths of
signals through the system are followed, the input/output relationship can be
obtained
y = H 4H 2H 1u + H 4H 2H 1v + H 5H 4H 2y (1.45)
where
v = H4"ly (1.46)
The same system can be rearranged differently as shown in Figure 1.19b
supposing T = Hi! and 4 = H4"l, or after solving equation (1.45) as
F = (H 1H 2H 4)/[1 - (HIH2H4)(H5/Hl + H3/H4)] (1.47)
as depicted in Figure 1.19c.
5
a) I
I
v
2 4
---- y
,
b)
rl T
1 5
I
I
f1
I
u
~I 2
1 4
~~ .y
- -1
1 "3
- -1
·1
"5 =H- 1
4
"3 =H- 1
1 =H l' 4=H4' 5' 3
c) I--------PROCESS--~-~~---- --,
I ~
I
u ~I
1 2 4
I
y
I I
L _____ ~ ____ ~_~~ __________ J
r---~----~~--~---~~--~----~-I
r
I I
I I
I 1 5 I--- I
I I
I
I I
CONTROLLER I
I I
I
I
I
I
'--- 3 4 - I
I I
L ______ ~ _______ ~ _____ ~ _____ ~ _~~
Figure 1.19 Signal paths and structuring of a linear system.

The basic analytical concept of process control systems 37
1.4.6 SYSTEMS AND AUTOMATA
Computers are used as process controllers. They exhibit different transfer

functions when connected to the process. By changing the computer transfer
function, significant benefits can be gained in process control. Computers are
digital automata which change their internal states according to previously
obtained states and actual inputs. At the same time the computer activates
certain outputs connected to other system parts. A simple set of computer
states is presented in Figure 1.20. The change of states happens according to
the change of inputs. By combining internal computer states, computer input
and output the response function can be obtained as required for a specific
process control. An example of input/output functions is given Figure 1.20b
and c.
1.5 THE BASIC ANALYTICAL CONCEPT OF PROCESS

CONTROL SYSTEMS
A control system is made up of a controlled system and its control part. The
control system can either be a closed-loop or an open-loop system.
1. The closed-loop control system, where the control action is made de-
pendent on the measurement of the controlled variable, usually adjusts
the manipulated process variable(s) according to the difference between
the measured value(s) and the reference value(s). The net effect of the
adjustment process is a change of some basic features such as response
time and disturbance sensitivity. The closed-loop control system can be a
feedback or a feedforward system.
2. An open-loop control system does not directly utilize the measurement
of the controlled variable. It consists of the same parts as the closed-loop
control system, but with no direct feedback (or feedforward) connection.
The simple process control system shown in Figure 1.21 is a feedback control
loop that affects the forward control element. The process part includes direct
and indirect controlled systems, while the process control part includes the
feedback elements, the summing junction and the forward control element.
The adjustment of the input variable by the forward control element produces
the controlled variables as an input to the direct controlled system. The direct
controlled variable occurs at the output of a direct controlled system. The
feedback element produces a feedback signal that is subtracted from the reference
variable value at the summing junction, thus producing an error control signal
which is fed into the forward control element to adjust the manipulated variable
to the input variable.
A direct controlled variable is fed to the indirect controlled system and,
together with the disturbance, produces an indirect controlled variable as the
a)
u= 0, y= (j u=(j, y=l u= 0, y=l
u=l, y=2
b)
STEP INPUT OUTPUT INTERNAL STATES
n Ul Yl II ql q2 q3 q4
0 '/1 (j 0 0 0 1
1 1 0 1
2 0 1 1
'"
'/1
'"
----------
''/"1
3 1 0 0 1
'" '/1
4 1 1 '0" ~
5 'I" 1 '0" 0 1 0
----
6 1 2 0 0 1 0
7 0 1 1 0
'" '"
r
c)
I 1 n
• 1 I n
~
'" 0 1 '"
Figure 1.20 A simple set of computer input, output and internal states.
The basic analytical concept of process control systems 39
PROCESS (controlled system)

direct
controlled
variable
disturbance disturbance indirect controlled
manipulated
variable
•
direct
controlled
system
input CONTROLLER
variable
t final
control
sensing
element
error
control
signal feedback
element
summing
junction
reference set-point
1--_11--
input element signal
Figure 1.21 A feedback process control system.
system output variable. The process is usually characterized by material, en-

ergy or information flow. The true process variables are therefore material
quantity, energy amount or information state, and are termed state variables.
Processes can be expressed analytically by deterministic equations and oc-
casionally by stochastic relations.
The development of sets of equations describing the process is known as
mathematical modelling. It is a specialized and skilled task as inspection of
any good text on the subject will readily show.
The first thing to remember when modelling is the aim of the exercise:
Why is the model being written? What do we want to learn from it? This is
important because it shows in which areas the model might be simplified so
that it can be written more quickly and use less computing time when run.
Building a model which is better than it needs to be for the given purpose
incurs more expense and usually takes longer.
The sets of equations defining the models vary greatly. Some may be as
simple as a single linear differential equation; others will be as complex as
sets of nonlinear partial differential equations. Most will lie somewhere in
between. However, it is worth noting that lumped systems are always modelled
by ordinary differential equations whilst distributed systems have to be modelled
by partial differential equations.
The computing effort required to solve a particular model is not straight-
forwardly related to the number and type of equations, although effort does
generally increase with size. The effort also depends on 'stiffness' which is
best described by saying that to obtain accuracy the computer is obliged to
use very small step lengths, thus requiring greater computational effort.
A good modeller is involved in making compromises between speed of
solution (very important if the model is to be used on-line for control) and
accuracy. For this reason modelling is often described as being both a science
and an art.
REFERENCES
Bulletin of Statistics on World Trade in Engineering Products (1989) United Nations.

Koch, G.R. and Hoffmann, R.H. (1978) Angew. Inf, 6, 248.
Siemens (1990) Course Material for Teleperm ME System, AS and OS, Karlsruhe.
Wolverton, R.W. (1974) IEEE Trans. Comput., C-23, 615.
2
The basic process unit
2.1 INTRODUCTION
A process control system comprises a process involved in a controlled system

and a process controller. Functions of the entire process control system are
described in this chapter. The basic process unit which integrates all essential
process and process control functions is identified as the basic building block
of the process control system, both for its vertical and horizontal integration.
The most critical parameter of the functioning of the basic process unit is its
data processing ability. Therefore the different functions of the basic process
unit controller are considered separately. The necessary data processing functions
of the controller are established and analysed as based on available data
processing time. The timing relations and informational features of process
input/output devices are also given because they limit the functional and
informational characteristics of the basic process unit and therefore the entire
process control system.
2.2 THE BASIC PROCESS UNIT
A verbal descriptive form of a control system includes all procedures, relations

and logics shown in a process. In order to systematize the description, Z.
Kehler (private communication, 1975) proposed the following postulates:
Postulate 1 The basic hardware elements of the production process and plant
are called the basic process units of the entire process plant. A basic process
unit generally consists of:
large mechanical process hardware that includes all process devices, sensors,
transducers and signal converters;
the controller for the basic process unit that includes all the hardware and
software, where all the procedures, relations and logics, as well as a
connection with the superimposed control system(s), are executed.
41
42 The basic process unit
Postulate 2 A basic process unit is responsible for at least one specific process
function. Any part of the unit is based, designed, manufactured, mounted and
put into action only if it completely fulfils its part in the functioning of the
whole unit.
Postulate 3 Common process functions of the whole process or plant are

executed in and by means of a process controller that unifies the hardware
and software into the following functions:
mutual functional interconnections and coordination functions between basic
process units;
acquisition of common process signals, execution of common shutdown
commands and issuing of permissions for the start-up of the whole process;
presentation of common functions to process operators.
The interaction of the basic process unit, process controller and superimposed
control system is shown in Figure 2.1.
Example 1
The functions of a crude-oillnatural-gas separator can be split up into the
functions of the controller for the basic process unit, the process controller for
the whole measurement station, the synoptics of the basic process unit and the
special device function as shown in Figure 2.2. The basic process unit is
responsible for three main functions.
1. Processing process signals A, C, D and E and command B, where
flow transmitter data (A) are transmitted directly to process controller
level;
level transmitter data (C and D) are processed for the synoptics of the
basic process unit with the logic functions
Cl = C . (Q21 + 1 Hz) (2.1)
Dl = D . (Q22 + 1 Hz)
where Q21 and Q22 are set/reset sequential functions generated at the
process controller and 1 Hz is a 1 hertz signal generated at the basic
process unit controller for the alarm warning purposes of the local
synoptics;
data from flow pulse transmitter (E) are preprocessed at the level of
special device functions where they are converted into pulse-cumulative
and flow-rate data;
the command (B) is issued according to the logic relation
B=A·L·J1+F (2.2)
where L and J1 are set/reset sequential functions generated at the
basic process unit controller and F is a logic variable generated at the
r-----------------------------------------------------------------,
I I
I I
I I
I superimposed control system I
I I
I I
:L ________________________________________________________________ ...JI
process controller
I
I
controller] -1 controller] controller] +1
basic process basic process basic process

unit] -1 unit] unit]+1
energy, material or information flow

Figure 2.1 The interaction of the basic process unit, process controller and
superimposed control system.
special device level
connection
IG
process 2~ " 21~
controller 00
"
en 00
-eN
functions ......
~eN
~
'"
eN
'-
~
00
eN
II U .....
eN
.....
'" 00
" 001
'" '"
",,,, 0
00 r-
eN eN
~~
u,~
II
~F 00 h
connection
synoptics
of the
basic
tlrocess
unit
connection
functions II
.......
of the
...... + + + ......
~
basic
+ .......
'"'"01 :.:: r-
+2::: ~~ '"
"
process ...... eN ~
01 ~- ~
~
unit « :3' « 0
Uen'
+ -~ :i~
00
-0
~
,
U
0......
............- ......-
eN
c/+
......
« ~. " .....
U
.....
0 "
'-::t::
II
::::: .....
II II
,,2-
...... II
(f ! t !
II
:If ~:'::oo ~oo -,00
connection
process data and < ~ U 0 t.t,;;

commands FT VS .LT LT FPT
Figure 2.2 The functions of the basic process unit of a crude-oillnatural-gas

separator: FT, flow transmitter (signal); LT, level transmitter (signal); VS, valve
shutdown (command); FPT, flow pulse transmitter (signal); 0, panel signalling
lamp: 0, panel pushbutton; D, panel symbol for the flow pulse transmitter.
The basic process unit 45
synoptics panel when pushbutton 'MAN' (manual operation) is de-

pressed.
2. Generation of set/reset sequential functions where
K function is set with a signal HI from pushbutton SO (shutdown)
and reset with a signal II for the 'release of emergency reset' from
pushbutton AD;
L function is set with the signal from pushbutton AD and reset with
a signal G I for signalization of the shutdown condition.
3. Generation of signalling functions J I and G I for the synoptics of the
basic process unit where
J1 function is set with the logic function
S = (C + D) . L27 (2.3)
where L27 is an 'alarm acceptance signal' generated at the synoptics
of the process controller and reset with a signal II from pushbutton
AD;
G I function is given with the logic function
GI = C + D + K + L27 (2.4)
Figure 2.3 shows the main data processing and synoptic system blocks. The
basic process unit logic controller is connected to the process, the synoptics
unit of the basic process unit controller, the process controller for the entire
plant and to special devices such as set-point regulators, counters, integrators,
etc. The process controller for the entire plant performs the following functions
for the basic process unit of the crude-oillnatural-gas separator:
receiving process signals A2, C2, and 02;
issuing set/reset sequential functions Q21 and Q22 in order to control the
behaviour of signalling lamps via the basic process unit controller;
receiving signals L27 and L28 from other basic process units for overlapping
actions.
Example 2
The functions of a basic process unit for a street traffic control system are
shown in Figure 2.4; the synoptics of the basic process unit and corresponding
process controller functions are also given. The three main functions of the
basic process unit are:
I. Processing process signal A and commands B, C and 0
the data from traffic pulse transmitter A are passed directly to a special
device level where they are converted into traffic flow cumulative and
traffic rate data;
,• synoptics of
the process
controller
~
process
i
,
controller of
the whole plant
from/to other
I j from/to other
basic process basic process
units units
,• synoptics of
the basic
process unit
1 1
II
•
II
synoptics of
the special devices
~ i
logic controller of
the basic process
- special devices
(set-point regulators,
unit counters, integrators)
1 t ~ i
basic process unit:
hardware, process devices, transmitters
Figure 2.3 The structural scheme of main data processing and synoptic system
blocks.
special device level
connection
process
controller
functions
connection
yellow flash
synoptics
of the ··808··.
EJ o~ ..~ BEl 0
basic
process
unit
connection ~
functions ..... -<J .....

c-s- +
"";Iii:
of the <J
+<J<J
~
Ij
.:::,:+
U1...:I
basic +
... _+ ~ ._
: + .... ~ ...
-....
c:l ~
process .....
c-s
C\I_ ...
c-s ..... - ~ II II <1 II .... en en ;;~~
unit
<11- ....;15 "'-
~ ~~==:
~
1I~;::: c-s
..; v ~ .....
-
II ::: II II II
..... c-s II II U.lii: ..... u. ..... C'-Ic<")
..... c-s .......... - N .....
~
c-s II c<")c<") .....
z II II 1I1u. .... II
; ;;
~ ~~ ~~z
~ <J <I<J<J r.:Q:::E uc:lI:::EZ .... ::E <J<J
connection
process
1
<t;
TPT
r.:QUc:l
TLO TLOTLO
data and
commands f13 '" (E21 !:, E21) MFI Gl + (E31 d E31) LFI + (Kl d Kl) M +
(E32!:, E32) LFI + (11 Eli) M + J1 1 Hz F1 + (E22 L'1E22 ) MFI Gl
Figure 2.4 Basic process unit functions for a street traffic control system: TPT,
traffic pulse transmitters (signal); TLO, traffic light on (command); D, panel
pushbutton; 0, panel signalling lamp; d , panel symbol for the traffic pulse
transmitter.
- commands for traffic lights B, C and D are generated according to the

following logic functions:
B = F1 . ~ K1 + LM + M . F1 . G1, for green light signalization

C = E21 . ~ E21· M . F1 . G1 + E31 . ~ E31 . L· F1 . K1 . ~ K1
. M + E22· ~ E22· M· Fl· G1 + E32· ~ E32· L· F1
+ II . ~ II . M + 11 . 1Hz· F1, for yellow light signalization
and
D = F1 . ~ II + [ . F1 + M . F1 . Gl. (2.5)
for red light signalization where L and M are set/reset sequential
functions, F1, Gi, HI, II, J1 and K1 are signals generated at each
pushbutton pressed at the synoptics of the basic process unit, E31 and
E32 are signals from the special device level when the change to
green and red respectively is to take place (at 'local mode' of work),
E21 and E22 are signals from the process controller when the change
to green and red respectively is to take place (at 'remote mode' of
work); logic variables designated with ~1 and ~2 are time-delayed
signals used for changing the traffic lights from green to red and from
red to green light; 1 Hz signal is used as a flashing signal of a yellow
light.
2. Generation of set/reset sequential functions Land M where
- L function is set with a signal
s = ~ E31 . G1 (2.6)
and reset with a signal
R= ~ E32· G1 (2.7)
- M function is set with a signal
s = ~ E21 . HI (2.8)
and reset with a signal
R = ~ E22· HI (2.9)
3. Generation of time-delayed signals ~ E21, ~ E22, ~ E31 and ~ E32 from
signals E21, E22, E31 and E32 according to relations
~ E21 = E21 (t + ~1)
~ E22 = E22 (t + ~2)
~ E31 = E31 (t + M) (2.10)
~ E32 = E32 (t + ~2)
where ~1 and ~2 are given time intervals between turning on the green
and red lights (~l) and between turning on the red and green lights (~2)
Basic process unit data processing 49
generation of signalling functions Bl, Cl, Nll, N12 and N13 for the
signalization of traffic lights and working modes of the basic process
unit.
The process controller for the entire system performs the following functions
for the basic process unit of the traffic control system:
receiving process signal A and command states B2, C2 and D2;
issuing signals E21 and E22 from signals L21 and L22 received from
other basic process units.
The split of functions given in the above examples enables consistent process
control of process parts, signalization of states to process operators and
meaningful integration of these basic process units into the entire process.
2.3 BASIC PROCESS UNIT DATA PROCESSING

As already described, each basic process unit controller performs control and
logic functions. The postulates are as follows.
Postulate 4 The speed of data processing VDB of each basic process unit
controller is faster than the speed of data processing of the superimposed
process controller VDC concerning data processing required for this particular
basic process unit, that is
(2.11)
expressed as the number of logic/arithmetic operations in a time unit.
Postulate 5 To satisfy a particular production function, each basic process

unit controller communicates with the superimposed process controller by
some chosen, summary and global control and supervisory signals, commands
and set-point values. Thus, designating the number of process signals and
commands at the level of the ith basic process unit controller as nBCl and the
number of process signals and commands communicated to the ith basic process
unit from the process controller as npCi' the following relation applies
nBCi > npCi for each i = 1, 2, 3, ... =m (2.12)
where m is the total number of basic process units in the process.
Postulate 6 The quality of all functions of the process depends solely on the
quality of the functions performed at the basic process unit. The quality of
functions performed at the basic process unit depends on the data processing
ability of the control unit with other parts of the basic process unit being
equal.
Of the many parameters characterized in the basic process unit controller,
the following appear to be the most important:
Table 2.1 Response time for some types of process transmitters
Type of transmitter Response time or equivalent Type of output signal

freuency response
Pressure transmitter* 5 Hz fiat response Analog 4-20 rnA

(electronic)
Flow transmittert 3 s - 10 s response time Analog 4-10 rnA
Temperature transmitter:j: 4 s - 40 s, 63% response time Mechanical recorder
Level transmitter~ f s - 100 s, 63% response time Analog
* Fischer & Porter, Type 50 EP/EN/EQ

t Fischer & Porter, Oscillating Ball Flowmeter
:j: Fischer & Porter, Liquid Filled System
~ Leeds and Northrup, Differential Pressure, model 1911(118)
response time to process variable change;

accuracy of process variable processing;
process safety and primary process protection;
process stability
The response time to process variable T R can be defined for each process
variable, and principally includes the following time requirements for data
processing at the basic process unit level:
TR = Ts + Tc + Tp + TD + TpR (2.13)
T s is the time interval from the change in process variable to the response in
the sensing device (e.g. transmitter), which includes the complete conversion
into an electrical or optical signal at the transmitter output. Table 2.1 outlines
some response times for different process transmitters. T c is conversion time,
i.e. the time interval from the appearance of an electrical or optical signal at
the input of the control unit of the basic process unit to the conversion of this
signal into the appropriate digital code of the control unit. For analog signals,
this equals the filter time constant, Tf , multiplexer scanning time, T M' and analog-
to-digital conversion time, TAD' i.e.
(2.14)
Filtering time is the time interval from the filter input signal to the output
signal. Different filtering time constants depend on signal-to-noise ratio of the
process variables and usually assume some of the following values: 1 ms,
2 ms, 10 ms, 20 ms, 50 ms and 100 ms. Some commonly used filters and their
transfer functions are shown in Figure 2.5.
Multiplexer scanning time is equal to the time interval needed for input of
a process signal through some multiplexer circuits to the data processing unit.
Response Transfer Terminal

Type of filter Electrical equivalent scheme time function condition
Common passive
filter
2 ~= 1 ZI = 0
Passive T type 't = RC VII + 2 RCs Z2 = 0
RC filter
V2 = see (*) ZI = 0
Active RC filter
Vl Z2 = 00
Figure 2.5 Some commonly used filters and their transfer functions: A, amplifier
gain factor.
Scanning time differs for various types of data processing systems, but usu-
ally there is a difference in analog and digital input scanning, where digital
signals are scanned mostly upon a change of signal state, whereas analog
signals are scanned at regular time intervals. Thus, scanning time for digital
signals can range from 10 /-lS to 10 ms while scanning time for analog signals
is usually 100 /-ls to 1 s, depending on the number of inputs scanned.
Analog-to-digital conversion time is the time interval needed for the
conversion of the analog value at the scanner input into the digital value at
the analog-to-digital circuit output. Table 2.2 shows the analog-to-digital
conversion time for different types of device, together with their important
parameters.
The data processing time T p is the time needed for the conversion of data
from their raw form (i.e. as they appear in the system after filtering, scanning
and digitization) to the appearance of a digital signal after data processing, de-
pending on each process variable. Data processing algorithms, and their time
responses, are discussed in Chapters 4 and 5. T p can range from 1 ms to 100 s.
The response time T D of a process device after receiving an output com-
mand from the control unit usually consists of:
reception time of the output command. When the output command is given
in a coded form, it also includes the conversion of this code;
Table 2.2 The parameters of analog-to-digital converters
AID converter type on Accuracy bit Conversion speed Integral linearity (%) Differential linearity Conversion error bit
principle of (J1Slbit) (%, LSB)
Successive approximation 8-14 0.l-15 < 0.1% ± 1 LSB ±t LSB - ± LSB

One slope integrator 8-12 0.3-10 < 0.01% < 0.1% ±t LSB - ± LSB
Dual slope integrator 8-12 0.1-3 < 0.01% < 0.1% ±t LSB - ± LSB
LSB - least significant bit value

Table 2.3 Response and settling time for different process devices
Type of process device Response and settling time
Electrohydraulic linear actuator* 1-2 Hz, a 3 dB upper limit frequency

response
Flow rate controllert Total response and settling time 1-2 s
(Maximum differential pressure 84 bars)
Switching power transistor Total response and settling time 0.3-1.2 Ils
(switching power 100 W - 10 kW)
* General Controls ITT, Type AH90

t KATES - Automatic flow rate controller
conversion of the received command into the powered signal to the proc-
ess device;
response of a process device to the powered signal.
Table 2.3 gives the characteristic response times for different process devices.
The process response time TPR is the time interval that characterizes particular
processes. For most processes, T PR lies within a wide range from 1 ms to
1000 s and usually represents the longest time constant of the whole basic
process unit.
To enable the basic process unit to function, the control unit with its synoptics
and special devices receives the process data, processes it according to given
algorithms, issues commands and set-point values to the process devices and
communicates with the superimposed control level. While the control unit is
performing these tasks, the process parameters change and, due to the serial
nature of logic and arithmetic functions of the control unit, there is always a
time delay between the real and computed process variables of the basic process
unit. To preserve the accuracy of a process variable, the speed of the data
scanning should be given. For a process variable y(t) scanned by the control
system at regular time instants Tj, i = 1, 2, 3, ... , given the accuracy e' of the
observation of a variable y(t) and a maximum speed of change cm = [dy(t)/
dt]max of process variable y(t), the following relationship exists
i=I,2,3, ... (2.15)
where k is a data acquisition factor, usually between 3 and 5, that depends on
the number of time instants taken in variable calculation. For input variable
with cm = 2 Vis, e' = 0.02 V, and k = 3, a data acquisition interval tj+! - tj
equal to or less than 3.33 ms is recommendable.
Digital signals do not operate like analog signals, and are usually of a more
spontaneous nature with their acquisition time being much shorter. A higher
speed of data processing is also required for a digital signal in the primary
protection and process safety of a basic process unit. The count pulse data
differ in their rate, usually between 10 per second and 1 per minute, and are
processed by the control unit of a basic process unit. Pulse signals of a higher
rate, and pulse signals of a special process, are usually preprocessed at special
device level and then fed back to the control unit (see Figure 2.4).
Control functions are performed according to a required process control
algorithm and the precision of their functioning is essential for the overall
response time, stability and process control accuracy of the basic process unit.
A real process control system is described as a linear time-invariant discrete-
time dynamic system with concentrated parameters
X(t k+ 1) = Ax(tk) + Bu(tk)
y(t k) = Cx(tk) + Du(t k) (2.16)
x(to) = Xo
and is theoretically controllable only if its mathematical model (2.16) has the
following five features (Novakovic, 1981).
1. Complete controllability of states and inputs.
2. Complete observability.
3. The same number of control variables and state variables.
4. Minimum sensitivity to parameter and structure change, and to variable
disturbance change.
5. Maximum sensitivity to control variable change.
A complete controllability is usually not fulfilled, i.e. the dimension of the
measurable state vector A(tk) is 1 < n where n is the dimension of the state
vector x(tk). The state vector x(td can be reconstructed or approximated from
the observed variable if a functional F' exists
xlt k) = F'[y(t k'), to < t k' < td, tk ~ to (2.17)
so that xltk) =:: x(tk) where xr(tk) represents a reconstructed state.
Definition 1 The n-dimensional system

X(tk+l) = A(t k) x(td + B(tk) y(td + C(td u(td (2.18)
is a full order observer for the n-dimensional system
X(t k+ 1) = A(tk) x(tk) + B(td u(t k) (2.19)
y(t k) = C(tk) x(tk) (2.20)
if
(2.21)
implies
x(td = X(tk) tk ~ to (2.22)
for all u(td, tk ~ to
Theorem 1 (Kwakemaak and Silvan, 1972) Equation (2.18) is an observer for

equation (2.19) only if
A(tk) = A(t k) - K(tk) C(t k)
B(t k) = B(t k) (2.23)
C(tk) = K(tk)
K(tk) is an arbitrary time-varying matrix. As a result, the full order observers
have the following structure
X(t k+1) = A(tk) x(t k) + B(tk) U(tk) (2.24)
+ K(t k) [y(t k) - C(t k) x(t k)]
When applying the control laws in the form
U(tk) = -F(tk) x(t k) (2.25)
and substituting equations (2.25) in (2.24), the controller takes the structure
given in Figure 2.6. A simplified structure of the controller is shown in Figure
2.7.
For the discretized continuous process with
X(tk+1) = a. x(tk) + f3 u(td (2.26)
(2.27)
consisting of a series connection of second order elements and an integrator
with
a. = (:~: :~~ ~], f3 = (~~:] and C = (001) (2.28)

a31 a32 a33 b31
the observer has the following structure
X(t k+1) = a. x(t k ) + f3 u(t k) + K [y(t k) - C x(t k)] (2.29)
and the control laws are given by
u(t k) = -F X(tk) + rk (2.30)
The choice of K is done by subtracting equations (2.26) and (2.29)
x(tk+l) - X(tk+l) = (a. - KC) [X(tk) - X(tk)] (2.31)
r- - - - - - - - - - - - - - - ; - - - - - - - - - - - - - - - - - - - - - - ,
).l (t k ) .:(tJ\ + 1) Lla,l).t :
I
I
Figure 2.6 The structure of the full order observer .1t =t k+1 - tk = constant.
The dynamics of the error given in equation (2.31) can be modified by the
appropriate choice of K. Usually the observer has to be faster than the plant
so that the results of the states may be supplied quickly to the process. This
is especially important when some external disturbances exist, or the plant's
initial conditions are not known. At other times, the observer may smooth the
measurements contaminated by noise thus being slower than the plant.
The proper K may be solved by using the values of the poles of closed-loop
dynamics described by the equation (Kwakernaak and Silvan, 1972)
(2.32)
The desired poles of equation (2.32) are three eigenvalues ZJ, Z2 and Z3 found
from the equation
det(zI - a - ~C) =0 (2.33)
PROCESS
----------------------------- ~
r--------------------------------------------
state estimator
_________________________________________ J
..:~~~l~n! ~~~ _______________________ -.J

Figure 2.7 The simplified structure of the controller.
The dynamics of the controller in equation (2.31) is

X(t k +1) = «l - KC) x(t k ) (2.34)
The solution of equation (2.34) for K with modified eigenvalues Zlm' Z2m and
Z3m from equation (2.33), due to known observer error dynamics and with
given values of (l, is found from
det(zI - (l + KC) =0 (2.35)
or
(2.36)
The block diagram in Figure 2.8 for the case in equation (2.28) with known
K requires 21 multiplications and 19 additions during each sampling interval
in the controller.
Controllers can be designed for optimal control of the process from three
main purposes.
zero order x,y

x = Ax + Bu
hold
- 6
----------------------------~------- ..... --------- D
!
sample and
hold element
a = exp All controller synchro

13 = J~ a13 dt
Xk + 1 = aXk + 13Uk + K(Yk - Yk)

Yk = CXk
~6~__ r
reference value
~--------------------------~rk
Figure 2.8 Complete block scheme of the computer,controlled continuous process:
C, equation (2.28); K, equations (2.33) and (2.35).
1. Relocation of poles of system response to more favourable positions to

obtain a control law
(2.37)
that minimizes the cost function J for any initial condition (Katz, 1981)
J ~ 1/2 (% xi Aw Xj + uJ Bw Uj J (2.38)
where N is the number of steps and Aw and Bw are constant positive-definite

symmetric weighting matrices of appropriate dimensions. For N increasing
to infinity, a steady state is reached and FK = F = constant and a con-
troller is named a regulator. Various methods are used to minimize equation
(2.38) (Gorez, 1981).
2. Proper response to the input command in the model following method of
control design; input commands are applied to the process and model and
the control law is determined so that the process output is similar to the
model output.
3. Proper response to large disturbances (e.g. system start-up or shutdown)
by proper design of system control through system eigenvalue synthesis
(Katz, 1981).
Each control system when observed as a data processing system follows the
simple and approximate calculation of data processing that can be expressed
as available data processing time for a single processor in its basic process
unit. The available data processing time can be estimated differently for vari-
ous controller functions. When all processing functions have the same priority,
the simple relation for total available time T TOT can be expressed as
N Mi
LL zi.j . 't i.j ::; TTOT (2.39)

i=l j=l
where Zi,j is an input, output or data processing variable to be processed in
the time interval T TOT and 't i,j is the data processing time required for the
processing of the corresponding variable Zi,j in the time interval TTOT; N is
the number of groups of data processing variables and Mi is the maximum
number of data processing variables in the ith group. Figure 2.9 illustrates the
total processing time for this case. When more processors are involved in the
data processing of the basic process unit, equation (2.39) can be applied for
each processor if the data processing requires the same priority. Total avail-
able time T TOT can be estimated from the critical process control functions of
a system. A primary protection and process safety time response is usually
more critical than a control algorithm. The data processing function must then
fulfil the most critical function inside its total available time T TOT'
A more precise calculation of data processing parameters of a basic process
unit requires the connection of the data processing algorithm with the sup-
posed hardware configuration (see Figure 2.10). The execution of any hard-
ware data processing function is negligible when a calculation is made ac-
cording to equation (2.39).
Analog data inputs are scanned at time intervals ti+l - ti according to
equation (2.15). The values of analog inputs are converted into digital values.
Input wordlength C w of each signal is given with the expression (Ahmed,
1976)
(2.40)
where
XSAT and XTH are the saturation and threshold levels of the analog-to-digital
converter; usually XSAT = 10 Y, XTH = 5 mY;
Fq is the quantization noise value in dB; usually 60-80 dB;
82 is the variance of the input samples to the analog-to-digital converter,
usually around 10-3 y2; with these data C w is between II and 22 bits.
Digital inputs are scanned in groups in case a change occurs in one of the
input channels. Each group of input channels consists of 4, 8, 16 or 32 inputs.
Digital signals (being a stochastic variable) appear at the rate of one per hour
analog c=:::=v> analog
inputs outputs
data flow path
digital
inputs
input output
scanner control and scanner
and data processing and : digital
count pulse
and process o outputs
inputs data data
algorit~ms
adaptor adaptor
communication communication
I
inputs I outputs
I
I
I
total available data processing tir'ne -l

~ r
--)
_________________
analog input ....processing
________ ..... ' / I ......
~'-----
-"-...-.. ....._on/off command processing
o
°t l O t 0 / I '-... -"-...-.. processingoTseti>Oint-----
d Igl a mpu processmg /
_____________________________J /1 ................... ..... ................ -..-. .................... '-'-.. , values
_______________ _
I ' ...... ...... ......
count pulse input processing /
___________________
1-'-...-..
I..............
-""" ...
-"-'-...-. ..... """-... ... communication output processing
- -- - - - - - - - - - - - - - - -
o 0 0 0 / -"-"-. "-"-. '-...-.. process safety and protection
_______ ~~~~~~l~~!~O_n_~ny_~t_~~<:':.~s~~ngj -"-"-"-.. -"'-""-' -'-"'-...Jl!'~~~~ ___ _
............... ..... .....
-"'-""-.. -"-'-"-.. process control algorithm processing
................... ---------------------------"-
"-"-.. reserved data processing time
'------------------~--------
Figure 2.9 Total available data processing timeo

data flow path
hardware-type II I
I
software/ data processing I I hardware-type
I I
data processing : I
I
I I data processing
analog input I
I
1
: data acquisition algorithm
I
I
: analog output
I II.)
module 1 II.) u : module
digital input 1 u process control algorithm JS.... I
I JS I
I .... II.) ~ ....
II.)
module I II.) ....
~ C<l
digital output
I ....., C<l process safety procedure module
count pulse I .S ~
.~
"04-<
~
input module I "04-<

I ~ ~
~ 0
C<l on
communication II 0011.)
C<l 1 process protection plan 001
~ II.)
binary output
input I
I ~ ~ ':;;J ~ module
hardware I
I 'B .6
....
auxiliary data processing II.)
~"O
~
I
'2~ ; communication
I I
binary input I o..C<l
~
I
module
I ';:: ...c
local display protocols I "5.-C : output module
I "5
I
operator
1 S I
I S
I I II.)
I
I "0
keyboard I
I
communication protocols I : local display
module iI
I
I t module I
i I
I
I
I I
system bus
supervision
Figure 2.10 The hardware configuration and the calculation of data processing
parameters.
to one per second, depending on the type of process controlled. The probabil-
ity of data acquisition upon request by the ith group of n digital input chan-
nels equals
'A'j = n'A' + (n - 1)'A'2 + ... + 2'A'n-l + 'A'n (2.41)
where 'A' is probability of the occurrence of an event in any of n input chan-
nels, all probabilities are equal. For a small amount of probability, equation
(2.41) can be approximated with the first two terms.
Each event is tested for plausibility as there is a risk of errors due to noise
and cross-interference.
Count pulses are acquired at regular intervals because they must be checked
for the accumulated pulse number. When count pulses are directly connected
to input hardware, a provision is made to enable the collection of all count
pulses. This is usually the case for lower and regular count pulse rates and
mandatory for point of sale sites.
Communication inputs are processed with a speed equal to the maximum

bit rate on communication line when one processor concept is accepted (see
Figure 2.9). The communication hardware usually performs the necessary
additional functions (e.g. synchronization, coding and line protocols) so that
the processing of communication data for the control system only demands
the necessary processing power needed for the data messages.
Analog signals are filtered. For example, when recursive filtering is ap-
plied, then the impulse response of the filter can be given following the relation
for
I (n T) = 0.5 n n = 1, 2, 3, ... (2.42)
where T is the sampling interval.
The z transformation of the transfer function of the recursive filter is ex-
pressed as (Bogner and Constantinides, 1975)
Hr(z) = Yr(z)/Xr(z) = z/(z - 5); z = esT (2.43)
Thus an analog value At from the time t is taken as a time series (AJ2,
AJ4, AJ8, ... , AJn) decaying in geometrical progression and added to analog
values from other points in time.
2.4 PROCESS HARDWARE FOR DATA INPUT
Standard process equipment includes transducers, sensors and signal convert-

ers for input of process data into a computer control system. According to the
type of electric or optic output signal from these input devices, a status de-
vice, pulse-count device or analog signal device may be used. These devices
are installed at the process site and their signal adapters or amplifiers mounted
at the control point, or anywhere between the process and control point. The
functions of input devices are (1) power supply of transducers and transmit-
ters, (2) converting the process input value, (3) testing the process input value
(state), (4) forming the limit signal value, (5) calculating the proper process-
ing variable, (6) process signal distribution, (7) reporting wire breakdowns
and (8) simulating the process value. The basic features of the process input
devices are specified by their manufacturers as power supply data, device
input data, device output data, construction and mounting data, cabling and
safety data.
2.4.1 MEASURING TRANSDUCER
A measuring transducer is the process input device that accepts process infor-
mation in the form of a physical quantity as its input variable and, according
to a definite law, converts it to information in the form of the same or another
physical quantity.
Process hardware for data input 63
There are many different types of measuring transducer, depending on the

conversion sensitivity to specific physical phenomenon, such as:
displacement, angle, velocity, rotation speed, acceleration measuring
transducers;
- flow measuring transducers or flowmeters;
measuring transducers for electrical quantities;
temperature measuring transducers;
force and torque measuring transducers;
- pressure and differential pressure measuring transducers;
deformation measuring transducers;
- level measuring transducers;
analytical measuring elements for chemical and physical measurements.
2.4.2 SENSOR OR DETECTING DEVICE
A sensor or detecting device (transmitter) is the part of a measuring trans-

ducer that converts the input process signal into a form suitable for measure-
ment.
The relationship between the input and output variables of a sensor is es-
sential and cannot be altered externally other than physically or functionally
modifying the device.
A general relation for continuous-type, one-signal parameter sensors can be
given as
n n
i=l i=l
where f(v) is the dependence of output physical quantity f on the quantity v
to be measured and Vb v2, ... , Vi are measurement perturbance factors that
modify the output f(v) in an undesirable way. It is necessary for a linear time-
invariant response to possess a sensor where the second, third and all other
terms on the right-hand side of the equation (2.44) are negligible compared
with the first term, e.g. less than 0.1 %.
2.4.3 SIGNAL CONVERTER
A signal converter is an input device that changes a standardized signal,

the physical nature of the output being the same as the input. A standard
process measuring model is shown in Figure 2.11. The output of a measuring
transducer can, for instance, be a standard pneumatic signal 20 684.28 -
103421.4 Pa (corresponding to previously standard 3-15 p.s.i.g.) and the out-
put of a signal converter can be a standard electric direct current signal of 4-
20 rnA.
v(%)
100 T ___ _'"*'1

...- ....
disturbance v I / "
xlOO / "
/, ,,"\
/
\ disturbance
"" '- -- v2x100
o I~ I~ f(%)
o 100
process
physical
quantity
signal physical
L quantity
v sensor sensor output standardized input

input variable (f) physical signal
variable suitable for quantity
measurement
Figure 2.11 A standard process measuring model.

The process signal (input signal)

The process signal is obtained by a transducer or measuring device, and is a
measurable process variable whose one or more parameters carry out instruc-
tions on one or more process variables represented by the signal. These are
termed information parameters of the signal. For example, for an ampli-
tude-modulated sinusoidal carrier the instantaneous amplitude is the information
parameter of the signal. Information parameters of intelligent transmitters are
described in Chapter 3.
2.4.4 STANDARD PROCESS INPUT DEVICES
Standard process input devices are categorized according to the types of sig-
nal they issue. These are set out below.
1. Binary device. A device that issues a process signal with a binary quantized
information parameter conventionally denoted as 0 and 1 (binary signal).
2. Digital device. A device that issues a process signal with a quantized
information parameter represented numerically (digital signal).
3. Analog device. A device that issues a process signal with the information
parameter which may assume all the values of the given range (analog
signal).
4. Sampled device. A device that issues a process signal with the informa-
tion parameter which is only intermittently observed and represented
(sampled signal).
The description of standard process input devices and their signals are given
in Table 2.4.
Process input devices are usually installed on site where the process vari-
ables can be detected. Input signals are transmitted from the process into the
control room, or cubicle. For larger distances or where only the sensors are
mounted on site, the amplification of input signals is performed between proc-
ess and control points. Typical solutions for the connection of various input
devices are outlined in Figure 2.12. Input devices usually fulfil the following
functions.
1. Power supply of transducers and signal converters. The supply of power
to transducers is, for the most part, organized separately for each meas-
uring variable through the corresponding devices; transducers are con-
nected by two, three or four wires to the power supply device, thus enabling
individual voltage supply; through power supply a live zero processing of
current signals is possible ('0' = 4 rnA).
2. Testing of process input value (state). The voltage or current of power
supply for transducers is controlled; the limit value of input signal
is controlled against low (or high) preset value; thus the breaking of
Table 2.4 Standard process input devices and their signals
Input device Input variable Information parameter Input signal Common connection characteristics
Binary device Any process variable Quantized information Binary signal Closing contacts
parameter (0, 1) Insulated from any ground circuit or
conductive parts
Volt-free
~ 10 7 operations
Digital device Any process variable Quantized information Digital signal BCD accumulator up to six digits
parameter (digital Insulated from any conductive part
number)
Volt-free
Normally open contacts
Contacts with breaking capacity 10 VA
Minimum breaking current 0.5 AI
48 V DC
Contact resistance:;; 100 mn, ~ 10 7
operations
Analog device Any process variable Any value of a given Analog signal Floating loop (~ 100 mn; ~ 0.5 kV)
range with no common point
Standard DC current 4-20 rnA,
0-20 rnA or voltage 0-10 V
Controller load not exceeding 500 n
Sampled device Any process variable Intermittently observed Sampled signal Closing contacts
information parameter Duration of intermittent signal ~ 50 ms
(analog or digital)
Volt-free (for digital signal)
Floating loop (for analog signal)
process variable instrument adaptor transmission control place
resistance
temperature
ftJh~e~r:m~o~m~e~t~e~r}-~=========:________________
L
-1 A
strain
~ strain ~
. 1L.._g_a_ug_e_-,~----------------<0
flow ~_ flowmeter
electro-
magnetic
_
H3> A ---------------------------------o
_ _ _ _ _ _ ~~ Jdifferentiall
displacement ,----~~--~ - ~transforme~~----------------------------~O
pressure
.~
-1 transmitter
pressure r---- o
direct current
temperature
-~thermocouPle~-------_____i~
1 _ DC/DC L-,
~Iconverterl
Figure 2.12 Typical connection of various input devices:~,

amplifier; - - , DC connection;~, input variable.
~ z=Va.~
c z =Vt, PlJ
Flow
o -lOY
Figure 2.13 The complex calculation of flow: ~, transducer or signal converter;

EJ '
calculation unit; t1p, differential pressure; p, pressure; v, temperature;
I, DC current; u, DC voltage; z, calculation law of the flow.
conductors from power supply to transducer and short circuit of conduc-

tors and functioning of transducers are controlled.
3. Formation of limit signal value. For devices with an input signal limit
value (e.g. analog devices) two independent limit values are usually set,
denoted as MAX (IMUM) and MIN (IMUM) value; these values are set
in the range of 3% and 100%, with hysteresis ranging from 0.5 to 10%;
limit values are usually provided with hardware limit circuits, automati-
cally put into logic zero each time transducer breakdown occurs.
4. Calculation of proper process variable. Such data processing is, for in-
stance, effected in the following cases:
the calculation of alarm binary value out of two or more analog sig-
nals;
the calculation of flow or mass-flow out of pressure, differential pres-
sure and temperature signals (see Figure 2.13);
level measurement with pressure signal;
sum or mean value calculation of more measurement signals.
5. Process signal distribution. The distribution of every analog signal is usually
done with a chosen signal current level such as 0-20 rnA or 4-20 rnA which
enables the following to take place:
short-circuit protection of the output signal;
voltage drop between power supply device and transducers need not
be considered;
from transducer indicators

and
.~ recorders
'LgJ-
for other
functions
to limit circuits
Figure 2.14 The plant distribution of process signals circuits: 1LPF I. low pass filter;
~. voltage to current converter.
compensation from the current supply is made for slow varying volt-
age disturbances induced from other conductors or transducers;
the distribution of signals is performed by independent circuits (see
Figure 2.14) which enable the independent and isolated use of one
input signal for more plant functions.
6. Breakdowns report and simulation of process value includes the following:
testing and indication of input signal value and internal power supply
voltage;
lock-in of regulators and limit circuits when disturbance occurs, or
when simulation of process variables takes place; simple transducers
and signal converters (as temperature signal converter, for example)
possess connector pins for the control of input signal, for the calibra-
tion and for the simulation of process input variable; the connection
of test instruments to complex devices, as for example signal con-
verter and distributor, enables the measurement of input signal, simu-
lation of the input signal, simulation of process variable, simulation of
signal for limit circuits, measurement of the state of limit circuit el-
ements and the simulation of binary signals from transducers.
Electric system data of input devices
These include the following features of transducers:
1. Rated voltage of 24 V DC for power supply Us and for control test
voltage Vc.
2. Allowed voltage span for Us and U c from 19.5 V to 30 V.

3. Analog input variable as a DC signal from 0 or 4 rnA to 20 rnA with a
100 Q load.
4. Analog input variable as a DC voltage signal from 0 V to 2 V, or from
o V to 10 V.
5. Binary input signal as potential free contacts, 24 V, 48 V, 110 V DC, 6
rnA.
6. Analog output signal from the device as a DC voltage signal from 0 V to
2 V or from 0 V to 10 V.
Construction and mounting data of input devices

These include the following:
mounting data of transducers or sensors on process hardware elements
such as dimensions of holes for mounting parts, type of mounting (screw,
welding), recommended position of sensor on pipes or reservoirs;
cabling data from mounting place to the control room or cubicle including
the type of cable (twisted pair cable, coaxial cable), type of cable protec-
tion/coating, the way of cable lay-down, cable shelf, distance require-
ments between power cable and data processing cables:
construction data of transducer housing, signal converter housing and data
processing cubicle such as dimension of housings and cubicles isolation
type, the lay-out of electronic modules and the way of their mounting and
the connection of input device cables to electronic modules.
Safety of data processing equipment

This is described more fully in Part Four, but for the purposes of this chapter
the general requirement is that the equipment be designed and constructed to
ensure that there is no danger to persons or environment during normal and
abnormal use. Special attention should be paid to process input devices, in
view of their work in hazardous areas. (Testing of input devices is also de-
scribed in Part Four).
2.5 PROCESS HARDWARE FOR DATA OUTPUT
2.5.1 FINAL CONTROL DEVICES
Final control devices are executive elements of a control system that directly
change the value of the manipulated variable. Final control devices are both
continuous and discrete. They are made out of drives and executive devices.
Examples of drives are valves, actuators, valve positioners, speed variators,
solenoid actuators and servomotor actuators. Drives are connected to the ex-
ecutive devices: shutdown valves, transformer tap positioners, air dampers,
gearing mechanisms and electrohydraulic valves.
The transfer characteristics of continuous-type drives and executive devices
Process hardware for data output 71
feedback.
device
".I:; .,.'
Figure 2.15 The overall control scheme for final control devices.
are usually nonlinear. To obtain the required precision of manipulated vari-

ables an overall control scheme is applied (see Figure 2.15) for an
electrohydraulic regulator. By applying a computerized scheme even intelli-
gent positioners are becoming feasible .
An executive device is an element of the control path that directly changes
the amount of manipulated variable. Executive devices are generally (1) me-
chanical types such as valves, dampers, vices, (2) electrical types such as
generator exciters, solenoids, magnetic amplifiers or (3) optical types such as
laser exciters.
An actuator is a final control device and a mechanical executive device
that can be used for positioning throttling devices. Typical specifications of an
actuator are water-tight enclosure; hydraulic integral system; electric motor-
driven 225 VA; electric power input - single phase; control signal input - 4-
20 rnA; mode of operation - direct or reverse action of shaft upon control
72 j The basic process unit
Table 2.5 Electronic control devices for current
Control device Switching Switching Remark

current range voltage range
Power transistor 2-100 A 60-1500 V Approximate switched

power 100 W -10 kW
Thyristor (silicon 1-1500 A 50-2000 V Average current
controlled rectifier)
Triac 3-40 A 100-600 V RMS current
signal increase; power or signal failure - spring return; stroke range up to 10

cm; stroke speed 0.9 cm/s; operating ambient temperature -40°C to 60°C,
humidity 0 to 100%, atmosphere (industrial including hazardous), vibration 0
to 120 Hz up to 7 G's, nuclear and seismic capabilities (IEEE standard No.
344. 1973).
A valve is a final control device with an executive device modifying the
rate of flow of a fluid. The control of the moving part of the valve can be
manual or automatic, and be carried out in a continuous or stepwise manner.
A typical specification of a valve for the control of piston-operated line valves
is: 6 part 4 way valve, 6 mm; working pressures up to 8 bars; coil supply
voltage from 6 V to 250 V DC; enclosure - waterproof; humidity - submers-
ible with suitable cable gland; manual override and reset.
A valve positioner is an auxiliary control system designed to increase the
accuracy and speed of the positioning of the executive part of the valve by a
feedback mechanism similar to that shown in Figure 2.15.
A speed varia tor is a coupling device actuated to vary the ratio of the
angular velocities of a driving shaft, and of a driven shaft, or the ratio of
available torques on these shafts.
A solenoid actuator is an actuator where the displacement of the control-
led part is obtained by the attraction of the armature of a solenoid.
A servomotor actuator is an actuator where the displacement of controlled
part is obtained by the movement of the motor part, which is energized by a
power input (e.g. 120 V AC) through relay contacts. The position of the
actuator is sensed by a feedback signal that is compared with the set-point
input signal. Any imbalance of the set-point value causes the appropriate motion
of the motor.
Electronic control devices using current are active electronic components
with at least three electrodes acting as a current path between two electrodes
when an appropriate signal is applied to the third or control electrode. Elec-
tronic control devices include the thyristor or semiconductor controlled rectifier,
the triac and the power transistor. Examples and a description of these devices
are given in Table 2.5.
Auxiliary data for process control 73
For the range of applied final control devices, a whole set of supporting
hardware components is required to enable the proper functioning of these
output devices. Special features of output devices are devoted to operator-
process communication functions. These functions are:
1. Display and manipulation of set-point and output signals. An analog in-

terface unit generally provides the operator with a display of process
variables, indication of the set-point value and indication of the output
signal; the operator is also able to manipulate the set-point value.
2. Mode selection and indication. The operator is able to change the mode
of control from manual to automatic, and vice versa, by throwing a switch.
3. Bump-free transfer to manual and automatic mode of operation. Analog
and digital control systems and regulators enable bumpless, balance-free
transfer between automatic and manual modes of operation according to
ramp or other type of transfer function.
4. Single supply operation. Most systems operate from a single 24 V DC
power supply, permitting simple battery back-up.
5. Special output features. Manual adjustments of executive devices are of-
ten enhanced by variable speed incrementation of the output; such a system
offers large fluctuations of the process variable within a reasonably short
time, with no damages incurred.
Ordinary signals for the hardware of final control devices are given in Table
2.6 for different types of output devices and connections of output devices to
data processing equipment. (Testing and safety of final control devices are
described in Part Four.)
Various commands issued from the control system to final control devices
and other parts of the control system are indicated in Table 2.7. For a basic
process unit, Figure 2.16 shows the standard data on input/output and process
equipment (PI-diagram).
2.6 AUXILIARY DATA FOR PROCESS CONTROL
Auxiliary data for process control are obtained during data communication
between a process control station or remote terminal units. These data are
issued either on the side of higher or lower hierarchical level, and either as
process information or process control information (see Table 2.8). Process
information includes:
acknowledgement information given as a response to an interrogation

command and referring to a process state;
interrogation command issued as a general interrogation command;
station interrogation command or specific interrogation command.
Table 2.6 Standard process output devices and their commands
Output device Output variable Information Output Common

signal connection characteristics
Binary-type Any process Quantized Binary Closing contacts

output device variable information signal
Insulated from any
parameter
ground circuit or
conductive parts
Volt-free
~ 10+7 operations
Contacts with breaking
capacity
10 VA, 48 V/O.S A
Analog-type Any process Any value in Analog Floating loop
output device variable a given range signal (~ 100 M Ohms,
~ 0.5 kV) with no
common point
Standard DC current
4-20 rnA, 0-20 rnA or
voltage 0-10 V
Controller load not
exceeding 500 Ohms
Process control information includes:

change of state information issued on check command;
transmission error alarm issued when erroneous communication of data
occurs;
equipment breakdown information issued by process control equipment in
case of detectable breakdown of equipment.
Table 2.7 Different types of command
Designation Feature
Single command Type of command with one output signal, usually volt-free
open contacts
Double command Type of command with two output signals in series, usually
volt-free open contacts
Pulse command Command signal usually of shorter time duration, i.e. 5 ms
to 500 ms
Maintained command Command signal usually of longer time duration, i.e. 500 ms
to 100 s
Persistent command Command signal continually applied to an output device
Starting command Command signal applied to start a plant or a part thereof
Stop command Command signal to stop or shutdown a plant or a part
thereof
Switching command Command signal to switch over the process apparatus from
one to another working state
Set-point (adjusting) Command signal to adjust the state of process apparatus to
command appropriate state
Persistent regulating Command signal of continuous type to keep the process
command regulator in a given operating state
Regulating step command Command signal issued in step-by-step manner
Command signal to enable a selection of specific process
Selective command apparatus
Group command Command signal issued to a group of process apparatus
Broadcast command Command signal issued to all peripheral process control
units
Instruction command Command signal issued to the control system operator
Function command Command signal of a specific system function
411 8------------.. 510
412 8--------------..----;
I
413 Q ------------.. -----~-----------,I i-----·

~ I I:
414 0-------------~----+-----------L---j ~
415 8------------.. 510 1
start stop
I
I
I
I
416@ 416©
RTK 410
Figure 2.16 Process and signal paths in a basic process unit: - - , crude-oil
flow; ----, signal flow; 0, transmitter; ©, pump motor; 510, another basic
process unit.
Table 2.8 Auxiliary data for process control
Process information Process control information
Acknowledgement information Change of state information

Interrogation command Check command
Station interrogation command Equipment breakdown information
General interrogation command Transmission error alarm
References 77
REFERENCES
Bogner, R. and Constantinides, A. (1975) Introduction to Digital Filtering, Wiley, New

York.
Gorez, R. (1981) Int. J. Systems Sci., 12, No.8, 907.
Katz, P. (1981) Digital Control Using Microprocessors, Prentice-Hall International,
London.
Kwakernaak, H. and Silvan, R. (1972) Linear Optimal Control Systems, Wiley
Interscience, New York.
Novakovic, B. (1981) Proc. of 26th JUREMA Symposium, Part 1, 43, Zagreb.
3
Stratification of control tasks and
data communication
3.1 INTRODUCTION
The task of managing a large company is achieved by a structure in which

responsibility is delegated to individuals or groups. In tum these may delegate
to others until the management structure resembles a tree.
Exactly the same approach is relevant to process control systems and this
chapter is concerned with the different levels or strata of control which are
normally identified.
Examples of the different control tasks are regulation, process supervision,
process measurement, process protection and on/off control. A particular
computer-based functional system (or systems), performing a specific task, is
associated with each control level. Computers are also used for processing
input and output data, and for process instrumentation. A survey of applicable
process control systems is also presented in this chapter.
Computers at different control levels are connected via communication lines.
All process control systems receive input data from transducers, issue control
signals, interact with an operator and print reports. These tasks all involve
forms of communication and apply even to modest installations. As size
increases, the computing function is usually split between a number of inter-
communicating processors. This offers advantages for modularity, system in-
tegrity and expandability. More communication is the price to be paid for
these advantages.
There are many approaches to the communication problem in process control.
The first is to use leased lines. This is only for point-to-point communication
and was originally developed for other purposes and adapted for process con-
trol use.
The second approach is more particularly engineered for inter-processor
communication and usually interconnects a number of processors.
New trends towards communication standards such as seven-layer MAP
(Manufacturing Automation Protocol) allow interfaces between different control
equipment suppliers. By using two of its layers a breakthrough in process
control emerges promising the connection of single devices (smart transmitters
and valve positioners) directly to the process computer.
78
Stratification of computer tasks 79
3.2 STRATIFICATION OF COMPUTER TASKS

Seven main and independent computer tasks are recognized in the realization
of process control systems;
1. Process instrumentation.
2. Process regulation.
3. The acquisition of measurement data.
4. Process supervision.
5. Protection and process safety.
6. On-off process control.
7. Set-point process control (governing).
Particular data processing tasks are performed on specific control variables,
but essentially all are carried out on the same set of process data at the level
of the basic process unit. For each particular process (e.g. electric power
distribution system, thermoelectric power plant, production, transportation and
distribution of natural gas and crude oil, large water supply system, chemical
process industry, numerical control of machine tools, cold strip mills, paper
mills or a cement factory), the same procedure of task stratification can be
applied.
A bottom-up approach to the specification of process control systems takes
into account the function of the basic process unit and unifies large mechanical
process parts, instrumentation and control devices. The control of complex
processes forces the coordination of functions of basic process units and
represents a top-down approach to the control of complex processes. These
bottom-up and top-down approaches must be complied with as the realization
of control algorithms for complex processes takes place. Two examples are
given below for the realization of process control algorithms of two different
electric power supply stations where the analysis of control functions of the
basic process unit differs considerably.
3.2.1 EXAMPLE 1
An analysis of control functions of the basic process unit for power supply of
a crude-oil measurement station is given in Figure 3.1 and, as seen from the
synoptics of the basic process unit, the voltage is applied through a voltage
transformer TRI to:
an ACIDC converter and to the field instrumentation as 24 DC power
supply;
- a separate branch controlled with the voltage switch VS3;
a branch fed uncontrolled to other parts of the field consumers.
The consumption of supplied electric power is controlled through a Wb-coun-
ter on a special device level acquiring pulse count signals C. The signals of
supervisory
level
connection
remote
terminal
unit
connection
t t
kwh
t t
special
counter
device
level I DJiJiJ I
"7
connection
process
controller
functions
connection
synoptics
of the
basic
process
unit
connection
functions of N
::r:: ..... .....
the basic ..... + + + +
process unit +
..... ~
~
c<')
~
'<t'
~
>n
~
9 9 9 9
~
9 ....
<: CQ 0 ~
II
,.....
Y
II II
connection
[ ~~
L! L L u
Q~ ~~
....
process data <: ""
CQ 0
vs vs eet vs vs vs
and commands
Figure 3.1 Analysis of control functions for the power supply of a crude-oil
measurement station: VS, voltage switch (signal); EET, electric energy pulse
transmitter (signal),
voltage switches (A, B, D, E and F) are processed in the basic process unit
and fed to the process controller, where all the necessary set/reset functions
and a connection L28 to other basic process units are executed. For the purposes
of process supervision, all the signals from voltage switches are then fed to
the remote terminal unit.
Figure 3.1 also shows the following main data processing tasks:
integrity, protection and safety of the basic process unit;
- supervision of basic process unit;
- measurement and presentation of process data.
Simple analyses of process control functions are largely dependent on the
structure of the controlled process. Thus, a relatively simple power supply unit
that is independent of other process parts in a crude-oil production field, which
supplies only one measurement station in the field, can be protected and
supervised in a particular manner. Equally, this unit is also included in higher
control levels by sending its data for further data processing to these levels.
The results obtained can be used for the control of other power supply units
in the same crude-oil production field, or even for a group of production
fields. There is no direct automatic action on this basic process unit from
other power supply units, since it is not needed. The provision of direct automatic
action from other basic process units would complicate the structure of
algorithms performed on the process controller at the level of the basic process
unit, as can be seen from Example 2 below.
3.2.2 EXAMPLE 2
The algorithm for control actions of all distribution substations (see Figure
3.2) put in one control centre is highly complex. This is due to the complexity
of the problem, implying that the total power system of generation, transmission
and distribution of electric energy should be represented and computed in the
real-time environment, that is at the time of data acquisition. The proper conlrol
of such a system requires the monitoring and processing of real-time power
system variables, and then the execution of the desired feasible control actions
directly on the system's side remote control equipment. Instead, considering
the system's spatial structure and local control objective characteristics, the
overall network can be reduced to a set of distribution subnetworks as shown
in Figure 3.3. Each subnetwork is provided with independent local control
hardware which consists of a distribution substation feeding one radial load
bus centre. A distribution substation consists of a set of m parallel transformers,
capacitors and reactors.
The power flow is controlled by transformer tap positions, and by switching
capacitors in or out to control the power factor. The aim is to maintain safe
stable operation within the constraints. The load between phases must be kept
82 Stratification of control tasks and data communication
IIOkV bus
transmission
network
,..- ,..
-" .... ",," ,.. .... ,.. "- A ....
llO/35kV
transformers
35kV bus
distribution
t capacitors/
~ reactors
substations
load
} centres
1 to N
Figure 3.2 One-line presentation of a distribution network.
in balance. The voltage must be controlled and heat and transmission losses
minimized by adjustment of the power factor.
Each subnetwork is decoupled from the rest of the network. Two assumptions
are made.
1. Constant load demand during data collection, data processing and process
control implementation.
2. Constrained requirements of localized control actions .
•
The formulated control actions are performed as a set of noninteractive local
controls, with little overall effect on the system operations.
The overall control strategy of the radial load bus subnetwork enables the
control to be reduced to its priority functions which are executed independently
and in order of priority. The priorities and their criteria are described as follows
(Arafeh, 1978).
Priority 1 control function Transformers leading power factor control

transformers, capacitors and reactors which are controlled for a specified leading
power factor angle limit a UM so that the ith transformer complex power
flow Sj satisfies
generator
1
generation bus
voltage ratio
transformers
phase angle
transformers
radial load bus
load centre
I capacitor bank
Figure 3.3 A set of distribution subnetworks.
(3.1)
where 1m and Re denote reactive and active parts of the complex power
respectively.
Priority 2 control function Load centre voltage control for specified voltage
schedule V sc ' voltage tolerance VTOL and calculated load centre voltage VL ,
transformers, capacitors and reactors are controlled so that the change in load
centre voltage satisfies
Priority 3 control function Transformers MW feedback control the ith

transformer to satisfy
Re(Sj) ~ 0 for i = 1, 2, 3, ... , m (3.3)
where Re (Sj) denotes a real part of complex power flow Sj.

Priority 4 control function Transfonners MV AR feedback control the ith

transfonner to satisfy
sign {Im(S;)} = sign {Im(~ S;)} for i = 1,2,3, ... , m (3.4)
Priority 5 control function Transfonners MV A balance leading control the

ith transfonner for specified MV A tolerance MV ATOL to satisfy
At; = {[Abs(S;)M ~ Abs(Sj)] -I]} ~ MVA TOL for i = 1,2, 3, ... , (3.5)
where ti is the ith transformer distribution factor based on the parallel

transfonners MV A ratings MV ARi and is given by
m
ti = MVARJI, MV A Rj for i = 1, 2, 3, ... , m (3.6)
j=1
and dti is the deviation of the ith transfonner from the balanced condition.
Priority 6 control function Substation bus lagging power factor controls a

capacitor or reactor which is switched on/off if a capacitor or reactor exists
with a reactive power value Qc satisfying
m
(3.7)
i=1
Respective constraints on transfonner tap position change and capacitor or
reactor switching operations are summed to be embedded within each control
function. The control procedure tries to fulfil the control functions according
to their priorities. This means that certain priority level functions are dealt
with once all higher priority level functions have been satisfied.
Analysis of control functions of the basic process unit of a voltage ratio
transfonner in a distribution substation is shown in Figure 3.4. The vector
diagram of voltage changes as the result of transfonner tap movement is
illustrated in Figure 3.5, and calculations for a subnetwork control are further
outlined in Table 3.1. The control algorithms of a process controller are
summarized in Table 3.2. The values of equations (3.1) to (3.6) are calculated
according to the commands given in Table 3.1, and, where the feasible solution
for the network applying all rules cannot be found, the subnetwork operator
is infonned accordingly.
A relatively more complex task of the computer control system is the control
of the thennal power plant. Control functions depend to a great extent on the
power generation part, and Table 3.3 outlines the typical distribution of on/off
control functions and regulation of a 650 MW thennal power unit (Reding,
1979). A systematic stratification of control functions for a part of a thennal
synoptics
of the substations
t
operator
t t t t
process
controller control algorithms are as given in Table 3.1
,
I I"
# L ~ L ~
I I I I I
basic basic protection and control algorithms
process are as given in Arafeh, 1978
unit
,
i ( t t'" t t t'"
~ ~ ~
synoptics
of the
basic
process
unit
process
device
level
v = Vn+ no~ V n; V 0 = V L + Is / Y f
~V=~ +~V
n n
~ n = n - no
Figure 3.4 Analysis of control functions of a voltage ratio transformer: n, new tap
position; no, instantaneous tap pQ.sition; f/' feeder admittance; Y s ' feed to load
admittance; Y, transformer admittance.
v o +!:::,V0
reference vector Vo
Figure 3.5 The voltage vector diagram of transformer tap movement.
power unit is shown in Figure 3.6. A great many specific control, regulation,
supervision, measurement and protection functions are attributed to different
levels and parts of a thermopower unit, in which numerous functional systems
are installed with thousands of control loops.
3.3 CONTROL LEVELS AND COMPUTER INPUT/OUTPUT

HARDWARE
Although a wide variety of functional systems exists, it seems that five

independent levels of control are recognized in computer controlled process
control.
1. Complex process instrumentation and process device level.
2. Process controller level.
3. Level of data acquisition or remote terminal unit or direct digital control.
4. Level of process control centre or supervisory control or district centre.
5. Level of despatching centre or management control centre.
Input/output hardware for these levels is described below, together with their
performance, and is also illustrated in Figure 3.7. The specific tasks of each
Control levels and computer input/output hardware 87
control level and corresponding input/output hardware facilities are given in

Table 3.4, and particular features and types of input/output circuits and modules
are outlined in Tables 3.5 to 3.7.
3.3.1 LEVEL 1
Functional systems at level 1 can accomplish some or all of the following

functions:
- Receiving signals from signal transducers or signal converters.
- Issuing control signals to various process drives.
- Issuing converted process signals to level 2 or 3 systems.
- Receiving commands from level 2 or 3 systems.
- Complex process data calculation.
- Indicating the state of signals and drives to process data synoptics or to
indicators.
- Regulator functions.
3.3.2 LEVEL 2

functions.
- Receiving signals from signal transducers or signal converters.
- Issuing control signals and commands to various process drives.
- Issuing and receiving data and commands from levelland 3 systems.
- On/off control, logic· and arithmetic functions for process control and
protection.
- Indication of process state to synoptics of the controller.
3.3.3 LEVEL 3

functions.
- Receiving process data from signal transducers or signal converters.
Issuing control signals and commands to various process devices.
Connection to level 1 systems with simple series communication of signals
and/or with bus or coded signals.
Connection to level 2 systems with standard series communication of signals
and/or with bus or coded signals.
Table 3.1 Calculations for subnetwork control
Process Approximate process model Process constants Process variables

1
Power a =- for Hi....{) -16 voltage rate (VR) a - transfonner type constant Vj - transfonner internal voltage at
distribution 160 transfonner neutral tap (ith transfonner)
m - number of transfonners in
subnetwork
parallel Sj - the ith transfonner complex power
-a = -1601 (1-2 + J.,f3J
2
'
- for 16-0-16 flow
Vo - the reference voltage
phase angle (PA) transfonner (measured substation bus Yj - the ith transfonner admittance
output v~ltage at
11Vj - the ith transfonner internal
-Vj 1
( Vo + ~S+ ; i J the secondary side)
= _ =1, 2, ... m voltage change per tap
1 + ailloi No Yj nOi - tap setting before move-
S~ - conjugate complex of the Sj
ment of the ith transfonner
!!VJ = a;Vi; 1,2, ... m
Y, - substation load admittance
nj - tap setting after tap
movement Q - capacitor or reactor reactive
LSj Yf - the admittance of the
power value (MV AR)
y =...i=l.- + jQ
, VJ feeder to the load
11Vo - substation bus voltage incremental
change
!!n. = n. =+= no - incremental (decremental)

Lj I1n j!!VjYj - Vo!!Y,
I1Vo = ..2.
...::.c'..."....._ _ _ _ __
, 'tap mbvement of the ith transfonner
LY;+Y,+I1Y,
i:=1
~Y, - the change of substation load
~Si = (VO + ~VO)(~ni~Vi - Vo)*Y'*; admittance
VL - the load centre voltage

~VOSi .
+--; 1= 1,2, ... m
Vo
v, Vo _
~ ~ (t S: + 1 jQ
Y1 Vo
~v
_[ i = \_ ,=1 st + jQ! ~'=I _ _
[i. ~ni~ VSi ~
L V2 m
YI 0 ~ - -
.£... Yi + Y,
1=1
Table 3.2 The control algorithms of a process controller
Procedure The content of operation procedures

steps
Control of transformers leading power factor (3.1)

2 Load centre voltage control (3.2)
3 Transformers in feedback control (3.3)
4 Transformers MVAR feedback control (3.4)
5 Transformers MV A balance leading control (3.5)
6 Substation bus lagging power factor control (3.6)
Rule number Heuristic control rules
Switching on/off of capacitors - for load power factor change

2 Switching on/off of reactors - for load power factor change
3 Parallel change of tap movements - for bus voltage magnitude change
4 Increase tap positions on the feedback transformers - for decrease of
MW and MVAR feedback
5 Phase angle transformers control the real and reactive power flow while
voltage ratio transformer control the reactive power flow
6 MVA loading is increased by increasing the tap position
7 Intended opposite tap changes control the power distribution in parallel
transformers keeping voltage magnitude constant
Table 3.3 Typical distribution of on/off functions and regulation of a 650 MW thermal
power unit
Functions System part Control system
On/off control 650 MW turbo set and On/off control system of steam
generator cooling system turboset
Feedwater pump turbine
Air turbines On/off control system
Main boiler including
oilbumer and coal pulverizer
Gas turbine On/off control system for gas
turbine
Regulation 650 MW turboset Turboset regulator
Feedwater pump turbine Regulator of the feedwater pump
turbine
Air turbines Turbine regulators
Boiler Boiler control
-- --
thermal control of
power thermal
unit power unit
level
--- I
I
boiler
control
I
---I 1---
group of
I I
combustion
function feedwater
process
level control
control
___ .-11 1L ___
L~::-
group of group of
burner 1 burner 2
control control
---~ ~---
burner 1 - 1 burner 1 - 2
control control
---
I I
process I I
burner 1 burner 1
devices and
safety control protection
level devices devices
Figure 3.6 A stratification scheme for a thermal power unit.

complex data despatch centre-
level 5 presentation management
'"
...s
and control control
r:::: <L>
~~-
U
process control-
'">-
'"
~
supervisory ~
process data supervisory ...u

.g
level 4 access and control- ~
~}]
control district centre
U
data acquisition- ~.~
~
operator's remote terminal
level 3
console unit-direct f4
digital control joI
U
process data
level 2 synoptics or process controller
indicators 4
t- ~ 11 1 t
complex process instrumentation-
level 1 regulator
- process drives control
. . =-
--------
signal transducers and drives
process hardware
coded or bus digital signals; -+,

Figure 3.7 Five levels of control. ~, connection of control levels through
connection of control levels through digital or
analog signals transmitted in series.
Table 3.4 Tasks, control levels and input/output hardware
Control level Tasks Input/output hardware
Complex process Calculation of complex Analog and digital input/output

instrumentation and process variables from given circuits
smart transmitters process signals Pulse input circuits
Calibration and expandability Serial bus connection
of measuring range
Digital input and output
circuits
Pulse input and output circuits
Standard serial interface to
remote terminal unit
Process controller Logic and protection functions Digital input and output
circuits
Control of basic process unit
Coordination of a few basic
process units Standard serial interface to
Communication with remote
terminal units
Data acquisition Control of process variables Analog input and output circuit
Remote terminal unit
Communication: with Digital input and output
Direct digital control
controller, complex process circuits
instruments, other RTUs, and
superimposed digital computer
Standard interface to controller
Support of operator's console
Operator's console interface
functions
Communication interface
District centre Optimizing functions Communication interface
Supervisory control
Process control centre Process data issuing Operator's console interface
Data logging Standard data bus
Expert system software
Engineering calculation
Account issuance
Program compilation
Data supervision
Dispatch centre Broad band optimizing Standard communication

Management control functions interface
Expert system software Opertor's console interface
Account issuance Peripheral devices interface
Off-line programs
(simulation and modelling)
Program compilation
Overall process supervision
Table 3.5 Analog input circuit (according to VDIVDE 3552) applied to Teleperm ME
AS220E analog input module
Types of analog input*

A B C D
Types of circuit disposable (Y, N) N Y N N

2 Galvanic isolation (Y, N) N
3 Number of digits of digitized input value (bit) 12
4 Maximum speed of input value (value/s) 330
5 Percentage error at maximum speed (% of
total span) NA
6 Maximum scanning rate within item 5, given
percentage error (kHz) 0.250
7 Possibility of external triggerring (Y, N) N
8 Maximum percentage error (% of total span) NA ± 0.4
9 Speed of input value corresponding to item 7
(value/s) NA
10 Number of measurement regions (number) 5
minimum voltage measurement value (mV) NA
maximum voltage measurement value (V) ±1O
minimum current measurement value (rnA) 0 4
maximum current measurement value (rnA) 20
minimum resistance measurement value (Q) NA
maximum resistance measurement value (kQ) NA
11 Programmable switchover of measurement's
region (Y, N) N
12 Automatic switchover of measurement's region
(Y, N) N
13 Rated voltages (number) 2
14 Block length if block read-in possible (value/
block) 4
15 Allowable overloading (% of maximum value of
the measurement region) 10
16 Automatic error diagnosis included (Y, N) Y
17 Minimum configuration (number of inputs)
18 Maximum configuration (number of inputs) 4
19 Minimum step increase of inputs (number of
inputs)
* A = closed contact immediate B = closed contact integrated

C = open contact immediate D = open contact integrated
Control levels and computer input/output hardware 95
Table 3.6 Digital input circuit (according to VDIIVDE 3552)
Types of digital input*
A B c D
Types of circuit disposable (Y, N)

2 Self-indicating (Y, N)
3 Galvanic isolation (Y, N)
4 Maximum input speed with changing address
(kbit/s)
5 Maximum input speed with constant address
(kbit/s)
6 Number of information bits per input word
(number)
7 Block length if block read-in possible (word/
block)
8 Read-in of one word at time of input change,
hardware disabled (Y, N)
9 Rated value of digital input voltage (V)
10 Allowed overloading (% of the rated value of
digital input voltage)
11 Automatic error diagnosis included (Y, N)
12 Minimum configuration (number of digital inputs)
13 Maximum configuration (number of digital inputs)
14 Minimum step increase of input (number of digital
inputs)
* A = static closed contact C = dynamic closed contact

B = static open contact D = dynamic open contact
Connection to level 4 and level 5 systems through bus or coded commu-

nications.
Connection to other functional systems.
Connection to operator's console to indicate process states and receive
commands for process control.
Data acquisition.
Direct digital control.
Remote data acquisition and process control.
Algorithms for on/off control and process protection.
Table 3.7 Digital output circuit (according to VDIIVDE 3552) applied to Teleperm
ME AS220E on/off control module 6DS 15I0-8AA
Types of digital output*
A B
Types of circuit disposable (Y, N) N Y

2 Galvanic isolation N
3 Maximum output value speed with changing
address (kbit/s) NA
4 Maximum output value speed with changing
address (kbit/s) NA
5 Number of information bits per output word
(bit/word) 24
6 Block length if block read-out possible
(word/block) NA
7 Rated value of digital output voltage (V) 24
8 Allowed overload (0) Overload Protected
9 Switched power (VA) 3
10 Automatic error diagnosis included (Y, N) Y
II Minimum configuration (number of outputs) 2
12 Maximum configuration (number of outputs)
13 Minimum increase of outputs (number of
outputs)
* A = relay output
B = electronic output
3.3.4 LEVEL 4

functions.
Connection to levels 3 and 5, and other functional systems through standard
bus or coded communications.
Connection to operator and system supervisory consoles.
Supervision and coordination of process variables.
Data acquisition and modelling of the controlled process.
Optimization of process work and process control.
Front-end processing towards the level 5 system.
Characteristics of process control computer systems 97
IG
.-16 Mbit
/",/ I RAM
./
,,"
/,.-"2 Mbit
// I RAM
:
" I
" 1
1M /,/'
:
,," 1
/"
"64kByte I
/1 I
/ / : memory :
/' I I
,,," I I
/ I 1
/' I I
..-lkMOS I :
lk ....
/",/ I
I
RAM iI 1
I
,/,/: I
/ 1
.fourbit I
/ / I TTLcouhter
/f ~T~ ?istable I
Clrclpt
:
// 1 I
,.RT~:
// : logic gate 1 1
~--~-=~~-L---1--~----r---~--~------~-------r--~year
1958 1964 1969 1974 1979 1984 1989

Figure 3.8 The density of components in integrated circuits.
3.3.5 LEVEL 5

functions.
Connection to level 3 and level 4 systems through bus or coded
communications.
Connection to display units for complex data presentation.
Issuing commands to subordinated control levels in order to coordinate
and control overall system performances.
Modelling and optimization of controlled process and process control.
Process prediction calculation, planning and system development calculation.
Management control of the system.
3.4 CHARACTERISTICS OF PROCESS CONTROL COMPUTER

SYSTEMS
Since the production of the planar transistor in 1959, the number of elements
in advanced integrated circuits has, according to the prediction of G.E. Moore
(Noyce, 1977), doubled annually (see Figure 3.8). Less intermediate testing is
needed in control system production because the correct functioning of the

complex integrated circuits has already been ensured. The end user now needs
less floor space, less operating power and less air-conditioning for the control
equipment. The most cost-efficient design between high assembly costs (incurred
at a low level of integration) and high scrapping costs (incurred at a high
level of microchip integration) has higher scrapping costs. The linear di-
mensions of the circuit elements can probably be reduced to about one
fifth of the size it was in 1977 (i.e. about 0.06 mm for transistors (Noyce,
1977» before any fundamental limits are encountered. Global characteristics
of computer systems for process control (e.g. dataflow (number of job in time
unit), reliability (number of failures in a year), serviceability (l/mean repair
time), allowable installation conditions (temperature span; sensitivity to dust
and aggressive atmosphere» are highly influenced by the state of solid state
technology. Global characteristics of computer systems for process control are
given in Table 3.8. Particular characteristics of computer systems for process
control are given in Tables 3.9 to 3.13.
3.5 A SURVEY OF PROCESS CONTROL COMPUTER

HARDWARE
Computer systems for process control have been developed based on the con-
cept of functional systems. Such systems are recognized as basic building
blocks when complex process automation takes place. Manufacturers of
computer process control systems share different approaches to these, which
are as follows.
1. Series or parallel data communication concept.
2. Different operator/system communication facilities.
3. More hardware- or more software-oriented implementation of functional
specifications.
The main features for two different manufacturers of computer hardware for
process control are summarized in Table 3.14. A more universal picture of
computer systems for process control can be given as a four-level control
system, although the approaches to the implementation of hardware for these
manufacturers represent a functional system concept. The basic configuration
of each functional system is shown in Figures 3.9 to 3.12, and their main
features are given in Tables 3.15 to 3.17.
Microcomputer-based systems for distributed data processing and
minicomputer systems need fast buses (up to 20 Mbyte/s) for internal data
communication (Warren, 1983). There are more than 140 different bus designs
identified in Europe alone. The most popular are the Multibus IEEE-796, the
VME buses, the VERSA bus and the S-100 (IEEE-696), which are extremely
important in internal system design. Hardware design is influenced by the
need to achieve cost/performance optimization using the standard modules on
Table 3.8 Global characteristics of computer system for process control
Control level Minimum system requirements Characteristics

and possibilities
Process controller and RAM and ROM memory only Number of inputs: 25-64
instrument multiplexer to ensure maximum reliability
The length of analog input
Interface hardware designed connections < 30 m
in the range of computer
Input scanning rate 1-10 ms
hardware
Temperature span 0-50°C
Active error protection
Humidity span up to 100%
relative humidity condensation
free
Mean time between failure
6000 hours
MTTR 6-24 hours
Degree of protection IP 54
Process computer RAM and ROM memory only Number of inputs = 250
to ensure maximum reliability
Remote terminal unit Memory cycle 0.2 ~s
Parallel job execution
System organization - parallel
Disposition of hardware for
Mean time between failure
fast program changes
5 -10 000 hours
Interface hardware designed
Memory capacity up to 256 k
words
hardware
Word length - up to 32 bit
Doubled systems or error
protection when in direct MTTR 6-24 hours
digital control
Degree of protection IP 54
Temperature span 0-50°C
Humidity up to 100% relative
humidity, condensation free
District centre RAM, ROM and disc memory Memory cycle 0.1 ~s
Supervisory control Parallel job execution Mean time between failure
5-10 000 hours
Despatch centre Disposition of hardware for
fast program changes Memory capacity 1-4 M
Management control
words
Doubled systems and doubled
discs when in direct digital Word length 32 bit
control
Disc capacity 30-400 M
Interface hardware designed words
MTTR 6-48 hours
hardware
Temperature range 20-2YC
Humidity range up to 80%
relative humidity
MTTR = mean time to repair

Table 3.9 The characteristics of main semiconductor memory
Parameters of main memory A* B*
Absolute length of memory address (bit)

2 Number of word information bits (bit)
3 Number of parity control bits (bit)
4 Basic memory capacity (kwords)
5 Maximum memory capacity (kwords)
6 Minimum extension steps of memory capacity
(kwords)
7 Different types of memory according to memory
cycle (I-ls)
7.1 ROM
7.2 EPROM
7.3 PROM
7.4 RAM
8 Combination of A and B type memory Yes No
9 Number of kwords reserved for user Basic Maximum
9.1 ROM (kwords) version version
9.2 PROM (kwords)
9.3 EPROM (kwords)
9.4 RAM (kwords)
10 Number of addressable memory locations
10.1 Direct
10.2 Indirect
10.3 Relative
10.4 Indexed
11 Autoincrement/decrement addressing Yes No
12 Virtual addressing Yes No
13 Memory protection Write Read
13.1 Minimum protected unit (word) protection protection
13.2 Maximum protected field (word)
14 Time interval of power supply back-up for
battery-protected memory (hour)
A* - Semiconductor memory of the type EPROM, PROM, ROM

B* - Semiconductor memory of the type RAM (DRAM, SRAM)
Channel capacity 101
Table 3.10 Processor characteristics
Total number of instructions

2 Number of input/output instructions
3 Addition time for fixed-point arithmetic when
operands and result are one memory word (~s)
4 Execution time for the calculation of R = Operand length

(A + B . C)/D, when operands and result are given 1 word 2 words
as memory locations (~s)
4 Fixed-point arithmetic 2 words 4 words
Floating-point arithmetic
5 Total number of registers available for programming
Operand registers
Address registers
6 User-defined micro programming instructions Yes No
7 Interrupt signal of power supply disturbance Yes No
the market. An example of cost/performance compromise in hardware is

obtained by using two or more buses in a system where a dedicated backplane
is used as the primary bus and the standard double serial communication is
used as the connection with other distributed data processing resources.
3.6 COMMUNICA TION CODES AND CIRCUITS
Two types of data communication exist in a process control system: one between
the process control system and simple process devices such as transmitters
and final control elements, and the other between various parts of the process
control hierarchy. Data communication in the process control system is based
on the messages appearing at the transmitter and received at the receiver. A
definite trend in technical systems is towards messages transmitted in characters
coded in a binary form from the binary message sources. Essential for the
proper functioning of process control systems is their error-free operation. But
the communication of data is influenced by different noise and disturbance
effects in communication circuits and media. Therefore the coding of technical
messages is of prime importance if control systems are to function correctly.
3.7 CHANNEL CAPACITY
An information channel is a physical system enabling transmission of messages

from the message source to its destination. Figure 3.13 illustrates the elements
Table 3.11 Interrupt unit characteristics
Waiting time (~)
2 Recognition time (Ils)

3 Recovery time (Ils) With Without
processor processor
4 Total number of interrupt inputs
5 Numbef of interrupt levels
Minimum configuration
Maximum configuration
Minimum expansion steps
6 Number of interrupt inputs for each level
Minimum configuration
Maximum configuration
Minimum expansion steps
7 Processing of interrupt for each level
According to time of occurrence Yes No
According to subpriority Yes No
8 Hardware solution of priority comparison between Yes No
new event and present processor priority
9 Register word saving at the interrupt
By software write-in in the main memory Yes No
By hardware write-in in the main memory Yes No
Skipping to another register word Yes No
Number of registers for the switch-over
10 Pushing of interrupt signals External mask Internal mask
Bit manner
Group manner
Total
11 External setting of mask
12 Memory elements of interrupt signals when Yes No
read-in automatic recovered
13 Supervision of processing of demanded interrupt Yes No
signals inside given time interval
14 Instruction for obtaining interrupt signal Yes No
15 Interrupt after each instruction possible Yes No
Table 3.12 Input/output structure characteristics
Data lines Control lines
1 Structure of I/O connections

Bus connection Yes No Yes No
Star connection Yes No Yes No
2 Number of I/O connections in basic configuration Input Output Total
(without extensions)
3 Number of I/O connections that can simultaneously
work sending character/word
4 Maximum data rate at I/O connection when
One I/O connection works
All I/O connections work
5 Execution time of I/O instruction
Normal condition - mean value (Jls)
Error condition - maximum value (Jls)
6 Peripheral's controlled data block transfer
Program controlled with the interrupt system Yes No
Through I/O processor channel structure Yes No
Through direct memory access Yes No
7 Switch-over of peripheral units between two central
processor units
Instruction controlled Yes No
Stationary switch-over Yes No
Table 3.13 Auxiliary field characteristics
Possibility of auxiliary elements protection when

unauthorized usage Yes No
2 Auxiliary functions Read Write
Access to central processor unit (CPU) registers Yes No Yes No
Access to main memory Yes No Yes No
Autoincrement Yes No Yes No
CPU hold at previously defined instruction address Yes No
Cyclic repetition of instruction Yes No
Start of firmware global test routine Yes No
Table 3.14 Main features of the system hardware of two computer manufacturers
Features System manufacturer
Siemens Honeywell
System type TELEPERM ME TDC 3000

System concept Two-three-level hierarchy Two-three-level hierarchly
System structure Serial doubled bus with timer Redundant serial bus
line (MAP: Token bus)
Basic functional system SIMATIC S5 family of PLC Process input/output unit
AS 220 family of control Multifunction controller
systems
AS 231 protocol system Application module
OS 250 operator configurator Universal station
station
SCICOMP surveillance
system
Data communication Serial token passing concept, Serial token passing concept,
between functional 340 kbit/s 5 Mbit/s
system
I/O unit:
Basic process interface Analog input 4-8 up to 256 analog input
unit (number of
Digital input 6-32 or up to 512 binary input
channels)
Digital output 1-2 or up to 96 pulse (counter)
input
Analog output 1-2
or up to 128 analog output
or up to 254 digital outputs
Modules for:
analog input 16
analog output 8
digital input 32
digital output 16
Operator interface - local single device - local interface
interface (AS 220) universal station
- local group device application module
interface (AS 220)
- operator control (OS)
- information system
(SCICOMP)
- structuring level (AS 231)
Scanning rate 2800 analog inputs/second 0.5 seconds for all 110
or signals
38 400 binary values/second
Main application field Thermal power plant Chemical batch process
Refinery
Chemical batch process Thermal power plant
Channel capacity 105
mICro-
computer +
timer
2 X serial
communication
channel
------
~
RW memory + ROmemory+
I/O ports I/O ports
~ ~
I/O circuits I/O circuits
for digital for digital
signals signals
] m
process onloff
operator's
devices and
panel
drives
•
•II
t
process
operator
Figure 3.9 A microcomputer-based process controller.
of the information channel. An output converter generates the message signal

suitable for the message receiver or user. The transmission system is com-
posed of a transmitter, transmission medium and receiver. There are always
stochastic uncontrolled noise signals (see Figure 3.13) that disturb the reception
of signals in a real information channel. Noise is a stochastic and uncontrolled
disturbance signal inseparable from the transmission system, which essentially
limits performance.
The channel capacity is a maximum speed enabling the transfer of
transinformation through the channel. Channel capacity is limited by the channel
noise. Channel capacity is
C, = max k'H(X,; Y,) bitls
(3.8)
P(X'i)
where
x" Y, are character sets of channel input/output;

k' is the number of characters transferred in the channel per second;
operator's II
•
console ~
micro- read-write read-only system serial

•11# her
serial ess
computer memory memory supervision communica- interlace trol
unit tion unit unit
/'L- system bus

~
ebus
analog digital analog digital special serial er
input input input input purpose communica- s
subsystem subsystem subsystem subsystem interlace tion interlace I
s
lb;>
process - drives and transmitters
f Figure 3.10 A microcomputer-based process control system.
to other
• system console floppy disks process
control
systems
central pro- main memory

IA
cessing unit and control
system bus
analog input digital input analog output digital output

subsystem subsystem subsystem subsystem
~
2x disk drive f process - drives and transmitters "/ process
control
systems
Figure 3.11 A small-to-medium microcomputer system.

console
subsystem
central diagnostic floppy
control microcomputer r------I'!
disc
AI
parallel
communication
communication
interface
QcomOi' interface
terminal I
cache bus peripheral devices bus
memory adaptor
interconnection bus series

process I/O
communication
subsystems
flOppyllfloppy
bus adaptor disc disc
6"~'~1
•
mass memory bus II •
U ,
Figure 3.12 A medium-to-large computer process control system.

Table 3.15 Main features of microcomputer-based process control systems
Number of units Speed characteristics
Central processor unit with main

memory (kbits) 64-256 :0; 0.2 Ils
Serial communication unit 2-4 300-9600 bd
Serial interface unit 2 300-9600 bd
Console unit 300-4800 bd
Analog input subsystem (channel) 64-256 :0; 30 ms/channel
Analog output subsystem (channel) 16-64
Digital input subsystem (channel) 64-256 :0; 20 ms/channel
Digital output subsystem (channel) 8-64
Serial printer ~ 100 char/s
Table 3.16 Main features of mini-supermicro-computer system
Computer units Number of units Speed characteristics
Central process unit with

main memory (kbyte) 128-512 :0; 0.1 s
Optional peripheral memory
(Mbyte) 1-2 :0; 20 ms data access time
Floppy disk unit 1-2 3i inch/5~ inch
Printer 1-2 ~ 200 char/s
Console unit 1-2 ~ 300 char/s
Serial communication unit 2-8 300-9600 bd
Analog input subsystem 256 :0; 30 ms!channel
(channel)
Analog output subsystem 16 :0; 30 ms/channel
(channel)
Digital input subsystem (channel) 256 :0; 10 ms/channel
Digital output system 64 :0; 10 ms/channel
Incremental clock :0; 20 ms per increment
Table 3.17 Main features of medium-to-large computer process control systems
Computer units Number of units Speed characteristics
Central process unit with 1024-8112 :;::; 0.05 s

main memory (kbyte)
Peripheral memory (Mbyte) 50-200 :;::; 20 ms
Peripheral memory (Mbyte) 80-1400 :;::; 80 ms
Printer 2-8 2:: 100 char/s
Line printer 2:: 600 line/s
Floppy disk unit 2-4
Alphanumeric display panel 2-8
Graphic display panel 1-3
Series communication unit 1-8 9600 bd-2 MBd
Analog input subsystem 512 :;::; 30 ms/channel
(channel)
Analog output subsystem 16-218
(channel)
Digital input subsystem 2048 :;::; 10 ms/channel
(channel)
Digital output subsystem 256
(channel)
Incremental clock :;::; 20 ms
Real-time clock
Watch-dog timer :;::; 1 ms
Tape unit 2-4
H(Xj; Y j ) = H(Xj) - H(XdYj ) is a transinfonnation;

P(XIi) is the probability of appearance of a given character.
The channel capacity of a binary symmetric channel is given by the expression

C] = k'(1 =1- q
+ p ld P + q ld q) bit/s, p (3.9)
where for a symmetric channel (Yll = XII = '0'; Y I2 = X I2 = '1 ')

P = P(YIl/xIl) = p( YI2/XI2)' q = p( YI2/xIl) = P(YIl/Xj2)
The entropy of a binary information source is given as
H = P ld P - (1 - p) ld (1 - p) = s (p) (3.10)
or
Cj = k'[1 - s(p)] bit/s (3.11)
Types of connection and communication hardware III
equi ocation irrelevance

H(x/y) H(y/ x)
(a)
transmiller rec iver
(b )
Figure 3.13 Elements of the information channel: (a) transinformation;

(b) functional elements of the channel.
When the transfer of signal with a power level S is disturbed with a white
noise of power N having the frequency bandwidth B then the channel capacity
is given by the expression (1.14). The power level can be calculated by squaring
corresponding signal and noise envelopes.
3.8 TYPES OF CONNECTION AND COMMUNICATION

HARDWARE
Two main communication modes are distinguished according to the basic
types of the timing of data sending a synchronous data communication and an
asynchronous data communication.
A synchronous communication transfers characters continuously in a time
series beginning with the start sequence and finishing with a stop sequence
(see Figure 3.14a). The sequence of 'idle characters' is transferred if a new
start sequence does not follow immediately after the stop sequence (see Fig-
ure 3.14b). Start, stop and idle character sequences must be unified in a mes-
sage. Main characteristics of synchronous data transfer are complexity, large
amount of information flow and use of higher data transfer speeds. An
asynchronous communication enables the transfer of data characters in a row,
with each character represented as an n-bit dataword. The transfer of character
begins with the bit that serves the synchronization of the receiver, followed
by the dataword, and concluded with bits of stop combination. The duration
(a)
~ Ise~~a:~ce I C C C C c Ise~~~~ce I J
t
~
(b)
Figure 3.14 A synchronous communication of (a) characters and (b) idle

characters: C, transfer character; IC, idle character.
(a)
communication channel
(b)
(c)
Figure 3.15 (a) A simplex mode of data transfer; (b) a half-duplex mode of data
transfer; (c) a duplex mode of data transfer.
of stop combination can be deliberately chosen (e.g. 4.15 bit intervals). The
main characteristics of the asynchronous communication are simplicity of
hardware, low amount of information flow suitable for small datawords from
5 to 9 bits, and use of transfer speeds of up to 19200 bit/so According to the
types of communication regarding the direction and possibility of simultaneous
reception and transmission, three modes of communications are distinguished
as shown in Figure 3.15.
Types of connection and communication hardware 113
1. A simplex mode of data communication uses one channel for each

direction.
2. A half-duplex mode of data communication uses one channel with the
facilities to change the direction of communication on a channel.
3. A duplex mode of data communication uses two communication chan-
nels each for one direction of the communication.
Communication networks for data transfer fall into three main types (see
Figure 3.16).
Data terminals when connected into a given configuration comply with certain
rules to enable good communication, regardless of the terminal configuration,
that is:
only one terminal is a control terminal;
a control terminal is activated for the system start-up and in the case of
system fault;
a control terminal defines a master terminal during normal operation;
a master terminal addressing slave terminal(s) responds to its requests in
normal mode.
By combining the three communication modes (i.e. simplex, half-duplex and
duplex) and three main communication configurations (i.e. point-to-point,
mUltipoint and loop) nine different networks can be obtained (see Figure 3.17).
The communication availability of these nine networks is very different,
being very low for simplex loop configuration and relatively high for duplex
point-to-point configuration. The length of communication lines differs in these
networks. The most practical network communications are shown in Figure
3.18, and exhibit a different length of communication lines, with the multipoint
configuration having the shortest lines. The number of communication units
also differs for the connections described in Figure 3.18. Multiple point-to-
point connection possesses the biggest number of communication units. Loop
configurations with duplex communication can exhibit high availability when
they are capable of communicating in a loop in a reverse direction should
breakdown on communication lines occur (see Figure 3.19). The change of
communication direction is possible by using additional hardware and soft-
ware in the data terminals.
In order to make a distinction between correct messages and those corrupted
by noise and interferences on the transmission line, data transmission security
is essential for a data terminal on the receiver side. Data transmission security
is obtained by adding security bits to the message. The security bits increase
the redundancy in such a way that minimum code distances of message
characters are increased. A code distance (Hamming distance) is defined as
the number of bit positions in characters or in a message where two bits
differ. Minimum code distance D enables a safe error detection and data recovery
for all bits in a character and in a message with D - 1 or less erroneous bits.
(a)
(b)
(c)
Figure 3.16 (a) Point-to-point communication configuration; (b) multipoint

communication configuration; (c) loop communication configuration.
simplex half-duplex duplex
point-to-point
multipoint
loop
Figure 3.17 Nine different data communication modes: C, control terminal;
M, master terminal; S, slave terminal.
Data recovery is usually a retransmission of a message where the error had

been detected.
A communication link for data transmission is established either through a
physical line, private or leased, or through a radio link that can also be both
private or leased. Normally, equipment for the connection to leased line
equipment is the same type used for leased line equipment.
Physical lines can be used for the transmission of data on one channel as
well as time or frequency multiplexed connections in multipoint or ring
connections. The use of multiplexers is explained in Figure 3.20a. Frequency
division multiplexers divide the bandwidth of a high speed communication
star connection
multi-point connection
Figure 3.18 The most practical communication networks.
facility into separate subchannels for each low speed terminal (see Figure
3.20b). Time division multiplexers divide a high speed communication into
time-sliced subchannels, as indicated in Figure 3.20c. A dedicated time slot is
provided for each terminal connected to the time-divided multiplexer. Equipment
for a radio link can be divided into radio equipment, multiplex equipment,
antennas and leased-line connection equipment.
I
I
I
I
I I
~ line breakdown
I I
I I
I I
I I
Figure 3.19 A breakdown case on communication lines.
The basic analog transmission modes are baseband signalling, amplitude

modulation, phase modulation and frequency modulation; the line signals for
each of them are shown in Figure 3.21. The baseband signalling is usually
applied for shorter distances. Signal levels and timing are specified in the
industrial standard RS232 or RS485. The baseband signalling is realized in
baseband modems. A binary rectangular baseband signal S(t) is given in Figure
3.2Ia:
S(t) = L ajg(t - iT) (3.12)

i=-oo
where aj is the amplitude level of the ith pulse in a pulse train, g(t) is the
common function of the pulse shapes and T is the interval between pulses in
seconds. The spectrum density (Gro) of the baseband signal is given as the
Fourier transform of g(t), that is
G(ro) = Jg(t) e-jOltdt (3.13)
the amplitude modulation is rarely applied in digital data transmission because

of its sensitivity to level changes during the trasmission. An amplitude-modulated
signal is given in Figure 3.2Ib. Another version of the amplitude modulation,
named 'vestigial sideband amplitude modulation', requiring the transmission
of one sideband and two pilot tones (for carrier reconstruction), is used when
high efficiency of communication is required. The amplitude-modulated line
signal is given the relation
v(t) = [s(t) + c] cos roct (3.14)
data terminals
~
T1O--i-+-,
I
,
data terminals
,.-i'----c T 1
T2 V--~--j 1----l--<lT2
communication
channel I
T n o--~,+--'
I
_I
I
\'--------
(a)
low-speed
comm unication
channels
N - 1
~
,channel
f (Hz) I channell "channel N
f (H ) guard bands
z / (channel 2 '\ channell
I I "
I I
I I
I I
:
I I
I I
I I
(b) channel / (c)

separation
Figure 3.20 (a) The use of multiplexers; (b) a frequency division multiplexing;
(c) a time division multiplexing.
The constant c is so adjusted that the quantity s(t) + c always remains positive
so that the simple envelope signal detection suffices to demodulate s(t).
The phase modulation is usually applied in the following form of the line
signal
v(t) = L g(t - iT) cos(wct + <I>j) (3.15)

i=-OCJ
where g(t) is the desired envelope shaping of the passband pulse, T is interval
between pulses in seconds and <I>j is phase of the ith pulse. In binary phase
[G(ro)]2
-~
~ro
,
1t/T
.. ro
Figure 3.21 (a) A binary rectangular baseband signal; (b) an amplitude-modulated

signal; (c) a binary phase-modulated signal; (d) a frequency shift keying signal;
(d) a digital data transmission signal.
modulation <l>j = 0 or n, and the line signal is as given in Figure 3.21c; in

four-phase signalling <l>j = 0, n12, n or 3n12. The detection of digital signal is
efficiently performed by using differentially coherent reception in which the
input is so coded that the phase sample <l>j_1 compared with the previous phase
sample <Pj gives the input information.
The frequency modulation is performed by the baseband signal s(t) control
of the frequency of a carrier signal, that is the line signal v(t) is given as
v(t) = Af COs[O>,t + d [ 1
six) dx (3.16)
device bus
I
I I
individual memory
I A/D and interface I microcomputer
- communication facility
- device memory
- sensor linearization
Il
- rerange
II - engineering units
I - diagnostics
Ocmpern<me
sensor + +
1\ I D/A
I
digital
communication conversion
:~Ci~
I
i
communication output
~ell sensor I
I
-+-- optional digital

~ ~~ , /
-
connec tion to remote
,,
I I
process -- interfa ce device

/
/
)
\
4-20 rnA signal and digital signal
to control system
Figure 3.22 A block scheme of the differential pressure 'smart transmitter'

Model 3051.
The change of frequency occurs in the line signal with the change of the
binary state (Figure 3.2Id); this is usually called a frequency shift keying
(FSK).
The Bell 202 standard frequency shift technique, which uses high frequency
binary signals superimposed on the instrumentation standard 4-20 rnA
transmitter output, is applied for The Rosemount Model 3051 differential
pressure 'smart transmitter'. The microcomputer controls the operation of the
transmitter and in addition performs the calculations for sensor characteristics
linearization, rearranging, engineering unit conversion, transmitter self-
diagnostics and binary communication. With the addition of a temperature
sensor for higher measurement accuracy it achieves an accuracy of ± 0.1 % of
calibrated span and stability of 0.1 % of upper range for six months. A block
scheme of this transmitter type is given in Figure 3.22.
The digital data transmission can be so arranged that binary O's are trans-
mitted as 0 volts and binary I's are alternatively transmitted as positive and
negative pulses (Figure 3.2Ie). Thus, signals can be easily regenerated as
opposed to analog data transmission where they are to be amplified.
Modem technology is based on its task to form and acquire an analog
signal to transfer the data under the constraints of the communication facility.
Practical suggestions and recommendations 121
source modulator .--------.'-----__ J modulator sink
sink ,--.,.~-"demod ula tor demodulator source
data terminal modem modem data terminal

communication
line
Figure 3.23 The connection of modems to a communication line.
These constraints can include signal level, limited bandwidth, noise level and
nonlinear characteristics. The connection of modems to the communication
line and to data terminals is given in Figure 3.23. An example of essential
parts of a frequency modulated modern is given in Figure 3.24. The low pass
filter removes the high frequency components of the rectangular pulse-shaped
data input. A binary' l' causes the mark frequency to be sent and a binary '0'
the space frequency. The band pass filter limits the frequency to the bandwidth
of the transmission facility. The receiver band pass filter removes noise and
interference outside the signal transmission band. The limiter sizes and shapes
the input signal to the frequency demodulator circuit where the original signal
is then recovered in a discrimination process of the demodulator. The low
pass filter removes unwanted high frequency and noise components and then
the output signal of the low pass filter is fed into a slicer performing the
slicing function to again produce rectangular pulses.
The term 'data communication facilities' refers to the transmission lines
using any technology (e.g. cables, radio-connection, satellite-connection,
microwave link) and being capable of carrying data signals.
Communication hardware for distributed process control is organized as
part of a packet switching network (Figure 3.25) either on bit-oriented data
link procedures like HDLC or SDLC protocols (Figure 3.27) or on coaxial
cables like PROCONTROL P12 as shown in Figure 3.26.
3.9 PRACTICAL SUGGESTIONS AND RECOMMENDATIONS
It has been recognized from everyday practice (Plecko, 1984) that radio
equipment is lacking in some aspects of process data transmission. For instance:
1. Squelch circuits for automatic correction of amplification at the input of
the low-frequency circuit of radio receivers have too long a time constant
(100 ms), which causes slow response of the receiver; therefore, for a
duplex mode of data transfer with transmitters always in work, squelch
circuits have to be put out of work; for a half-duplex synchronous mode
of data transfer with transmitters intermittent in work, the unfavourable
influence of squelch circuits has to be compensated by idle words added
transmit section receive section
___________________J\ \
r
frequency frequency low pass

low pass band pass line
I band pass
,modulate limiter demodul- slicer
filter Ioscillators filter filter filter
data ator ~_ _ _-II data
in out
r
nOise
Figure 3.24 Essential parts of a frequency-modulated modem.

RTU (DTE) RTU
packet
switching
network
RTU
RTU RTU
Figure 3.25 Distributed process control system organized as part of a packet

switching network: NC. mode computer; CC, centre computer; RTU, remote
terminal unit; DTE, data transmission equipment.
8-bit address 8-bit information word
parity bit
I I
00 0 1 1 00 II!
I
,,
I
+ 15v ,,
I
,
,
I.
...-----;'
::- 15v
:,
"
100001 110
Figure 3.26 Signal form in a coaxial cable for PROCONTROL P 12 serial bus 18.
I
control field bits
8 7 6 5 4 3 2 1
flag information
direction transfer
comma~f' response N(R) P/} N(S) 0
of
address (1 frame
transmission
control
876 5 432 1
supervisory
(:omma~fs/responses N(R) P/F SSO 1
information S frame
unnumbered
comman~~s/responses MMM P/F MMI 1
(U frame. f
FCS
flag
Figure 3.27 Basic structure of an HDLC data packet: flag, flag sequence
01111110; address, secondary station address field, 8 bits; control, control field of
8 bits; information, information field; FLS, frame checking sequence; N(S), send
sequence number; N(R) , receive sequence number; Sibits, specification of a super
visory function; Mlbits, specification of commands and responses; PIF bit, check
pointing bit (command = 1 ~ response = 1).
to the message; the duration of idle words has to be larger than the sum
of duration of the transmitter time (10 ms), to be put into 90% of power,
and the squelch time constant (100 ms).
2. When using communication equipment both for data transmission and
speech transmission squelch time constants of all repeators have to be
added to the time constants discussed above; therefore, extra idle words
of data transmission messages have to be added to decrease the data
transmission speed.
3. The use of a public telephone network on occupied lines can sometimes
cause additional inconveniences where induction telephones and improper
telephone equipment damage the modems.
4. The use of a public telephone network on leased lines can induce prob-
lems, since the declared quality of lines may not be reached and the
quality of lines is controlled only by the owner of these lines.
3.10 OPEN-ENDED COMMUNICATION IN PROCESS CONTROL

In order to enable efficient communication between process computers in
industry a standard MAP (Manufacturing Automation Protocol) has been
launched by General Motors and about 2000 companies around the world.
Communication of smart process devices 125
The bases for MAP are the standards of the OSI (Open System Interconnection)
model established by the International Standard Organisation (ISO).
The OSI model possesses seven conceptual layers covering all aspects of
communication. MAP standards on lower levels of OSI are relatively firmly
shaped, while some higher levels lack descriptions (Peter, 1989) (Table 3.18).
The physical level defines electrical connection between local cable networks
including frequency converters and modems in each mode. Thus a
communication on coaxial cable is defined according to the IEEE 802.4 Token
Bus standard. The main characteristic of such a connection is that only one
communication point can transmit a message which possesses a token (firm
defined bit combination - password). The token is passed to the next
communication point after the message has been sent. The communication
speed on such connection is up to 10 Mbit/s.
Data link level specifies protocols in the local network. These can be provided
either by a token bus or by a CSMA/CD (Carrier Sense Multiple Access with
Collision Detection) technique. A communication partner can only start
transmitting if there is no traffic on the bus. Should two partners try to talk at
the same time then each will soon become aware that the signal on the bus is
different from its own message. When this happens both partners break
connection and only re-attempt transmission after a short random time. This
leads to one device getting back to the bus before the other; provided that
some third partner has not gained access meanwhile. At network level the
transmission of data blocks between data sources and higher levels is
standardized.
- Transport level deals with communication aspects of remote terminal units
in the system.
- Session level standardizes the coordination of dialogs and synchronization
of application programs.
- Presentation level is foreseen for data interpretation and it is still not well
defined.
The highest level - application - is in final development. It should enable, for
instance, communication between the mainframe and each programmable
controller at the site (RS511 standard).
Figure 3.28 represents a block scheme of a fully automated MAP production
cell.
3.11 COMMUNICATION OF SMART PROCESS DEVICES
Using two out of the seven layers of the ISO reference model (Katz et al. 1989)
direct communication between 'smart' transmitters and valves and the process
computer becomes a reality (Blick1ey, 1990). The communication is standardized
under the name PROFIBUS DIN V19245 part 1 and 2, and is presented
schematically in Figure 3.29.
Table 3.]8 MAP specification and OS] model
Communication MAP protocols and Functions and services

layer and name realization
7. Application User Functions: system control, application

ISO-CASE control
ISO File Transfer
Series:
MAT Messaging
identification of communication
MAP Directory Service
partner
MAP Network Managing
definition of availability
authorization
validity check
Software realization
agreement on available resources
acceptance of the services
synchronization of the application
responsibility for error elimination
selection of type of dialog
checking that data is complete
adherence to data syntax
6. Presentation Not existing yet Functions:
request to set up and perform a
session
Software realization data transfer
syntax coordination and conversion
coordination and conversion of
presentation profile
Services:
data-syntax conversion
data formatting
selection of syntax
selection of presentation profile
5. Session ISO Session Kernel Functions:
coordination of session and transport
connection
Software realization session connection flow control
accelerated data exchange
restoration of session connection
session layer management
release of session connection
Services:
setting up and releasing session
connection
standard and accelerated data transfer
dialog control
synchronization of session connection
reparation of irreparable errors
Table 3.18 (continued)

4. Transport ISO Transport Class Functions:

setting up transport connections
data transfer
Software realization release of transport connection
transport layer management
Services:
set up services
data transfer services
release of transport services
3. Network ISO CLNS Functions:
routing/alternative routes
network connections and its
Software realization multiplexing
segmentation and blocking
error recognition/correction
sequence and flow control
high speed data transfer
resetting of network connection
service selection
network layer management
Services:
network addresses and connections
identification of connection endpoints
transfer of network service data units
parameters for service quality
reporting on irreparable errors
high speed transfer of data
resetting
release of the network connection
2. Data link IEEE 802.2 LLC Functions:
IEEE 802.4 Token Bus set up/release transmission link
ISO 8802.4 Token Bus framing and synchronization
error recognition and correction
Hardware (MAC) and identification and parameter exchange
software (LLC) monitoring the physical connection
realization data link layer management
Services:
transmission section
data link service data units
data link connection endpoint
identifier
error reporting
parameters of service quality

1. Physical IEEE 802.4 10 Mbit/s Functions:

Broadband/Carrier band activation/deactivation of the physical
connection
bit transfer
bit transfer layer management
Hardware realization
Services:
physical connections
physcial service data units
endpoints of physical connection
recognition of connection level
sequence control
error indication
parameters of service quality
Cell Order Order Order
Controller entry entry entry
Bull Nixdorf Olivetti Siemens
SPS7 1.-----11 Targon 35 M44 M70
802.3 Ethernet 10 Mbit/s
GEC Siemens
Router Router
802.4 Token Carrier band 5 Mbit/ s

I
GEM 80 Siemens NUM760

PLC 150U PLC CNC
AUTOMAX
Cutter Kit/
Machine
DekitRobot
Figure 3.28 A fully automated MAP production system/cell.

-
to process to process
device A device B
-
process device C
1 media access and control
2 field bus data link

l
MAP
layers
7a lower layer interface (LLI)
7b fieldbus message specification

J
7
"
( ~
source remote
object object
definition definition
device C device A
remote
object
)
definition
\
APPLICATION
LA\ER INTERFACE device B
APPLICATION PROCESS
control supervision
measurement
- - process --
Figure 3.29 Source and remote object definitions in PROF/BUS DIN V 19245.
REFERENCES
Arafeh, S. (1978) IEEE Trans. Autom. Control, AC-23, 2, 333.
BBC Brown Boveri (1982) Publication No. C-1-T lOa, 174ae.
Blickley, G.L. (1990) Control Eng., Jan., 74.
Karlsruhe, 11 Oct. to 17 Nov.
a) IAEA (1982) Course Material on FRAMATOME SPIN System.
b) Mitsubishi PWR Control System.
c) BWR Japan Control System.
References 131
Honeywell (1988) TDC 3000 - Prospect Material.

INTEL (1983) Component catalogue.
IPA (1983) Report on Process Control Systems. Bucharest, Roumania.
Katz, M., Biwer, G. and Bender, K. (1989) Automatisierungstechnische Praxis, 31, 12,
588.
Noyce, R. (1977) Scientific American, 237, 3, 63.
Peter, K. (1989) Siemens Components, XXIV, No.4, 135.
Plecko, V. (1984) Private communication.
Warren, C. (1983) Electronic Design, March 17, 101.
Part Two
The Role of Software in
Process Control Systems
4
The relative roles of software
and hardware
4.1 INTRODUCTION
Powerful central processor units (CPUs) and slow connections to the monitored
process enable process control functions to be executed using software, with
all the advantages and elegance of software problem-solving abilities. Therefore,
software takes a prominent role in process control systems, but, unlike hardware
where interfaces are standardized, only a few operating system software packages
are standardized predominantly as de facto standards like DOS 3.0 or UNIX
System V (Mikovic, 1983). Operating system software is an essential part of
process control software (Figure 4.1). Peripheral device drivers and handlers
are not usually standardized. They have to be tailored to the operating system
since even the most common physical connections like the RS232C interface
are by no means plug-in-and-go features on a universal basis.
Process control software is developed by operating system programmers,
system programmers and user programmers (see Figure 4.1). Some important
software parts have already been standardized and developed by computer
vendors. For example:
software tools;
compilers and interpreters;
- assemblers;
- linker-loaders.
The development and implementation of all the other software parts shown in
Figure 4.1 is discussed mostly in Part Two.
A process control task is implemented by an abstract automaton. For instance,
a logic or mathematical task may be either implemented in hardware or software
as there are no differences in system functioning. However, there are differences
between the software and hardware approach to implementation of tasks for
process control systems. Typically hardware functions are executed in parallel,
while software functions are performed sequentially. The basic software
components are recognized and compared to the equivalent hardware com-
ponents. This comparison can be made in terms of, for example, 'process
control safety', 'power consumption', 'space requirements' 'cost requirements' .
135
136 The relative roles of software and hardware
external
hardware system
software
border
special user
devices programs
~~~~~(-+_ user
i!!! interface
local-area data base
network
display ~§§=t.•_ system programmer
keyboard interface
RS - 232
process
in terface operating system
programmer
interface
interrupt RAM bootstrap
Figure 4.1 Hardware. software and user/programmer interfaces in software

development (Schindler. 1983): ~. common software tools.
The correct choice of software or hardware function in a process control

system results in a synergism which will multiply both function abilities. The
easier solution to this dichotomy points to the fact that all parallel data processing
functions, with the response time faster than one to five milliseconds, should
be executed in hardware. All other functions may be executed serially in
software. The most cost-effective system design may be achieved when the
same, or nearly the same, software is applied to several installations.
4.2 DA T A PROCESSING
Data processing is performed on the following.

1. Hardware input/output units and subsystems by analog and digital electronic,
optoelectronic and optical signal processing.
2. Computer hardware by digital signal processing based on programs.
3. Computer peripheral input/output units by analog and digital electrical
signal processing.
Data processing algorithms are usually described by outer or inner functional
specifications, which are based on the behaviour of the equipment as viewed
by the user or designer.
Data processing 137
Outer functions of data processing are usually defined in broad tenns in the
preliminary design of the data processing system, and in a more detailed form
in the main project itself. Outer functions are implemented by inner functions.
Inner functions of data processing are the detailed functions of data processing
hardware, i.e. firmware and software are the two main ways of implementing
them.
1. Firmware programs or program parts are implemented mainly on read-
only memories (ROM), programmable read-only memories (PROM) and
erasable programmable read-only memories (EPROM).
2. Software programs or program parts are implemented mainly on random
access memories (RAM).
A survey of different data processing functions implemented by different
hardware, finnware and software tools is given in Table 4.1, and a typical
assembly of a data processing function of a digital input signal is shown in
Figure 4.2, which also specifies the individual process functions, controller,
hardware input/output and computer-programmed functions.
4.2.1 HARDWARE
Hardware consisting of input/output units, input/output subsystems, computer

hardware and computer peripheral units is made of optoelectronic, electronic,
electrical, mechanical and optical components and devices (named parts) that,
designed together, possess certain infonnational characteristics which depend
on the particular parts contained within the design. The main electronic, electric,
mechanical and optoelectronic parts of process control systems and their
informational characteristics are given in Table 4.2. The definitions of the
informational characteristics are as follows:
Input time. The time delay between exciting the analog or digital input of a
device or unit and the formed information response ready for data transfer or
data processing at the device output.
Input rate. The maximum number of digital or analog data that enter
uninterrupted into a device in a given time unit.
Output time. The time delay between the input of an analog or digital command
to the device or unit, and the readiness of the unit to carry out the command
function at its output.
Output rate. The maximum number of digital or analog data that pass
uninterrupted through a device output in a given time unit.
Table 4.i implementation of different data processing functions
Data processing Hardware Firmware Software
Hardware: Computer
input/output hardware
units and units - central
subsystems part
Analog input
processing
grounding and optoelectronic

shielding galvanic
isolation
filtration and low-pass filter
pre-amplification pre-amplifier
scanning and analog input CPU and ROM PROM operating
multiplexing multiplexer memory input! system, driver
output channel handler, driver
NO conversion AID converter
scaling CPU and ROM PROM real-time
memory software
application
programs
Analog output
processing
scaling CPU and ROM PROM real-time

memory software
application
programs
D/A conversion D/A converter
demultiplexing demultiplexer CPU and ROM PROM operating

memory input! system driver,
output channel driver handler
smoothing sample-and-
hold circuits
amplification amplifier
Hardware: Computer
subsystems part
Digital input
processing
grounding and optoelectronic

shielding galvanic
isolation
filtration low-pass filter
scanning digital input CPU and ROM PROM operating

multiplexer memory input! system driver,
counting counting CPU and ROM PROM operating
circuits memory input/ system driver,
Digital output
processing
output scanning CPU and ROM PROM operating

memory input! system driver,
output timing timer, time real-time clock ROM PROM operating

delay CPU and system driver,
memory driver handler
output reed relay/
amplification relay, contactor
Hardware: Computer
subsystems part
Printing, listing,
key-in
printing serial printer CPU and ROM PROM operating

line printer memory I/O system
channel application
program
listing serial printer CPU and ROM PROM operating
line printer memory I/O system
channel application
program
keying-in serial printer CPU and ROM PROM operating
video terminal memory I/O system
channel application
program
Displaying
character alphanumeric CPU and ROM PROM operating

display memory I/O system
channel application
program
function display semigraphic CPU and ROM PROM operating
display memory I/O system
channel application
program
graphic display graphic display CPU and ROM PROM operating
memory I/O system
channel application
program
Data processing 141
Hardware: Computer
subsystems part
Other functions
voice pattern voice pattern CPU and ROM PROM operating

recognition recognizer memory I/O system
channel application
program
plotting plotter CPU and operating
ROM PROM
memory I/O system
channel application
program
process variable CPU and application
calculation memory program
database
software
Multiplexer or demultiplexer frequency. The maximum number of commutated

multiplexer or demultiplexer channels in a given time unit.
Sampling frequency. The number of signal samples taken in an information

channel in a given time unit.
command digital
issuing output
driver I
I
!electromagnetic
/' I ~ ;/-.)'·1 digital message : valve
input I
-if 0 :=::=:r --- analyser I
driver I
I
I
message printer
print-out driver
I
, I
I
I I
process: controller digital operating application program operatin~ printer : process
signal I display input system system : command
I
subsystem I
process controller hardware firmware software function firmware hardware process

function i function I/O unit function function I/O unit function
function function
I
,- computer function
•
Figure 4.2 A typical example of digital input signal processing.
Data processing 143
4.2.2 COMPUTING POWER
The computing power of a processor largely depends on the type of computer

used, its speed, instruction repertory, the skill of the programmers and the
characteristics of the operating system used. Basically the programmed data
processing functions enable the design of a process control system of smaller
physical dimensions at the expense of the speed of data processing due to the
serial nature of computer arithmetic and logic operations. The main
characteristics which influence the computing power of the processor are cycle
time, word length and instruction repertory and addressing modes.
Cycle time is the execution time of basic instructions in a computer.
Microcomputer execution time is usually longer than the memory cycle. Bigger
computers have a cycle time equal to, or less than, the memory cycle. The
cycle time depends on semiconductor technology and on the number of parallel
and serial operations in the central processing part of a computer. At present,
basic instructions for most microcomputers and minicomputers are executed
serially. The data on cycle time for different microprocessor technologies are
shown in Figure 4.3, as compared to the operational speed of different hardware
components of the control system.
Word length is connected to the computing power of the processor in many
ways, and only one word can be processed during a memory cycle. Therefore,
word length determines the number of bits processed. The comparative data
on approximated computing power for a bit, byte and word processing of 4,
8, 16, 24 and 32 bit/word computers are shown in Figure 4.4.
Instruction repertory and addressing modes are directly related to the word
length, since long words enable the definition of more diverse computer
operations. Eight bit microcomputers can encode 8, 16 or 24 bit instructions
in 1, 2 or 3 memory cycles. A more powerful instruction repertory reduces
program length'and execution time, thus increasing the computing power.
The computing power of microcomputer or minicomputer installations depends
on six mutually exclusive and independent parameters (Evans, 1982):
ability of data communication, byte/s, through process input/output hardware,
videoterminals, displays, printers and communication lines;
Table 4.2 Characteristics of some hardware components, circuits and devices
Hardware characteristic Informational characteristics

parts and their symbols
Hardware input! Symbol Input time Input rate Output time Output rate Multiplexer! Sampling
output units and demultiplexer frequency
subsystems frequency
Optoelectronic < 0.5 ms < 1 kHz 0.5-1 ms < 1 kHz

coupler tif~~
Low-pass filter > 20 ~s < 100 kHz 1-20 ms < 500 Hz
i~~ ~
Pre-amplifier 1 ~s-l ms 2 kHz-2 MHz
1 0
(A-gain factor) .f>.
Analog input in~ MUX put < 100 kHz ~ 100 kHz
multiplexer
Analog/digital < 50 kHz 1 kHz-100 kHz

converter i13
Digital/analog < 100 kHz
converter i~
0
Demultiplexer < 100 kHz ~ 100 kHz
i+EMU~
Output filter < 100 ms < 500 Hz 1-100 ms < 500 Hz
i@
~ c0
Output amplifier i-G>-0 < 10 ms 10 Ils-lO ms 50 Hz-50 kHz

(A-gain factor)
Counter inl in/N out < 5 Mimp/s < 5IN Mimp/s

(N-division factor)
·8
Time delay . I6 out in(t) in(t-6t)
In
{~J 6t = 1 ms-100 s
speed of component
__e_oTf__m_i_c_ro~p~ro_c_e_s_s_o,f____________,-____________.-_
IOOOs r-0_r_c~y_c_l_e_t_im
ls r-----------+-----------+-f
1msr-__________-+___________
l~sr-________~------_+.
Ins~~--------~~--------~-L~----------~----~p-e-r-c-o-m~p~o-n-e-n.t
InW ImW lW lkW
Figure 4.3 Microprocessor cycle time and speed of control system components
(American National Standards Committee, 1978).
Software for process control data processing 147
3 relative
computing
power
suppositions b:. 200%
2 - equal cycle time
- equal technologies
- equal instructions repertory b:. 150%
1 0 @] ~100% @] @]100%
50% 050% b:.
25%b:. computer
~--~---+--------~--------+-------~---word
4 8 16 24 32 length (bit/word)
o = bit processing
o = byte processing
A = 16 bit word processing
Figure 4.4 Approximated relative computer power (from (1977) Electronic Design,
21, 11,54).
capacity of data storage and data retrieval media, byte, such as ROM and
RAM, external memory (e.g. floppy disk and disk memory unit);
computing power of the processor, instruction/s, expressed as the number
of instructions actually executed in a given time unit;
speed of program exchange, program/s, that describes the ability of a
computer installation to change the program already in operation;
program efficiency (expressed as instr./byte x instr./error x progr./instr.)
that describes the number of bytes in an average used instruction, multiplied
by the average number of instructions referred to a program error and
multiplied by the average number of instructions in a program;
program span expressed as the number of different programs that are
available in the computer installation.
Tables 4.3 and 4.4 give basic data on computing power for three microcomputer
and minicomputer installations for process control, and it will readily be seen
that, despite the different hardware, all come out about equal when applied to
process control.
4.3 SOFTWARE FOR PROCESS CONTROL DATA PROCESSING
The software for data processing in the process control system comprises a
set of programs, procedures, rules and documentation. The extent of software
differs largely for each particular functional unit, which depends on its
hierarchical level, its role in process control and its connection with other
Table 4.3 Computing power data of different process computer systems
Microcomputers Minicomputers Mainframes
Data word size 4,8, 16 12, 16, 18, 24, 32 24, 32, 48, 64
(bit/word)
Instruction word 8, 16,24 16, 32 24,32,48
size (bit/instruction)
Add time register 0, 1-1000 0.OS-2 0.002-0.3
to register (micro- 1 0.1 O.OS
seconds per data
word)
Number of 10-unlimited 70-unlimited 2S0-unlimited
instructions 70 150
Hardware BCD No Yes No Yes Yes
arithmetic
Maximum input/ 30K-4M lOOK-ISM 3M-30M
output rate (words/ 50K-IM 300K-2M 10M
second)
Direct memory Standard optional Standard Standard
access
Numbers printed in italic type are the most common values or ranges
Table 4.4 Computing power of three process control installations*
Feature Process 8 bit RTU 16 bit

controller microcomputer dispatching
minicomputer
fl - ability of data communication 7600/3.88 11 SOO/4.06 11000/4.04

byte/s, log (byte/s)
f2 - data storage and retrieval byte, 2000/3.30 40000/4.60 22000000/7.34
log byte
f3 - data processing speed instr/s, 1 000 000/6.00 SOO 000/S.70 1 000 000/6.00
log instr/s
f4 - speed of program exchange 3600/3.56 60/1.78 1/0
progr/s, log progr/s
fS - program efficiency instr!byte x 20/1.30 20/1.30 33/1.S2
instr/error x progr/instr, log fS
f6 - program span number of 10/1.00 30/1.48 300/2.48
different programs/log f6
Total feature - ~ log fi 19.04 18.92 21.38
* Total feature for a personal computer is approximately 20

Software for process control data processing 149
functional units. Figure 4.5 shows the main software functions of most advanced
The executive part of the real-time operating system coordinates the work
of all the software. Requests for activation of a particular program are the
result of a process event or time-programmed function. The operating system
assigns the permission for program execution on the basis of the time of
request occurrence and on the requested program priority. The other parts of
the operating system are software modules termed 'drivers' which enable the
connection of the computer process and a peripheral unit to the central processor
unit. Each particular peripheral unit, or a group of identical units, usually has
a specific driver which is part of the operating system.
4.3.1 PROCESS SOFTWARE
Process software enables all process data processing and operator-system

communication. Process data processing comprises process data acquisition,
process database refreshment, process data analysis concerning alarm states,
alarm issuing for alarm lists and data storage in historical data tables. The
operator-system communication comprises the software that enables the operator
to work with console, VDU and functional keyboards. Process software provides
the operator with a display of process data and facilitates the acceptance of
process commands from the operator.
4.3.2 INTERCOMPUTER COMMUNICATION SOFTWARE
This ensures the connection of hierarchically organized functional systems

into an organized entity that enables efficient exchange of data between
functional units. Data are transmitted in data blocks of limited length, and
each data block contains a block header, data field and end-of-block data. The
block header contains all data on data source, data destination and data
organization. A data field contains process data, messages or control data.
End-of-block contains the data that enable data block protection from
transmission errors.
4.3.3 MESSAGE-SWITCHING SOFTWARE
This software enables the connection of a message-switching function and

process software function into an entity. These two functions use the same
computer resources, and therefore increase its efficiency. The message-switching
software contains the features for acceptance of messages from message-
switching terminals, acceptance of local program messages, analysis of accepted
process units other computer
peripheral units systems
r-------- ------- -------- -- -------1

1 I
1 Ir I
1 I
1 peripheral :
: units and communication I
I process driver
1
I unit drivers
I
1
j
1
L---..
1
executive
:~------------~ operatmg
I
I
I
1
- part system
I
---------1
process intercomputer
software • communication
software
message
switching
software
expert
system
software
extended
real-time
software
-
Figure 4.5 Main software functions of process control systems.
Software versus hardware 151
messages, direction of messages and communication of messages to message-

switching terminals, to other computer functional systems in the control system
or to local computer programs.
4.3.4 KNOWLEDGE ENGINEERING AND EXPERT SYSTEM SOFTWARE
This enables pattern recognition on a plant simulation database and intelligent

display system (Bailey, 1988) that: (1) displays malformations and suggests
control action to resolve the problem; (2) triggers plant control takeover under
threat of catastrophe; and (3) detects and identifies the nature and gravity of
incipient hardware or software malfunction. Engineers are also enabled to
make calculations for process design, construction and technology using different
workstations connected to the system.
4.3.5 EXTENDED REAL-TIME SOFTWARE
Extended real-time software enables efficient simulation of various situations

based on process modelling, prognosis of material, energy and information
flow balance, change and verification of the process model, automatic follow-
up of set-point values (according to given prognostics) and automatic return
from the on-line to the off-line working mode (Bodlovic, 1977).
Operating systems and process software are the most important software
components of process control systems and are discussed further in Chapters
5 and 6.
4.4 SOFTW ARE VERSUS HARDWARE
Logic and arithmetic operations can be defined as a progression of memory

bit patterns. This implies the possibility of designing a universal logic system
- a microcomputer that can execute any kind of logic and arithmetic task.
Compared with other logic and arithmetic components, the microcomputer
possesses strong universal advantages, but it lacks lower execution speed and
is more expensive than usual hardware components for logic and arithmetic
operations. Compared with bigger computers, the microcomputer lacks larger
memory capacities, but it is much cheaper. The block scheme of a typical
universal logic system is given in Figure 4.6. The large scale integrated
programmed processor either reads or writes one word of the memory at each
clock pulse, or performs logic, arithmetic and input/output operations with
these data according to previously read instruction tlata from the memory.
The hardware of a universal logic system is designed to perform certain
elementary operations, as directed by the same operations that progress and
strob: ata} output

~::::~:::::~ and interface
address
Figure 4.6 Typical universal logic system.
change the addresses of memory work locations. Such an elementary operation

is called an instruction, and a sequence of instructions a program. The
programs are written for a given application and stored in the memory of the
universal logic system. The required instructions are stored and executed in a
sequence changing the content of the program counter of the universal logic
system. Program branching is executed by putting the new address into the
program counter. The conditional program branch is executed after the given
branch condition is fulfilled. Branch conditions can be arithmetic when, for
example, two location or register contents are equal (or less than or bigger
than) or logic (e.g. when the content of a given location is equal to logic 1,
or 0).
4.4.1 SERIAL OR PARALLEL ALGORITHMS
Process applications deal with data processing of a great number of process

points which change in parallel over time. Two extreme solutions are still
feasible:
1. single processor architecture where process points are transmitted in parallel
by the hardware and then serially processed by a kind of polling algorithm
in a common program loop;
Software versus hardware 153
coaxial Ethernet
~--.--~--- -- -. --
( )
expert
system
Dl II~ designer's
-
station graphic
• • •
II ~
I station
~ c:~ '-..
II ~
r-- / ~L;---·-7
special
fileserver = I!i IIlIII •

111111111
I purpose
multiprocessor
r------r
LJ ~
L!I!I_~~!lj /
fibre optic
Ethernet
I
/ / ._-
process I.- ! -H-
-IT---- remoteI/O
=
..
station '------.------i
=
I
• • • • OLe connection
I
= ~- .. IY
-_ drive system
I i -----.-
Figure 4.7 Distributed control system with communication and expert system
support.
2. parallel processor architecture where all process points are fed to the
common parallel processor and processed in parallel; in such a way critical
process states can be detected very quickly and fed to the other parts of
the system.
Practical solutions today use a third approach, namely intelligent input/output
modules where a type of multiprocessor architecture is applied but without
the lateral redundancy offered by a parallel processor architecture. The problem
with parallel processor architecture is the lack of operating systems (Brajak,
1990) and supporting software tools. Pending a standard practical solution, a
more parallel architecture for process control, supervising and fast expert system
environment is proposed as given in Figure 4.7 with the following components:
operating system UNIX System V, release 4, featuring standard industrial
communication based on ISO open system interconnection, transparent
file system, real-time capability and standard graphics, icon, motifs;
communication software environment that connects different stations ranging
from pes to mainframes enabling
Ethernet and X.25 as physical standard

TCP/IP as transport and network layer standard (see Chapter 3 for
ISO-MAP communication);
Network File System and Remote File System as transparent file systems
on a range of machines using UNIX;
X-Look, X-Windows for transparent connection and application execution
on different machines.
4.4.2 COMPARISON OF HARDWARD AND SOFTWARE
It is also possible to compare hardware and software engineering methods and

products (see Figure 4.8). The final product of a computer control system is
either a hardware component or a program instruction. The initial hardware
(denoted as H) and software actions (denoted as S) are the total parts of a
system's actions (T) and can be stated as
T = S + H = const (4.1)
and
(4.2)
j
where Y is the production cost, i is the interest rate, Pj is the jth product cost,
kj is the capital needed for jth product design and w is the unit cost of labour
(Horvat, 1987).
For n identical systems, the total initial action Tn is a function of hardware
H and software S actions, thus
Tn = nH + S = const (4.3)
The minimum marginal costs are obtained from equation (4.2) when
nH =S (4.4)
Synergism of hardware and software is reflected in the amplification of system
functions by using proper hardware/software trade-offs. Fast process actions
like status signal processing are performed by parallel hardware actions. Slow
process actions like a set-point issue are performed by software actions. A
rough calculation shows that for each status process input approximately 50
to 100 bytes of program are needed for filtration compared to the hardware
solution (Huelsman, 1970) given in Figure 4.9 that costs approximately $10.
Thus, synergistic solutions that involve both hardware and software can be
cheaper than only hardware or, preferably, software solutions.
hardware system software system
circuit block scheme flow diagram
signal path on program path on

a block scheme a flow diagram
electronic module program module segment
electrical scheme detailed flow diagram
logic component program instruction
flow chart boolean

AND structure
yes
OR circuit f = f1 + f2 OR structure
A
B
C
D-
f E-
1 F-
G-
H-
A-
B-
C-
f2 Q -
E-
F-
e
H-
2.25 IC packages = 80 ROM bits
Figure 4.8 A comparison of hardware and software methods and products.

1 fJ.f
input 56kf2 output
operational
I
amplifier
O.39fJ.F 47NFI
Figure 4.9 Active filter-hardware solution for status input filtering.
REFERENCES
American National Standards Committee X3 (1978 ) Computers and information
processing, organization and procedures, 13/SD-2 CBEMA Washington, DC.
Bailey, S.J. (1988) Control Eng., July, 75.
Bodlovic, P. (1977) Koncar Information Globus Zagreb 24, I, 9.
Brajak, P. (1990) Personal communication.
Evans, C. (1983) The Mighty Micro. Publ Gollancz, London.
Horvat, B. (1987) Working theory of costs, Rad, Beograd (in Croatian).
Huelsman, A. (1970) Active Filters, McGraw Hill, New York.
Milovic, M. (1983) Personal communication.
Schindler Electronic Design (1983), March 17, 117.
5
System software
5.1 INTRODUCTION
The principle underlying the provision of system software is that it should

provide all those facilities which are likely to be common between applications.
In this way the user has only to write the application-dependent part.
System software includes the operating system, on-line and off-line software
support; it enables the development, implementation, operation and changes
of the software for data processing in process control application.
The move to provide an operating system itself imposes a way of doing
things on the programmer. In aiming to meet the requirements of different
uses the features included are inevitably more than are needed on one particular
application.
From this it follows that the operating system's function is to create an
abstract automaton that specifies data structures and operations with the level
of abstraction appropriate to the user. It also satisfies the automaton requirements
for user-efficient execution of data processing tasks. The subsystems (e.g.
database software, computer support or language interpreter) can overlap and
extend the operating system to make it much more powerful at running the
system's software.
The operating system is the key to process control since it does the following:
1. It acts as an intermediary between the (micro) computer and the high
level language program.
2. It manages the functions of the (micro) computer, e.g. I/O functions,
memory usage, control processor(s) use.
3. It provides the programmer with an idea of how the computer works, thus
making it 'user friendly'.
4. It provides a process for appropriate start-up and restart functions including
software for the recovery of errors in hardware and software, thus making
the computer 'process friendly'.
157
158 System software
programmed
tasks
•
operating
system
/ ~
programmed input/output
tasks units
system: ~ ~, two-way communication; --+,

Figure 5.1 The assigning of computer resources by the operating
one-way communication.
5.2 BASIC CONCEPTS OF REAL-TIME OPERATING SYSTEMS
Real-time operating systems are geared for large amounts of input/output

activity, usually involved in updating databases and executing process control
operations. There are two principal computing resources requiring data
processing: input and output units and programmed tasks (Sajfar-Tasic, 1981).
The problem of assigning the computing resources in a process control system
is solved by the unique managerial control of the operating system (see Figure
5.1).
The assignment of a programmed task is done indirectly through the operating
system in which the program requests its inclusion. The operating system puts
this request in a common waiting queue. The other common routine takes the
request out of the waiting queue and puts it into execution.
A program request routine forms a connecting request list based on the
priority of programs that request the processing on the first-in-first-out or
last-in-first-out principles.
A program starting routing takes requests out of the request queue
considering the situation in the interrupted program queue and includes the
first requested program for execution only if its priority is higher than that of
the program at the top of the interrupted program queue. This algorithm enables
the determinacy of the programming system executing first the interrupted
program from the interruption place, then the requested program is included
in the execution from the beginning with the new program input data.
The program that requests some computing resources is instantaneously
Basic concepts of real-time operating systems 159
interrupted. The request has the fonn of a programmed interrupt so that its
priority and the address of the continuation of the interrupted program are
automatically saved at the top of the interrupted program queue at the moment
of the execution of the request instruction. The table of interrupted programs
is usually of fixed length and it is filled bottom-up and emptied top-down,
thus enabling the inclusion of interrupted programs according to last-in-first-
out order, and saving the priority order of interrupted programs because the
last interrupted program has the highest priority in the queue of interrupted
programs.
When programmed tasks use some data input or output unit, they occupy
this unit by requesting it through the operating system that has routines for
input/output requests, for the extraction of the first request out of the queue
and for the omission of the processed request out of the queue. The request
for the input/output unit also contains a unit tag. A separate queue is formed
for each input/output unit by an acceptance routine. This queue is a list of
requests ordered according to the priority programs that request the input/
output unit and satisfies a first-in-first-out rule. Certain data should be given
to the operating system in order to accept input/output requests. These data
are:
tag of the requested input/output unit;
address of the working area for data including the direction of data transfer,
number of data or data volume and data code;
infonnation on execution of program after use of input/output unit that is
usually the relative address of the place where the program has to be
continued.
The program that continues after using the input/output unit is stopped until
the unit executes its task, thus disabling the change of the working area before
data are read in (transmitted) or written (accepted).
Figure 5.2 illustrates the routines for data input/output. The acceptance routine
of the input/output subsystem accepts all these requests and should this unit
be free the program proceeds to the program for input/output communication
and sets the integer variable of input/output occupation. Any input/output
communication program uses the routine for extraction of the first request out
of the queue at its beginning. Omitting a processed request from the queue is
done by the routine shown in Figure 5.2. The routines for input/output requests,
the extraction of the first request from the queue and the omission of the
processed request from the queue must not be interrupted until completed.
This feature is executed by these routines using the highest priority level.
An operating system contains the components that handle job management
functions (or executive functions), system management functions and data
management functions. Tables 5.1-5.3 outline the particular functions of an
operating system.
I/O unit request from programmed task
acceptance routine of
I/O request
(I/O RR)
no
I/O unit
I/O unit communication program
immediate free?r----------- - ------- - ------- --- - - -..,
return to I
I
programmed task I
I
I
I
outine for the extraction

of the first request out
direct starting of the queue
of program starting
routine
( monitor })oo...._-+---'y_e_s_ _<
request
processing
deletion of
processed
request out of
the ueue
yes no
call of program
program to be continued?
continued
I
I
I
I
1 execution of
I continued
I
I program
I (BPR)
I
i
1______________________________________ J
I
Figure 5.2 An example of the execution of I/O unit routines.

Basic concepts of real-time operating systems 161
Table 5.1 Operating system functions: job management functions
Global function Subfunction Division of subfunctions/programs
Executive function job control job scheduling

resource allocation
program loading
program termination
I/O control scheduling
data transfer
remote terminal support
system system start-up
communication job control communication
I/O stream control
resource status modification
system status interrogation
recovery processing checkpointing/restart
automatic restart routine
restarting
System resource main storage fixed foreground/background
management allocation areas
fixed independent partitions of
various sizes
free internal storage as a large
pool of storage
paging of main storage into small
fixed pages
virtual memory by expanding
storage to external memories
I/O devices preemptive occupation
allocation nonpreemptive occupation
dynamic assignment of I/O
devices
CPU time priority job allocation
allocation time slicing and partition
assignment
information files shared files and routines
reentrant routines
Operating systems for process control can be divided into time critical
operating systems and time uncritical operating systems.
The time critical operating systems are met in some process control systems
on the lowest hierarchy level.
Figure 5.3 illustrates how most systems specify typical time response ranges
(shown by a horizontal bar) and shows that the mean response times peak at
approximately 150 J.ls (actual data in Table 5.4).
162 System software
Table 5.2 Operating system functions: system management functions
Global functions Subfunction Division of subfunctions/programs
System generation error diagnostics error control routines processor errors

and processing main storage
support errors
I/O data errors
power failures
timing services interval timing service
real-time service
testing services independent program testing
resident program testing
debugging independent debugging
services resident program debugging
System maintenance updating of
operating system
changing operating dynamic changes of operating
system programs system programs
suspension of all processing
while changing operating
system program
Program maintenance updating the deactivation of program
program change of program parameters
activation of program
changing the dynamic changes on deactivated program
program testing deactivated program
activation of changed program
Compiler interfaces translator and
compiler usage
translator and
compiler exclusion
5.3 STRUCTURE AND FUNCTIONS OF REAL-TIME OPERATING

SYSTEMS
A real-time operating system can be presented as shown in Figure 5.4 where

the operating system uses data from tables and communicates with the user's
programs that use the data from the database. Data on every program are in
system tables, thus defining a programming system to the operating system
by activation or deactivation of certain programs or modules. The common
data of user's programs and the process data are in the system database.
Input/output drivers are coordinated by the operating system feeding data to/
from input/output devices.
Table 5.3 Operating system functions: data management functions
Global functions Subfunctions Division of subfunctionslprograms
File management file support for system on-line file location

functions files off-line file location
permission of user file access
restriction of user file access
backup file service
file restoration service
I/O support facilities individual data access global read macroinstructions
individual data global write macroinstructions
processing single buffering
data transfer double buffering
and buffering buffer pooling
dynamic buffering
Data management retrieving selected file creation
facilities portions of the file new file creation
displaying selected file merging and reorganization
portions of the file data selection
data computation
reporting
data management program
percentage of system
%
150J,ls
I ... .....
50
, -~
/ I ,
, I "
40 , I ,
, I ,
, " II \\
, I \
30
I : \
, ~I \
I I \
20 / ! \
, I \
/ I \
10 //
/ " .....
.,.'
"
5 10 20 50 100 200 300 1000
response time (J,I s)

Figure 5.3 Interrupt response of real-time operating systems
from Table 5.4 (distribution function).
Table 5.4 Mean response times for various systems
Company System name Supported Size (kbytes)* Speed (Ils) Supported Comments
processors languagest
Advanced OS Engine Z8002 8/16 40 to 400 None PROM-based kernel of MSP/

Micro Devices, Z8002
Santa Clara,
Calif.
2 Hemenway MSP 68000, Z8002 8/24 Not available B, P, Forth Has provisions for
Corp Boston, multiprocessing
Mass.
3 Hew lett-Packard RTE-A A600 30/50 55 to 140 B,F,P Languages, database,
Cupertino, Calif. networking optional
4 Industrial MOTOS/68-KF 68000 -/16 20 to 500 None Up to 2048 tasks; 68-KF is
Programming firmware version of 68
Inc.,
Jericho, N.Y.
5 Intel Corp., iRMX 286K iAPX 86.88 0.5/13 30 B, C, F, P, Cobol, Supports MMX 800 and
Hiltsboro, Oreg. iAPX 286 PL/M Ethernet; ROMabie nucleus
6 Intelligent SMT Plus 8086 LSI-II 4/20 >10 None Response time depends on
Industrial 680968000 configuration
Systems,
Secaucus, N.J.
7 Motorola Inc., VERSAdos 68000 12/60 200 F, P Modular construction; fast
Tempe, Ariz. floating-point package optional
8 Systems and REX-80 8086/8087, 4/- 100 to 350 C,P, PL/M Configurable in firmware with
Software, 8088 run-time binding
Downers Grove,
II.
* The first number gives the size of the kernel, the second the minimal size of the operating system
t B = Basic C = C language F = Fortran P = Pascal
Structure and functions of real-time operating systems 165
system user's database

tables programs
I/O system I/O process

devices devices
exchange; --+,
Figure 5.4 Global structure of real-time operating system.
program exchange.
~, data
waiting and serving

request queue
user's driver
program 1 1
user's driver
program N M
monitor interruptI
waiting
list
Figure 5.5 A real-time operating system: ....... , program communication; C=::>'

data communication.
A structural scheme of a microprocessor real-time operating system is shown

in Figure 5.5.
The input/output communication with external devices is organized by using
the interrupts initiated by external devices. The operating system disables
simultaneous acceptance of requests from several user's programs for the use
of the same input/output device. Other possible conflict situations are avoided
through the organization of software which activates only one program that
166 System software
requests the access to a certain input/output device. The communication is

usually executed under the control of the central processor unit in a single
processor installation.
The interrupts from external devices (peripheral devices) and process input/
output devices are first served on the basis of hardware priorities. A number
of hardware priority levels can be identified, usually not less than three.
However, for a simple, single processor installation, two levels are sufficient.
The interrupts from the devices raise an interrupt in the system and are serviced
by the activation of the corresponding interrupt service routine. All other
interrupts, even those of higher priority, are disabled during the short interval
of interrupt processing.
The highest interrupt priority levels are used for the system control functions,
e.g. power failure, time-out failure and system start/restart functions. Lower
interrupt priority levels are reserved for input/output devices, e.g. video terminals,
real-time clocks, process devices and system clocks. The interrupt routine
begins to process software priority levels after the acceptance of the system
clock interrupt.
The software priority levels available are usually not less than three. Data
acquisition programs, control algorithm programs and diagnostic programs
are executed at higher levels of software priority. Off-line data processing and
system programs for the communication of input/output devices are executed
at lower priority levels.
The execution of programs at software priority levels is interrupted by
system clock interrupts starting with the highest level of software priority (see
Figure 5.6). The execution of the program at the interrupted level (see level
5 in Figure 5.6) is performed from the interrupted point to the end of that
program and afterwards again for all the programs of the same priority level.
The execution of the program then proceeds to the next lower level.
The most critical parameter in real-time data processing is the time response
to process interrupts, and to interrupts from input/output devices. The time
interval attached to particular programs that are executed, or wait to be executed
at any of the software priority levels, is variable and depends on the organization
of internal program and external data processing demands caused by the process
interrupts. The programs at the lowest software priority level are executed
only when there is no activity at other software priority levels. For instance,
these programs are designed for background program activities.
Minicomputer and microcomputer operating systems have been developed
with more hardware priority levels and stronger software support. Figure 5.7
shows a simplified structural scheme of a minicomputer operating system.
The highest priority user program demands processing by issuing a TRAP
instruction. This is processed by the TRAP service routines. Input/output unit
drivers generate interrupts at a higher level than interrupts in TRAP processing,
and form an input/output servicing queue that is first served. After servicing
all input/output interrupts the program returns control via the TRAP processing
hardware highe r
in te rrupts priori ty
routines
ha.rdware
priority
lower levels
h ardware
priority 2
in te rrupts
rQutines
system :
clock to tl
3
~
inte rrupt
t2
4
I
:
software
priority
:::::: ~ t31 levels
5
I
- tI~~}~
:~t;:::::;
6
priority levels
J
3
4
to t1
t
sy stem cloc k interrupt
Figure 5.6 Program execution at software priority levels.

168 System software
return to interrupt (RTI - trap generation)

~-- ....... "
,
500lls \
\
\
I
i
I
I
start I
continue I
110 I
processing:
I/O processing I
via exclusive I
resources I
,
I
I
I
I
store 110 I
I/O queue I
.
I
I
I
I
I
r-------::;;o..r"I/O
queue
I/O processing empty yes
via common
resource~
Figure 5.7 Structural scheme of a minicomputer operating system (RSX 11 MIV.3

model from Digital Equipment Corporation applied on PDP 11 computer series).
routines. The sequence of saving and subsequently restoring all the contents
of processor registers is initiated after each change of processing control.
Data flow activities exist that are not processor controlled, e.g. data transfer
between memory and disk or floppy disk units. These transfers are executed
at the highest priority by using hardware (cycle stealing).
The program activities at different program priority levels for a particular
real-time system are given in Figure 5.8, where the scheduler is triggered by
a timer interrupt to perform either the interval control routine at the priority
level 3 when there happens to exist a real-time data acquisition, or to proceed
to level 4 priority when there happens to exist off-line data processing. Having
timer interrupt
, / ,
I ,
sc heduler , program progra program
I
I
,I
~~. yes (real-time : yes (real-time data
,i I
I I data acquisition) : acquisition) ~---
-1 - - - c- yes (real-time -- -t-- - - ----- ------

, I
data acquisition) ,
,,I
,, I
da ta , , measure- I
interval ,
ac luisition:, ----. control ment data ,I
routine read-in ,I
pr, >gram I
,I ,, ~
.. .. ,,
I
I
I ~ -- -- ---
I
I I
,,
,
:-1-
,
; I main user's
ime control I I
other I I program and
Iof I/O unit I
:,
programs: I , L.... read-in data ~
I ommunication ,
,, ,I processing
I ! and dia,gn. prog
I I I
,, -~~-~-.
, ,. I ------+
,,
mterval r -_ _,;...-_ _ _ read-m
control time interval result

data ...... interval 1 counters and tables data
data buffer
interval 2 diagnostics
data
\ 1
print-out \
P'UH-UU< \~header queu~s of header print-out
queues queues _ _ _ channel 1 read-m res~ channel 1 buffer
chan:nel 2 channel 2 ,
------i---------~==~==='-l----------------~=====~:--------------------.:
time:
level 3 priority level 4 priority level 5 priority
Figure 5.8 Some program activities at different priority levels (Mikovic, 1980).
170 System software
timer interrupt timer interrupt
A B c D
scheduler and
system programs
II
II
interrupts and
DMA transfer
I data
acquisition I data
processing
/
~.x_
~
I
I
I
I
I
( \
I
I
I
I
I
I
I
I
I
I
I
basic interval
....
I
I
'....
I
+"
timer interrupts
--==-==::.--~-~--------------- +
01 I I c I A B
basic interval
D " A I I I I[ ~t
B c
basic interval
D
If---- -~----iI
Figure 5.9 A particular case of distribution of tasks in a basic interval as shown

in Figure 5.8.
perfonned the time control of input/output device communication and the

diagnostics program, the scheduler is again put on the task to read in meas-
urement data, or to proceed to level 5 priority. The main user's program data
capture is executed at level 5 priority. The scheduler automatically proceeds
with the background activities at priority level 6 should off-line data processing
occur.
The calculation of time critical parameters for the process control system is
based on the worst~case time situation, since there is always the possibility of
a program at a lower priority level being interrupted by a program of higher
priority. The read-in procedure for each channel has to be perfonned as defined
in Figure 5.8. Thus a set of task intervals perfonned by the processor in a
basic interval has to be less than the duration of the basic interval. A simple
distribution of tasks in a basic interval for the case given in Figure 5.8 is
described in Figure 5.9. The data acquisition interval is split up into time
intervals needed for data acquisition tasks for a particular channel. The constant
parts of data acquisition tasks and other program jobs are grouped into
characteristic program tasks executed in a different schedule controlled by the
distribution of program tasks to various priority levels. The time occupancy
Structure and functions of real-time operating systems 171
end
prepare
block
~~._______
. un_b_IO_Ck____~C
Figure 5.10 State diagram of processed tasks.
scheme of program tasks varies for different real-time process control systems.
The stochastic part of program execution tasks lies in the process interrupts
caused by status and count-pulse inputs. Status data are usually processed at
priority levels higher than measurement data and count pulse data because of
the need to associate time tags of higher accuracy to status data to present a
chronological event record, or to intervene faster in the process after receiving
emergency status data.
Usually a task can assume four distinct states:
processed state when it is being processed in the central processing unit;
suspended state when it is put into inactive state;
ready state when it waits for the assignment of the central processing
unit;
waiting state when it is blocked waiting for other tasks to perform (i.e.
make free a certain I/O unit).
The exchange of states is performed through operating system routines (Figure
5.10), and an example showing the execution of three tasks is given in Figure
5.11.
A multiprocessor configuration involves the communication of many single
processors through a global memory, i.e. a memory accessible to all the
processors. The communication is based on send and receive mechanisms to
or from this global common pool. Figure 5.12 presents the difference between
priority
TASK
CD ® @Q 0@ G)®®
~ ~
p p
W
a S
p p
b R
c R
Figure 5.11 Task priority control: 1, starting task a at initialization time; 2, task a
starting task b; 3, task a starting task c; 4, call of task a waiting for ending of
tasks band c; 5, call of task b waiting for ending of task c; 6, ending of task a
call; 7, ending task a; 8, ending task c; 9, ending task b. Task state: S, suspended;
P, processed; R, ready; W, waiting.
local
end (A) mailbox
J
r ceive (A)
Figure 5.12 Difference between local and global data traffic.

Data and symbols for the operating system 173
local and global data traffic. The synchronization of tasks can be realized
through a polling mechanism where a cyclic task is started in a goal system.
This cyclic task searches for messages in the global mailbox and takes
appropriate actions upon them. The interrupt method makes use of the interrupt
(OUT-command) to start the task in the goal system whenever a corresponding
message has been found in the global mailbox.
An operating system makes an overhead in data processing systems that
can be calculated as 20% of the time (Mikovic, 1980) for a single processor
microcomputer installation. The time response overhead is approximately
1 ms in data acceptance, compared with 100 f.ls in a minicomputer operating
system (see Figure 5.7).
When written in C critical components of the operating system can add
approximately 20% more time overhead and make program writing a lot more
comfortable (Rus, 1990).
Multiprocessor microcomputer systems show approximately the same time
overhead characteristics (Borevkovic and Cvitas, 1983) since the most critical
time response functions are done at the highest priority level by a single
microcomputer in a system.
5.4 DATA AND SYMBOLS FOR THE OPERATING SYSTEM
Operating system data are used to build up and enable a flexible and modular
exchange of system functioning when applied to different process control
functions. For example, a microcomputer control system has the following
sequence of programs and data:
1. System parameters
2. Operating system
3. User's routines
4. RAM memory programs and stack
A more detailed sequence is given in Figure 5.13.
System parameters are set by the statement EQU. The symbols that have to
be defined are:
XSTL - stack length, 80 bytes plus maximum stack length for each
priority level - an approximate total between 150 and 200
bytes;
XPNl, XPN2, XPN3, XPN4, XPN5, XPN6 - maximum number of
programs for each priority level;
XRTC - address of system clock;
XRTW address of bus control watch-dog;
XBAS basic time interval (ms);
174 System software
sequence of programs and data memory type and

symb 0 Is address
( operating system programs

)A
1K
programs for I/O units
2K
user's programs for control
algorithm realization and
calculation (macros and algorithms)
3K
( initialization routine )
4K
function generation tables
and constants
KrROM
J
1-8
( system stack RAM
( working area of operating

system programs J
( common locations
)
working area of other
programs .1- 9K
I----
ri---'
I
Figure 5.13 PROM and RAM usage for a microcomputer control system
(from M. Mikovoc, Internal Report, No. 1651, R. Koncar).
XRWM - RAM space detennined for system stack, system tables and
parameters; the calculation of space is given according to the
relation
6
XRWM = XSTL + 108 + 3 L XPN (5.1)
i=l
System parameters may be set up in the following form
XPNI EQU 2; 2 programs at priority 1 level
XPN2 EQU 2; 2 programs at priority 2 level
System software 175

XPN5 EQU 2·, 2 programs at priority 5 level
XRTC EQU 327Q; address of system clock
XRTW EQU XRTC; address of bus control watch-dog
XBAS EQU 6; 6 ms for basic time interval
XSTL EQU 200; 200 bytes for stack length
The structure, volume and content of the operating system data depend generally
on the following.
1. Operating system concept: a small operating system usually requires a
small amount of operating system data and enables fewer modifications
than the large system with greater amounts of changeable operating system
facilities and volume of operating system data.
2. Type of application field of real-time systems: larger operating systems
applied to special data processing do not require a proportional amount of
operating system data as a general purpose operating system.
3. Type of data processing organization such as serial, tree, parallel,
hierarchical, mixed: the more complex the data processing organization,
the larger the amount of operating system data applied.
Operating system data enable communication between different levels of
operating system programs, system programs and application programs. A
hierarchical real-time system is given in Figure 5.14. The only exception to
this concept lies in the communication link 6 where the user's program demands
a process input/output unit. The data required for some program communication
links are given in Table 5.5.
A certain number of operating system data also exist and are arranged in
system tables where definitions, equivalences and constants necessary for a
successful functioning are given. The system tables are usually the only software
module to be reassembled before the new control system is built.
The standard list of contents of system symbols, constants, variables and
tables for a minicomputer operating system with a mass memory facility is
given in Table 5.6 and the global structure of the corresponding operating
system is shown in Figure 5.15. Distributed data processing implies independent
data recovery in operating systems (Skeen and Stonebraker, 1983).
5.5 SYSTEM SOFTWARE
System software for process control systems includes real-time system software
for on-line and off-line application. The on-line system software covers operator-
system communication, debugging routines, utility routines (mathematical
176 System software
system user's
programs programs
---- - - ---- - - - - -----~----- ------

'-3
le-YdT--r------ -
t--l ,-2 Y-4
I
read write
analysis read write analysis
and and
read-in print-out
,I
l
I
r
------- ---------- ---------- ---------- ------t ' -
+--5
level 2
driver
6,
communication
control
------------ --------- ---- ---------------- - --

level I
+-7 ,-7
driver I ... driver N
--~j~- ~A-~-----~-,--
-8 -8 - 8 -8
---------
1 -8
-
external external external external external

... unit
unit unit unit unit
Figure 5.14 The structure of a hierarchical operating system (Cihlar, 1980).
routines and conversion and fonnatting routines) and system generation

programs.
Off-line system software includes higher language facilities (e.g. compilers,
language processors and interpreters) and editing facilities (e.g. editing, file
manipulation and creation and program linking).
Programs and routines for operator-system communication enable:
System software 177
Table 5.5 Data for program communication links
Type of Program description Data description

communication
7 Real-time clock driver External symbols:

labels of locations in other system
programs
addresses of real-time clock registers,
priority data
Internal symbols:
driver labels, internal routine labels and
subprograms, logic unit number
Constants:
starting time counter values, codes of the
real-time clock status word
5 System supervisory External symbols:
program and error location labels of other system programs
print-out and routines, logic unit number in the
operating system and priority data
Internal symbols:
program labels
labels of the working
memory area
Constants:
priority position place, error codes, masks,
error table locations
Intercomputer Error message data
communication routine Routine parameter list
Parameter list for the message error print-
out
2 Read-in of external Error message data
data into system Logic unit table
buffer Driver initiator table of the peripheral unit
Parameter list of error print-out
Parameter list for floppy disk read-in
Parameter list of the read routine
system start-up, data acquisition start-up and shutdown of data acquisition,

read-in of headers, read-in and read-out of the date and time;
memory change, memory load, external memory load delete and change
(for systems with external memory facilities);
data exchange between external input/output devices;
use of a real-time language processor;
other communication options such as system restart communication, program
activation and deactivation, memory dump activation and on-line debugging
program start-up.
178 System software
Table 5.6 Definitions, variables and tables for a minicomputer operating system
System globals Content of system globals
Definitions Global symbols

Core areas for mass memory resident programs
Program numbers
Logic unit numbers
Maximum number of logic units
Maximum number of diagnostic timers
Variables Top of link for I/O buffers
Top of link for scheduler and timer requests
Counters
Tables Interrupt vectors
Scheduler stack
Standard program starting address
Program status
Program execution priority
Core map partitions
Priority for partitions
Starting address for partitions
Partition numbers for mass memory programs
Address and size for mass memory programs and files
Buffer areas
Table of driver table addresses
Top of link for logic units
Diagnostic timer address table
Driver tables for physical devices
Debugging routines enable:

trace function, for example by setting check points in a program by writing
a TRAP instruction that prints out the contents of the processor's general
registers at the time TRAP occurs;
system report program that prints out the messages and initiates the
program to be debugged;
disk or floppy disk file handler, file compare and file maintenance;
dump programs that execute different types of dumping (e.g. fast memory
dump, mass memory dump, post-mortem dump program and cyclic output
of fast memory).
Utility routines enable:
mathematical routines for multiplication and division;
conversion and formatting routines for various conversions such as BCD
to binary, and binary to BCD conversions, octal to ASCII and ASCII and
ASCII to octal conversions, integer to ASCII and ASCII to integer
System software 179
110 unit
interrupts
r- --- ~-----------------1
feedback
r-+-
I scheduler
1
I/O unit to user's
1
1 drivers f------- request program
1
f
I I
1
1
1
user s read/write I eXIt
I I.
program -----r request 1
to user's
program
request I processor I
f
1 1
1 I
I I
I request I
I ~ entry I
,
1
Ii
processor 1 scheduled
I program
I
I starts
user's
program I executive
control
1
I
U.
finished I program
1 restores
1
interrupted
1
I I program
1
1 • I
operatmg system
•________________________ J
Figure S.lS Global structure of the operating system.
conversions, output string formatter, move text string; librarian facilities;

editing facilities.
System generation includes linking facilities (linking, linking loader, on-line
loader), system tables generation and an initialization module.
Off-line system software enables the processing of source programs. There
is a language processor for each programming language supported by the
process control system. Each higher language statement is compiled into a
whole series of computer operations. The most usual are FORTRAN, BASIC,
C, PL/l and Pascal which are edited, compiled, debugged, stored and linked
together.
FORTRAN. This is prepared for a run and consists of three parts:

program data;
user's FORTRAN program given as a series of addresses of data and as
a series of FORTRAN system subroutines;
FORTRAN system subroutines.
180 System software
The activation of FORTRAN subroutines is performed by a separate FORTRAN

subroutine. FORTRAN is known for its unambiguous data processing features.
BASIC. This is executed in a conversational mode using the language processor

to edit, list, sum and save the program for later use.
The BASIC language processor is usually an interactive interpreter which
examines each program language statement and interprets and executes it before
proceeding to the next statement. A message about the error condition appears
on the system terminal; program execution stops. Basic is known for its
sometimes inexact behaviour.
Programming system and system software PROZA 11 D/R * for process
control based on the RSX-IIMN3 operating system for the Digital Equipment
Corporation series of minicomputers is presented in Figures 5.16 and 5.17.
PROZA IID/R was developed for dispatch centre computers, mainly in double
computer configurations.
5.6 COST, SAFETY AND RELIABILITY OF OPERATING

SYSTEM SOFTWARE
Real-time control in a multitasking, multiprocessing environment requires both

complex programs and efficient utilization of processor time if it is to work
effectively. System software is usually developed by larger companies (e.g.
Honeywell, Foxboro, Taylor, ABB, etc). Turnkey control system suppliers
offer integrated development and production support based on the programming
standard and using elemental building blocks to simplify and reduce the cost
of programming. A large turnkey company can support an application group
of control engineers whose job it is to get to know the process well. This
group can then develop sophisticated control strategies and hierarchical control
schemes to improve control economics. It is reasonable that a full line turnkey
company would be needed to implement the process control strategies.
The mass production of control software reduces development costs, and
standardized modules can be applied in many uses without difficulty.
Nevertheless, as a generalization, process control software costs are exponentially
related to package size. Software quality and reliability have an inverse
relationship to complexity and there is a need for turnkey companies to put
the control over the operating system software. This stems from safety and
reliability considerations to prevent system crashes that could result in injury
to personnel or loss of property.
Recently developed real-time operating systems like MINIX (Tanenbaum,
1987) on personal computers make possible the immense system support offered
* Trade mark of Rade Konear-Zagreb.

References 181
application
programs
Figure 5.16 Application programs and program mechanisms in the PROZA 11 DIR
(Debeljuh and Vranit, 1982).
by personal computers and provide fine and effective solution for many non-
risky smaller process control applications.
REFERENCES
ASEA LME Automation Production Information 0011, PK1 962, Nr 88610023. ALRX-
11 Real-time Executive Programmers Reference Manual.
Borevkovic, B. and Cvitas, Lj. (1983) Koncar Strucne Informacije 30, 1-2,47.
Cihlar, z. (1980) Proceedings of the 9th Yugoslav Symposium for Measurement and
Measurements Equipment, Prishtine, Part II, pp. 539-51.
Debeljuh, V. and Vranic, P. (1982) Koncar Strucne Informacije 29, 1-2, 4 .
182 System software
report definition
report key-in
report files
- acquisition
- timing
e--1-\
(
screens
1
report blocked - screen reports r-
report list r- - printer protocols
( keyboard
}t- +
( mimics key-in l J mimics
~
operator
list of mimics ) ~
display of mimics
command
dispatcher
database list formatting ~ +

r-
{ J
functions - mail presentation
- database lists - screen lists sound alarm
- extensions - prin ter lists I--
- processing
modifications
- blocking of
t M"
functions
~ database primary ( copies )
data processing
data key-in ~ t 'Z •

spooler )
timer update
I/O functions
- local process I/O
~.
- RTU data
communication
'1
J
- r.roc~ss control
timer update
J unctIOns
- timer processing
( printer
control
- control of system
JI
responsibility
modifications
of responsibility
1/ 0 hardware
Figure 5.17 Program connections in PROZA I I DIR application package.
Mikovic, M. (1980) Proceedings of the 9th Yugoslav Symposium for Measurement

and Measurements Equipment, Prishtine, Part II, p. 521.
Rus, E. (1990) Personal communication.
v
Sajfar-Tasic,1. (1981) Coordination of Programmed Tasks in a Real-time Computer

System, MSc Thesis, Faculty of Electrotechnics of the University of Zagreb.
Skeen, D. and Stonebraker, M. (1983). IEEE Trans. Software Eng., SE-9, No.3, 219.
Tanenbaum, A.S. (1987) Operating Systems, Prentice-Hall, Englewood Cliffs NJ.
6
Application programs and databases
6.1 INTRODUCTION
Each process control system demands its own set of application programs
and possesses its own database. Simple process control systems require a modest
volume of application programs - 2 kbytes (Mikovic, Internal Report No. 1651,
Rade Koncar), whereas complex systems require 250 kbytes (Debeljuh and
Vranic, 1982) or more. Many complex systems are split up into smaller func-
tional systems since the cost of process control systems increases exponentially
with program complexity (Laduzinsky, 1982). Application programs use program
modules already developed, or macro-instructions, higher languages (e.g.
FORTRAN, C, Pascal and Ada) of functional symbols (ASEA Information
YLDA 78-08; STEP, Siemens). Some can be produced using different tech-
niques, tested, linked together and tested again, but extended real-time programs
are usually written in higher languages.
Databases for real-time applications are normally small and fast refreshed.
Extended real-time bases and knowledge bases are complex and require a lot
of data processing to be properly updated and maintained. Distributed databases
are found in distributed process control which require special methods and
techniques of updating - these include independent recovery of site failures,
resilient protocols and a complete point-to-point communication system
(Martinovic, 1983).
6.2 APPLICATION PROGRAM TASKS
Application program tasks are in the functional specification of the system.

The basic role of these programs is given in the system requirement
specifications (Heninger, 1980) which are part of problem-oriented work
(Lundberg, 1979), and quoted as external functions of the system (Wolverton,
1974).
Application programs are produced to meet the requirements of a specific
user. The data specification given in Chapter 10 is usually sufficient for the
basic description of application program functions for controllers and simple
183
184 Application programs and databases
Table 6.1 Input and output signals
Signal descriptors Signal actual data
Input signal Output signal
Signal name Switch selector Flow error

Acronym /SWISE/ /FLER/
Process hardware Panel switch selector Flow controller
Description Panel switch selector shows /FLER/ controls the output
two modes of device work: flow of a flow controller
automatic (A) and manual (M)
Characteristics of Automatic 01 Unit: m 3/s
values Manual 10 Range: -3 to +3
Accuracy: ± 0.1
Resolution: 0.002
Instruction sequence READ 14 (channel 2) WRITE 200 to 400
(channel 5)
test carry bit = 0 for request
acknowledged; if not, restart
Data representation Bits 7 and 8 8 bit two's complement;
scale 256/3 = 85.3
Timing characteristics Acquisition time 100 ms Display time 2 s
Comments /SWISE/ = 00 for neutral Mechanical limit set for
position of switch selector range outside ± 3 m 3/s
real-time systems. This is because they answer questions on where input and
output data are fed and what logic or other data processing functions should
be performed in what time restrictions. Table 6.1 defines an input and output
data signal.
More complex applications, e.g. positioning systems, can be solved by
combining specific application programming modules in a given operating
system that also possesses driver routines for the connection with hardware.
Figure 6.1 shows a block scheme showing software control of the analog
output.
A more systematic approach to programming is necessary for a larger control
system where previous techniques might prove to be cumbersome. A special
set of symbols for use in program function diagrams has been developed by
ASEA, Sweden.
A language tool, TLM from Siemens, using a successful combination of
logic and arithmetic (analog) variables, enables structuring of control functions
and parametrizing of process values, checking of control functions, use of
graphic features and protocolling of process values.
Application program tasks 185
system initialization
,-
I
I
: {routine
: ---- active at
,--_ _-'-_ _ _-, level 3
calculation of
control function PAG
and ramp generator
synchronization for process
values output analog
variable [ control to
enlarged ,------1.-------------,-- operating
detail of MEH system
__ analog unit~_-I set-point for
output I analog output
control I unit
IL-----r___--'
I
____ {cont~ol to
I
I
I
I
I mom tor
'-
Figure 6.1 Block scheme of positioning system with a software detail of the analog
output control (from M. Mikovic, Internal Report, No. 1651, R. Koncar).
The development of minicomputer-based application programs is supported

by more powerful operating systems, and as application programs are written
in different programming languages, a standard system version supports the
FORTRAN and macro assembler. Pascal is also valuable, since it preserves
more modularity than FORTRAN, with a reduced speed of only 3-4% (Tunon,
1982).
The programming system PROZA II D/R preserves a powerful interaction
of application programs with real-time functions by a set of program services
(Figure 6.2), that allow the following:
I. communication with peripheral units;
2. database access;
3. connection with system programs;
4. interactive communication with the operator from the operator's control
place by means of video display and a keyboard unit;
5. organization of own application databases.
primary data
communication with
peripheral
units syst~~~_o_p_e_ra
~Place
.....tor's
communication
with
basic
data
Figure 6.2 Application programs and mechanisms in PROZA 11 DIR.
An operator's control place is connected with application programs by an

interactive communication keyboard and video display unit. More application
programs can be involved in the parallel communication that possesses the
elements shown in Figure 6.3 (Debeljuh and Vranic, 1982).
A multifunctional parallel usage of the operator's control place, organized
in the system PROZA 11 D/R, is enabled by the multi-purpose use of the
keyboard and the communication of more application programs with the
keyboard.
Application programs can redefine dynamically the set of keys they will
use in communication with the keyboard control program (Sajfar, J., private
communication). The commands issued are connected with the mimic diagram
on the video display. The application program can be initiated by:
/
software mechanisms /
\ /
\ /
/
- mimic refreshment
-.-.'---~ - initialization
- spontaneous program
initialization
- functions synchronization
automatic mechanisms
- program start-up by cursor

and functional key
- program start-up by virtual
keyboard
- search of mimic element
by key code
- search of mimic element
by cursor
interactive mechanisms
"-
"-
"-
'\
- data read-in by means of
\
\.
\ standard protocol on display \
\ - direct read-out of data
on display
- acceptance of characters
.. \
\
from keyboard
\
\
,
data read-in/out
mechanisms \
/
,/
,/
Figure 6.3 Interactive communication with operator's control place (Debeljuh and
Vranic, 1982).
calling the corresponding mimic diagram;

putting the cursor on the mimic element and activating the corresponding
keys on the functional keyboards;
a virtual keyboard (six keys) on the functional part of the keyboard with
multifunctional meaning (i.e. their meaning is shown on the screen).
The tasks of application programs differ in each specific application, as indicated
in Table 6.2.
Real-time and extended real-time application programs are executed where
the following requirements and constraints are to be defined and obeyed prior
to their definition.
1. Computer characteristics: if the computer is predetermined, a general
description of computer limitation is given; otherwise a summary of its
required characteristics as discussed in Chapter 3.
2. Hardware interface: a concise description of input and output hardware
units and devices where information is accepted and transmitted by a
computer (see Chapter 2).
3. Software algorithms: functions and data; what software must do to meet
its requirement specifications in various situations and in response to various
events.
4. Timing constraints: how often and at what time margin each program
function must be performed.
5. Accuracy constraints: how close the program output values must be to
ideal values to be acceptable for process requirements.
6. Response to undesired events: how the hardware performs and what
software is to do if, for example, sensors go down, if set-point positioner
does not respond, etc.
7. Program modularity: what parts of the program should be easy to re-
move in order to be substituted by other parts; also what parts of the
program should stay the same no matter what changes are made later on
in the programming system.
8. Program life cycle: what changes have already been made and which are
expected to be made later on.
Example
An industrial data transmission unit can be defined by the following charac-
teristics.
1. Computer: an M68000-based multiprocessor unit, up to 640K memory
and separate interrupt unit.
2. Hardware interfaces: communication unit, local console pushbuttons and
indicators, local printer for process data print-outs, bus extension, process
interface modules for up to 256 digital inputs, 128 analog input and 64
binary output, real-time clock.
Structure and timing requirement of application programs 189
3. Software algorithms, functions and data:

input data collection;
communication of input data summaries to superimposed computer;
message communication with superimposed computer;
local protocol print-out;
acceptance of data from local operator console;
in case of communication failure local work mode is available;
EEPROM store of configuration data.
4. Timing constraints:
all input data must be accepted, maximum input rate 0.1 Hz;
if after five trials there is no communication then switch to local mode,
and repeat the sequence each hour;
local protocol print-out each day at 8.00 a.m.;
change of console data accepted between 7.00 and 8.00 a.m.
5. Accuracy constraints: 0.1% at 35°C with a 0.0057%;oC temperature
coefficient.
6. Response to undesired events:
alarm to change of back-up battery voltage;
- alarm to change of process connections;
- automatic initialization after system failure.
7. Program modularity:
easily removable local console configuration and local protocol print-
out;
communication protocol and input data collection remain regradless
of other changes.
8. Program life cycle:
the number of process inputs may be expected to increase by a factor
of two.
change of communication coding can be expected.
6.3 STRUCTURE AND TIMING REQUIREMENT OF

APPLICA TION PROGRAMS
Real-time application programs usually take a simple chain-type structure,

which takes the form of a row of activities in sequential order that execute
data processing tasks one by one and usually do not return or re-enter the
beginning of the program. An application program can be activated as illus-
trated in Figure 6.4.
There are different activation routines for application programs and requests
from application programs in real-time operating systems. For instance:
activation request of a given application program into a waiting queue on
the priority basis and on the first-in-first-out principle;
Table 6.2 Application program tasks
Group of tasks Logic Measurement Protection Process control Supervisory

controllers system system system system
Real-time tasks
Instantaneous input of process data - analog + + + +
Instantaneous input of operator's data
digital + + + +
analog + +
Data processing
logic + + + + +
arithmetic + + + + +
digital filtering + + +
complex + + + +
Data presentation
real-time data presentation + + + + +
post-mortem analysis + + +
chronological event recording + + +
Data output and operator's intervention into
process
direct control of important process points + + + +
control of on/off process points + + + +
control of set-point values + +
remote control of process points + + +
Communication facilities
data and commands communication + + + +
message switching + +
Extended real-time tasks
Process state calculation and estimation + +
Process state simulation and modelling + +
Process identification and optimization +
Expert system tasks
Automatic start-up + + +
System parametrizing + + + +
Functional diagnosties + + + + +
/
./
./
./
./
./
./
./
./
activation ~
activation
l~/I ./
./
./
./~
t
device activation device application
driver driver programN
application
program
,
activation application
program 1
process activation~ time-delay
I/O unit module
activation
operator's activation activation
console
L/~:J
Figure 6.4 Activation of application programs.
activation request of a given application program into a waiting queue on

the last-in-first-out principle;
time-delayed repeated activation request into a waiting queue on the first-
in-first-out principle (the application program cannot be activated by any
other program during the time delay interval);
time-of-day activation request into a waiting queue on the first-in-first-out
principle;
process input or output action request on given process input or output
units (the program must not be activated by any other means during the
input/output action if after this action the program execution returns to the
same application program);
request for temporary use of given memory space from the operating system;
return of temporarily used memory space to the operating system;
request for temporary enabling and disabling and disabling of a certain
program;
request for obtaining the starting address of a certain application program;
request for obtaining the system timer data and time-of-day data.
An example of a flow chart of a real-time application program for digital
input/output processing is shown in Figure 6.5 and the corresponding task
activation
of digital
signal
processing
(PI)
change of
pointer of
digital input
table (DIT)
activated
by operator
print-out
program delete the
(P5) known message
out of digital
input table
Figure 6.5 Flow chart of digital input/output data processing.

procedures are shown in Figure 6.6. The programming chains for a specific
class of messages are given in Table 6.3 and a complete schedule of data
processing of digital signals is given in Figure 6.7. The four subsets of data
processing chains of the application program for the acceptance of messages
and the direction of data processing (see Figure 6.6) are:
API = (AI' A2, A3) (6.1)
AP2 = A4 (6.2)
AP3 = As (6.3)
AP4 = (A6 , A7 , Ag) (6.4)
A stochastic schedule of the execution of program subsets API, AP2, AP3
and AP4 exists, since the change of digital input signals is of a stochastic
nature. The program chains for already printed out messages to get out of the
queue are A 10 and All. The put-in and get-out of table (named DIT) is
completely synchronized if at the end of the data processing interval the number
of executions of subsets N(APl) and N(AP4) equals the number of chains
N(A IO ) and N(A ll ), that is, when
N(API + AP4) = N(A IO + All) (6.5)
The beginning of the execution of a program chain Ai denoted Al and the end
of the execution of a program chain Ai denoted Al are connected in the fol-
lowing manner
(6.6)
The given characteristics of the described operating system enable the inter-
ruption of a lower priority program by a higher priority program; however, it
cannot be activated until the interrupted sequence is completed. The overlap
of program chains inside the same program is therefore not possible and the
determinacy of the program P2 is preserved. Naturally the program chains are
serially connected so that after the last task of a program chain Ai the first task
of the program chain Aj is continued.
There are two ways of communicating between application programs and
operating systems: direct and indirect communication. The indirect commu-
nication of programs is performed via the database.
6.3.1 DIRECT COMMUNICATION
The direct communication is performed by a manipulation over an internal

variable of the program where the communication is directed; this variable is
named a semaphore (see Figure 6.8) and is used for the synchronization between
the transmission of a message from the operating system and the reception of
All Al A7 A3
A6
Figure 6.6 Application program P2.

Table 6.3 Programming chains for a specific class of messages
Class of data processing Subclass of data processing Chain of'tasks Number of tasks chain
Process interaction and print Message received in empty T] < T2 < T3 < Ts < T7 < TIO < T]4 < T]S < T]6 < T]7 A]
out digital input table
Message read-in into last free T] < T2 < T3 < Ts < T7 < TIO < TI3 A2
location of digital input table
There are still free locations T] < T2 < T j < Ts < T7 < T]O < T]4 < T]7 Aj
in digital input table
Message transmitted to Number of locations in digital T] < T2 < T3 < Ts < Tg < T]2 < T]3 A4
program for process input table unchanged
intersection through digital
outputs
Message transmitted to Number of locations in digital T] < T2 < T3 < Ts < Tg < T]] < T]3 As
program for production input table unchanged
registration
Message to be printed out Message read-in into empty T] < T2 < T3 < T4 < T7 < TIO < T]4 < T]S < T]6 < T]7 A6
digital input table
There are still free locations TI < T2 < T3 < T4 < T7 < TIO < TI4 < TI7 A7
in digital input table
Message read-in into the last TI < T2 < T3 < T4 < T7 < TIO < T!3 A8
free location of digital input
table
Message of nonconnected Number of free locations of TI < T2 < T3 < T4 < T6 < T9 A9
digital input digital input table remains
unchanged
Deletion of already printed After deletion of message the TI8 < TI9 < TI3 AIO
out message from digital digital input table remains
input table empty
After deletion of message TI8 < TI4 < TIS < TI6 < TI7 All
from digital input table the
print out of left over
messages is activated
program
acceptance of status priority
signals from process levels
data and proceedings of PI 5
,
processing level I status states to
application programs
data acceptance of message

in application program P2 4
processing level 2
system and direction
of data processing
t
data activation of output
processing level 3 on/off signals P3 4
~
data
production
processing level 4 P4 3
registration
~
data operator
processing level 5 print-out on teletype P5 3
~/~ P3
r- ::~~--.J P4
! """-..1 L
! 1 P5 r --I
I'- ___________________ _ _ _ _ _ _ _ _ J I
Figure 6.7 An example of a schedule of data processing:

- - .. , indirect activation.
-+, direct activation;
postpone
_~ send) (\
flag = I flag set
program PI
internal variable flag
message - I
flag = t,3 flag reset
EMTI 1 •
indirect transmission direct flag r eset
of message to P2 ifDIT< N
t
message
receiver ( DIT<N )
internal variable DIT
program P2
Figure 6.8 Direct communication between programs.
a message from the application program. The program PI as a part of the

operating system cannot use a direct search of the state of the table DIT
because the variable DIT is an internal variable of the application program P2
and it cannot be a global system variable according to the rules of a given
operating system. The synchronization of these two programs is done by
manipulation over a semaphore that is an integer internal variable of the
operating system program PI. This variable is automatically set to a logic
state I when the message is indirectly sent to the program P2. If the internal
variable of the program P2 is less than N, i.e. if the buffer in the program P2
is not full, then the program P2 directly resets the state of the semaphore
variable in the program PI. Thus the synchronization of the operating system
and the application programs is obtained.
6.3.2 MULTIPROGRAMMING CONSTRAINTS
Many critical system parameters can exist in a multiprogramming real-time

system environment concerning the application situation. The most critical
system parameters are system time response, memory usage, and device usage.
The system time response seems to be the most crucial parameter in some
real-time applications (Leinbaugh, 1980). To express the quantitative relations
of a guaranteed response time, the following characteristics of program task

chains are supposed:
each task chain consists of tasks running in a sequential order;
tasks may use devices;
the task chain may include critical sections where it exclusively uses data
areas and devices during task execution;
each task must be completed within its guaranteed response time.
In order to calculate a guaranteed response time, the following information on
each task chain is needed:
maximum processor time needed for each task;
maximum operating time of each device;
resource segments needed by each task;
position of device requests within each segment and position of the task
within each task chain.
Processing a task chain is as follows:
resource segments have a priority;
when only nonresource segments use the processor, each segment is run
at a specific rate of processor data processing progress;
device requests are handled on a first-come-first-served basis;
all resources needed by a task are simultaneously allocated and the earliest
task request is handled first;
the task is not to be started earlier than its guaranteed response time after
its most recent previous starting time.
The worst case time to run a task chain TTCi is:
TTCi = TPR i + TDi + TPNRJratei + Hi (6.7)
where:
TPR i is a total time requirement of the ith task chain for the processor
handling resource tasks;
TDi is a total time requirement of the ith task chain for the performing
of device operations;
TPNR i is a total time requirement of the ith task chain for the processor
handling nonresource tasks;
ratei is a proportion of the ith task chain that is needed for its handling
among other task chains; the sum of all task chain rates equals 1, i.e.
(6.8)
Hi is the worst case amount of time when the ith task chain is blocked
or slowed down by other tasks by using or needing to use the same
devices or resources.
task 1 task 2 t
resource Rl --"::"'1>
~'(~~~'in 11"""---IL-_d_e_Vi_c_e_D_l_---4t --- -1L-__d_e_V_ic_e_D_2_--t1

..::..:.:: resource
•• 4 6ms •• 4 lOms
....:...:.. non-resource
.,
processor time processor time
task 3 task 4
resource R2
.. lms ~. 3ms 6ms... lOms

~ resource ~ non-resource
task 5 task 6
lOms
.'4non-resource
6ms
Of • 4 ~resource3ms
, lms......
Figure 6.9 An example of a guaranteed response time.
The longest time is when the ith task chain is blocked or slowed down, and is
Bi = L TPR j [(TTC;I TTC j ) + 1] + BDi (6.9)
j=l
because the jth task chain can be executed (TTC;/TTCj ) + 1 times during a
given execution of the ith task chain (resource blockage); the blockage BDi
due to the use of devices by other task chains has to be calculated for each
particular case. Figure 6.9 shows an example of a guaranteed response time
for three task chains.
In a given example TPR i = 1 ms, TDi = 13 ms and TPNR i = 6 ms for i =
1, 2, 3. Bi = 2·2 ms + 6 ms = 10 ms for i = 1 and 2 because the blockage
caused by devices in other task chains is only 2 . 3 ms = 6 ms; the task chain
1 is, for example, slowed down once in its device D 1 action by the D 1 action
of the task chain 2 and once by the Dl action of the task chain 3, because the
total time requirement of task chains according to equation (6.9) equals 2·2
ms = 4 ms. B3 = 2·2 ms + 6 ms = 10 ms also for the task chain TC3. This
is due to the last task of the task chain TC3 being blocked by the first tasks
of both task chains 1 and 2. With the help of the relationship between differ-
ent tasks.
TTCj = X· (relative TTC j) (6.10)
where x can be chosen to select any set of times that maintains the relationship
between different task chain response times, equation (6.7) becomes
ratej = TPNR j . X· (relative TTCj) - TPR j - TDj - Bj (6.11)
By putting the calculated data into equation (6.11) and by letting relative
TTCj equal 1
ratej = 6/(x . 1 - 1 - 13 - 10) ms, for i = 1, 2, 3 (6.12)
When x = 42 the relative rates equal 1/3 and their sum equals 1. The guar-
anteed response time is 42 ms. If the task chains are actually restarted that
often, then the total processor usage is (21/42) . 100 = 50% rather than (7/
20) . 100 = 35% when each task is run separately.
Actual relations in the system time response include the operating system
overhead consisting of:
request handling time of the task chain that consists of the start task chain,
start task, request device, handle device interrupt and semaphore syn-
chronization time;
tennination handling time of the task chain;
timer interrupt handling time.
Each of these time-consuming operations has to be added to the required time
consumption given in equation (6.7).
When multiple or parallel (simultaneous) use of devices takes place then a
single semaphore is not enough. The problem when multiple or parallel
(simultanwous) use of devices takes place then a single semaphore is not
enough. The problem can be solved by using an array of semaphores
(Tanenbaum, 1987) which defines the state of all the devices in simultaneous
use.
6.4 DAT ABASE AND BASIC PROCESS SOFTWARE
Basic process software enables the connection of the operator, process and
superimposed or subordinated process control system by global system functions
such as those illustrated in Figure 6.10. The use of global system functions
depends on the actual process control application and the hierarchical level of
the process control system. The basic process software consists of:
database;
- programs for autonomous functions (process data acquisition, process data
(a) process data
process communication to
data higher
display hierarchical
level
process commands to
data process and control
acquisition of commands
execution
(b)~
I--------------r--l:~~:der
•
operator
•
I
process
Figure 6.10 (a) Main parts of basic process software. (b) The structure of a basic
process software.
analysis, processing of events and alanns, refreshing of video display units,

sensing of functional keyboards);
programs for operator-system communication (the supervision of operator's
consoles and system supervision, system control, operator-system dialog);
other programs (output of lists, statistical control of operators, automatic
recording of production or consumption).
The process database is a passive part of the basic process software and mainly
consists of two parts, the process data and system tables. Different program
interactions are centralized in a process database and consist of a set of tables
usually containing:
process data consisting of total and precise descriptions of each process
input and output signal or command, including all its parameters needed
for future data processing;
parameter tables for definition of data processing types;
parameter tables for definition of output lists, contents of particular mimics
and curves on video display units, contents of reports and contents of logs
on teletypes and line printers;
other tables for system definition parameters.
For the systems with external memory facilities, a part of the process database
is permanently located in the operating memory. The rest is on external memory
media such as disks, floppy disks or magnetic tape memory units. Process
data and parameters used for process data acquisition, processing and com-
munication, process data analysis and refreshing the video display units are
located in the operating memory. The connecting parameters enable direct
access to the process database, thus eliminating a time-consuming data table
search.
The process data can be broadly divided into four sections: single status
data, double status data, analog value data and counter data. One row of the
corresponding table is attached to each process signal, which corresponds to
each class of process data. Therefore, each process signal possesses an inter-
nal addressing system that attaches two coordinates to each process data point:
table address and position address inside the table. This is termed an internal
addressing system, but there is also an external addressing system since
principally each data point has to be accessible from another data processing
system within distributed data processing systems. The external addressing
system enables the connection of process data points to a distributed process
database. A data processing system with three levels of hierarchy can possess
the external addressing system in Figure 6.11.
The different status, analog and counter data and the corresponding database
assignments required for the supervision and control of plant processes are
given in Table 6.4.
Information on digital commands and set-points needed for databases in
process control systems is given in Table 6.5.
Database and basic process software 205
r assignment word
I
type of
. • process data
.
I
. , process data
. , I apparatus point marking
markmg 000 - 999
(r; -
level 3 /S - single status data
~
00 - 99
level 2 marking double status data
{
marking 00 - 99
A - analog data
0-9
!
C - counter data
level 2
level 3
the level of process apparatus and process data points
Figure 6.11 An example of the external addressing system.
Tables of connecting parameters (or connecting tables) are part of the

process data in a database. The connecting tables principally consist of the
following.
1. Connecting tables between communication programs for supervision of
data acquisition from lower level system parts and a process database; the
number of these tables equals the product of the number of lower level
computer systems and the number of data classes; these tables contain
internal data addresses from lower level computer systems.
Table 6.4 Status, anaolg and counter data and the corresponding database assignments
Name of status information Graphic representation Database parameters to be maintained for process
supervision and control
Status infonnation Instantaneous state, last known state, change of

state, active or passive state, alann state (0 or 1),
single point infonnation
cr---c/ 0--0 status code, code of location and apparatus internal
address, starting state message, codes of peripheral
units for infonnation output
double point infonnation "

''0--0
Event infonnation Instantaneous state, last known state, change of

state, active or blocked state, alann state transition
o to 1 or 1 to 0, status code, code of location and
0-1-0--- apparatus, internal address, starting state message,
codes of peripheral units for infonnation output
Counter data Instantaneous counter state, total counter state as
per day, actual counter state (counting, blocked at
serial counter data lntrLf---- dispatching level, manually preset state, counter
error), counter code, code of location and
apparatus, internal address, starting state message,
description (conversion factor, previous state,
-.-J/O~ previous total, time of manual input, time of
parallel counter data o manual input validity), codes of peripheral units for
information output
---J.
)'0>-----
Analog data A Instantaneous measurand, previous measurand,

upper and lower limit of measurand, upper and
lower limit of measurand scale, dead zone,
.1 measurand levels for alarm and critical alarm upper
A 0
fAIr and lower levels, binary status (measurand value
blocked at dispatcher level, manually inserted
value, alarm level, critical alarm level, out-of-range
level, change of alarm state), conversion factor,
analog data code, codes of location and apparatus,
internal address, adjustment, codes of peripheral
units for information output
Tables 6.5 Database information for process control systems
Name of command Graphic representation Data base parameters to be maintained

information for process control
Digital Instantaneous state, final state, type of

command time execution (pulse, maintained,
persistent), control time delay,
- - - - / : '0 - conditional status (and its code) for
I enabling command execution, the
D
state of digital status that must be
maintained after control time delay,
single command command code, control function (open,
close, start, stop, switch on, switch
off), code of location and apparatus,
codes of peripheral units for command
display and recording, display
description
I
double command
'I..I"
10--
I,
Set-point Instantaneous state, final set-point state,

command allowed limits, conversion parameters,
time parameters conditional status for
enabling set-point execution, set-point
code, control function (opening,
closing, starting, stopping) set-point
code, code of location and apparatus,
code of peripheral units for set-point
display and recording, display
description.
2. Connecting tables between programs for visual data display or data print-
out and a process database.
3. Other minor groups of connecting parameters and definition tables as
well as system tables, among them specifically:
lower level hierarchy computer systems;
process apparatus;
process data points for the historical periodic record;
video display and printer lists;
system configuration and dynamic system data;
definition tables of group data;
mimic displays;
trend displays;
operating
system
application application
program 1 program 2
database
table
Figure 6.12 Data sharing between two application programs:

path; ....... control path.
===>. data
consuming points;
system operator and dispatcher names.
6.4.1 ACCESS TO DATABASE
Data from a process database can be shared among many application pro-
grams. The support of database share can be organized by the program sup-
port inside the operating system and software mechanisms built in at the
beginning of those application programs that share a common database table
with other application programs. An example showing data sharing between
two application programs is given in Figure 6.12. A typical global structure of
a database, as shared by application programs in a teleprocessing environment,
is given in Figure 6.13.
6.4.2 BASIC FACILITIES OF AN ON-LINE DATABASE
These include:
1. Integration and control of data content: implies the coordination and
controlled integration across the number of application program areas to
SER FUNCfIO S
~ ,.,..~ TELEPROCESSING ENvrRO

:£,1'" NME rr
.,
•
\
\
II ,,
\
,
II
\ ,
\
~'''''''' \\
" .
'\
PROCESS I/ O
LOCAL DISTRIBUTED REAL-TIME DATABASE
Figure 6.13 Global structure of a distributed real-time database.
obtain efficient use and consistency of data structures. Thus, data dupli-
cation and data relationships are finally controlled.
2. Complex data structuring: implies some forms of complex file structur-
ing. For example, hardware pointers, inverted indexes or logic iden-
tifiers that provide a variety of entry points and access paths through this
complex structure to enable the efficient retrieval of data for a variety of
applications.
3. Retrieval and maintenance of data: provides the user with ready programs
of adding to or modifying the structure, retrieving and amending existing
data and storing row data in the database. The data manipulation facilities
should provide for the organization of parts of the database and for the
analyses and summarization of data contents and usage.
4. Data independence: implies the complete separation of the storage and
organization of data from the logic and processing of the various appli-
cation programs using these data.
5. Concurrent usage of a database: enables the sharing of the same data by a

number of users at the same time, without conflicts and inconsistencies
resulting from several users accessing the same data unit.
6. Integrity control: enables the control of the accuracy of data content after
entering the control system, and timeliness obtainable by data reconstruc-
tion after system failure.
7. Privacy control: preserves the confidentiality of the following:
contents of the database by preventing unauthorized access to specific
data units or data classes;
- use to which the data are put by application programs by the control
of application program algorithms;
- relating data from different parts of the database, especially historic
data, by the control of application program algorithms;
- right to modify existing data by the control of application program
algorithms.
6.4.3 DATABASE ORGANIZATION
The orgainzation of the database and database program support essentially

influences the response time and system extension capability. These are two
important features of process control systems.
The database subsystem is an ordered set of data partially stored in a fast
control memory and in a peripheral memory. The database of the real-time
software system PROZA 11 D/R (Debeljuh and Vranic. 1983) consists of process
data, process information system data (auxiliary data), derived data and user's
data.
Process data: the group of process data is formed from the data acquired
from remote terminal units or local process subsystems.
Process information system data: describe the state of particular hardware

and software elements, and communication lines. These data are obtained by
on-line autodiagnosis and supervision of the process control system.
Derived data: are obtained from raw analog and binary data (count and count-
rate data) by applying arithmetic and logic transformations to two or more
process data.
User's data: are put into the system by the user. Read-in of these data usually
includes a part of software that realizes this data processing algorithm.
The access and use of a database subsystem in the program system PROZA
11 D/R is given in Figure 6.14. The organization of the database follows the
hardware organization of the process control system because the data of each
data classes
status data
process ---1---- --- ---
analog data
data ---1--- --
counter data
plausibility of,
user's measurement data
data
auxiliary
data
derived
data
operator
class 1
addition of new addition of

data classes new data
Figure 6.14 Database subsystems in a PROZA 11 DIR sofware system and

database planning:--a-., software mechanisms;~, change of data
processing parameters; -<D-+, addition of new data;~, addition of new
data classes; --@-+' connection to mimics and mimics generation; --®-+'
manual data read-in.
process input are uniquely determined with their physical input into the sys-
tem. The data key for each process signal contains the data class (e.g. status,
analog or binary variable) and a description of three level hardware con-
figurations through which signals are fed to the system (see Figure 6.15).
The data are partially fed into the fast memory of the central computer and
partially on to system disks. Data descriptors and parameters that demand fast
access and frequent use are stored in the fast memory. These data are ordered
in the same way as keys in the key tree, i.e. according to hardware structure.
Data descriptors and parameters that are rarely used are stored in disk media.
The access to these data is also done through the key tree by hash ad-
dressing, which does not impose serious limits, especially when a hardware
specification is not well known. According to Figure 6.15, the key is formed
by three memory symbols Klo K2 and K 3 , each for every key level. The
pseudostochastic transformation of this address is done by applying the following
function
level 1
1 1 1 1 data class (process data)
- centralized computer
- communication unit
at centralized computer
- remote terminal unit attached

to the communication unit
.- hardware module in the

- signal channel on the hardware

module
Figure 6.15 The key of the addressing tree in the system memory.
R(K" K 2 , K 3 ) = ««(KI . M) mod 2W) . K 2) mod 2W)

. K3) mod 2w (6.13)
where M is the parameter of influence for the behaviour of function R, w is
the length of a memory location (w = 16 for a 16 bit computer word). The
goal address of the data is determined from the pseudo stochastic value by
applying the function
F(K K
" 2,
K)
3
= [R(K" K22w, K 3 )· A] + 1 (6.14)
Where the bracket [N] determines the greatest whole number of N equal to
or less than N.
6.4.4 CONTENTION RESOLUTION
A collision situation appears when two or more keys are transformed into the
same goal address. Two methods are applied against this process:
1. Connection of the whole primary bucket to each goal address. B keys can
be stored into each primary bucket. The value B is deliberately chosen.
The number of address keys AK, address space A and bucket length BL
are connected by
A=[:J (6.15)
where [N] denotes the smallest whole number greater than or equal to N.
2. When a primary bucket is full an overflow bucket of the length BL/2, is
reserved, where all overflow keys are stored. The address of the overflow
bucket is stored in the primary bucket. When the first overflow bucket
is full, a second overflow bucket is formed with its address stored
in the first overflow bucket. Thus, a list is formed for each goal
address that begins in the primary bucket with overflow buckets as its
elements.
The application of hash-coded addressing for 21 600 keys with the space that
is 24% greater than the minimal required memory space has shown the fol-
lowing:
10 18 448 cases it was possible to do a read-write operation with one
disk access;
in 2686 cases two disk accesses were demanded;
in 440 cases three disk accesses were demanded;
in 26 cases four disk accesses were demanded.
6.5 DISTRIBUTED DATABASE
A multiprocessor process control system can effectively communicate through

common memory locations, but a distributed process control system demands
a more restricted behaviour of the system, especially in a distributed database.
Data plausibility is not as decisive for distributed process control systems, such
as for distributed real-time systems, since the change in database updating is
more critical in the former. The reason lies in a relative low level of confidence
in process data acquired, since process sensors can send false or inaccurate
data, be unconnected to the system, and have no redundant measurement for
all process variables. Thus, plausibility criteria for measurement data of pro-
cess control systems are important in a distributed environment. The potential
advantages of a distributed control system are:
lower price (Le. cabling costs saved);
smaller time response due to computers' proximity to the process;
simpler adaptability of computers to process and control structure changes;
inherent parallelism enhancing process working quality.
Distributed database 215
The advantages of a distributed architecture have to be supported by efficient

interstation communication in the presence of delays, noise and system fail-
ures. The basic scheme of distributed control system is presented in Figure
6.16 with the following features (Martinovic, 1983):
- strong interdependency of function between controllers, and especially
between neighbouring controllers (process dependence);
- fast response time inside the system;
- safe scenario in case of process failure, i.e. slow plant is switched over to
manual control on critical basic process units and fast plant is to make
shutdown economically feasible.
Three types of failure can be distinguished:
communication failure;
- process connection failure;
- failure of the basic process unit controller.
Some common procedures for failure repair in case of strong functional de-
pendence or fast process response are given in Table 6.6. Total communication
failure is not an allowable event because it demands that the separated network
parts behave as separate blocks (Skeen and Stonebraker, 1983) or it demands
decision making in accordance with other parts of the network. When a part
of the database is temporarily not accessible from another part of the network
this part will lead to action as a separate block, thus disabling the system for
work (i.e. for permission issuance).
It is possible to compensate for the failure of the basic unit process controller
by a protocol resilient to the failure which fulfils the following two rules
(Figure 6.17).
Rule 1 If for every state s in the protocol set of states C(s), where protocol
transits from the state s, contains a commit state, then a failure transition from
s to a commit state is assigned; otherwise a failure transition from s to an
abort state is given; a failure transition thus denotes the occupation of initial
states which enable recovery at the moment of the return into the working
state of the controller and these states are process safe.
Rule 2 To each intermediate state s in the protocol possessing a timeout
transition to a commit (abort) state, sent by a failure site in the course of a
normal message, a timeout transition from state s to a commit (abort) state
must be assigned.
The working of a distributed control system with a mUltiple communication
connection and a controller failure is permitted (Martinovic. 1983) when the
consistent operation of a distributed database is by lateral controller connec-
tions and a distributed knowledge base made possible (Bailey, 1988; Jovic.
1989). Lateral controller connections are usual in distributed systems for process
operator to other
'data hiway'
•
II
1
~ , ,
operator's supervisory
interface computer
primary and secondary 'data hiway'

communication
) \
repeater
J
.
pnmary I ..
In
an d secon d ary commumcatlOn l'k
n
to other
process I/O
) subsystems
~ intelligent
~ r
in telligen t
operator's
I/O process I/O process
interface
unit unit
,
•
III
II J
-------- -------
processl/
Figure 6.16 The concept of a distributive intelligent I/O control system.
Table 6.6 Failures and repair design in distributed control systems
Failure type Process with strong Fast response process

interdependencies of basic
process units
Failure of process Lateral data transmission Transmission of critical

commands in safety system
Connection failure Work in a safe scenario working mode in n - 1 system
components
Controller failure Resilient protocol Figure Transmission of critical
6.16 commands in safety system
Communication failure Lateral hardware Local working mode
connections Lateral hardware connection
Resilient protocol, if only
feasible
Figure 6.17 Two phase commit protocol with timeout transitions and undeliverable
message transitions: ---, deliverable message states; - . - . - ., timeout
transmissions; - - - -, last message undeliverable.
centre 1 centre 2
+ centralized computer
TL TL
, I
I
I
I
+
,+
front distributor
: (doubled if necessary)
I
I
remote terminal
units
process
--------- - ,I
Figure 6.18 Hierarchical distributed system with data concentration point:

TL, multiple transmission lines; RTU, remote terminal unit (Debeljuh and
Vranic, 1982).
safety reasons (Lichtberger, 1990) while a knowledge base needs more ex-
planation - see Part Three of this book.
A distributed configuration with a concentration point is shown in Figure
6.18. A front-distributed concentrator, DS-8FD, is located between a network
of remote terminal units and a centralized computer. DS-8FD can also be
realized as a double configuration. The basic task of a distributed concentrator
is data acquisition and administration from remote terminal units and their
distribution to centralized computers. A summary of the basic functions of the
distributed concentrator is as follows:
data acquisition for centralized computers;
message switching to centralized computers;
supervision of data transmission through appropriate protocols;
transmission of on/off commands and set-point commands from the cen-
tralized computer to remote terminal units;
generation of time synchronization upon request from centres and its de-
livery to remote terminal units;
communication speed conversion.
The basic programming structure of a DS-8FD system is illustrated in Figure
6.19.
The database system is a queue-oriented system, each queue containing a
given number of queues of different priority for each remote terminal unit.
The contents of these queues are mapped into an independent queue-system
TL to centralized computer TL
communication
program C2
sorting
routine
request scanning
processor processor
TLn
to RTU's
Figure 6.19 Basic program structure of the DS-8FD computer: TL, multiple
transmission lines; RTU, remote terminal unit.
for each centralized computer. Thus, a complete state of all remote tenninal
units is mapped into a queue-system of centralized computers. The DS-8FD
computer contains two independent queue-systems each for one centralized
computer. The organization and appearance of queues in a DS-8FD computer
is shown in Figure 6.20. This organization enables selective data acquisition
in centralized computers according to data priority.
6.6 EXTENDED REAL-TIME SOFTWARE
Extended real-time software is a set of models, programs and routines used by

the system operators for better knowledge of the process state and easier
process guiding, control and planning. Extended real-time software can be
distinguished according to the following two parameters:
process structure, where the interrelations of various process parts, basic
process units and their functioning are studied and modelled;
process working mode, where different process modes (e.g. starting work-
ing mode, stationary working mode, quasistationary working mode) are
observed, studied and modelled.
The complexity of the process structure and the number of process working
states indicate the overall number of process models used. The total number
of programs and routines used depend on the operator-system dialog facilities
as well as the actual process application. Thus, the basis for extended real-
time software is a process model needed for better process description. Models
can be of a material or symbolic type, or of a static or dynamic type. A basic
survey of model types, with some examples used in process control systems,
is given in Table 6.7. A mathematical model executed in extended real-time
software can be described on the basis of the equation for the equilibrium of
mass, energy and information content being valid for a given process or a
process part such as:
dm/dt = mu - my (6.16)
dE/dt = Eu - Ey (6.17)
dI/dt = Iu - Iy (6.18)
where
m, mu and my are instantaneous quantities of accumulated, supplied and
removed mass;
E, Eu and Ey are instantaneous quantities of accumulated, supplied and
removed energy; and
I, Iu and Iy are instantaneous quantities of accumulated, generated and
processed information.
Extended real-time software 221
FIFO 3 - detail
FIFO 1
FIFO counter
FIFO 2 RTU
input index FIFO 3

centralized
output index
§1 ---...,.-'
RTU
computer 1
buffer 1
buffer 2
iiiil J RTUn
FIFO 1
FIFO 2 RTU 1
centralized
computer 2
buffer m
Figure 6.20 Organization and appearance of FIFO queues in a DS-8FD computer:

FIFO, first-in-first-out (queue); RTU, remote terminal unit; 1,2, 3, indices of
priority data levels; n, number of distributed remote terminal units; m, number of
buffers in a queue of each priority level.
The variables that characterize the state of accumulated mass, energy or infor-
mation are termed 'state variables' and denoted by x; the variables that char-
acterize the exchange of mass, energy and information with the environment
by means of control system hardware and software are termed 'interactive
variables' and denoted by y; the variables that characterize the quantity of
supplied mass, energy and information are termed 'input variables' and de-
noted by u. The models of real processes contain:
equilibrium equations such as equations (6.16), (6.17) or (6.18);
process state equations (see Chapter 2);
inequality relations that include limitations and specificities of the given
process (see Chapter 1);
probabilistic data (see Part Three).
Table 6.7 Models used in process control systems
Static model Dynamic model
Deterministic Stochastic Deterministic Stochastic
Material model copy geographic chart with

a distribution of
process
communication
microlocations
semicopy chart of a system consumer simulation presentation of system
communication chart consumption state
network
layout of a plant
plant isometry
scheme
analogy wall chart of a plant plant mimic diagram mimic diagram of presentation of
(stochastic part) system variable stochastic system
plant mimic diagram
model failure model
(static part)
Symbolic model descriptive rules for system report on system state report on system expert system for plant
governing variable trend start-up
imitative table of system table of consumer table of process instruction for
regular states policy policy emergency case
mathematical formal Ohm's law waiting queue of Navier-Stokes Bayes' rules
product transportation equation
The quantity characteristics of a given space and process are termed 'process
parameters', which influence the formulation of a model shape.
The mathematical model of a real process can thus be expressed as func-
tional L
L{ x(t), yet), p(Vi> t)} =0 (6.19)
where x(t) are state variable values, yet) are interactive variable values and
P(vi' t) are process parameters. Terms Vi (i = 1, 2, 3) are space coordinates
and t is a time variable. A static model of the real process is characterized by
dm/dt = dE/dt = dl/dt = 0 and equation (6.19) is simplified as a functional M
of the form
M{x, y, P(Vi)} =0 (6.20)
The process parameters p determine principally the relation between x and y.
The process parameters p are almost always deterministic quantities contain-
ing a small stochastic component like, for example, pipeline friction coeffi-
cient. Thus, a separate presentation of some examples of stochastic and deter-
ministic models (as applied to the process control systems in Table 6.7) is
taken as realistic only for cases where a deterministic model can be separated
from the stochastic model of a process.
The other part of the extended real-time software supports operator-system
interactions of a set of programs and routines that enable the presentation of
the process state and process variable trends in a way that is given in Table
6.7 as, e.g., consumption presentation, mimic and trend diagrams, failure model
presentation and system state and system variable trend reports.
The extent and amount of programming needed for the application of ex-
tended real-time software are estimated as being about ten times the efforts of
the complete process model software presently available. The main reason is
that additional adaptation work is needed. For instance:
fonnation, formulation and additional programming efforts for establish-
ment of the appropriated database of extended real-time software;
algorithmization and formatting of system tables for approximate operator-
system communication for the fulfilment of extended real-time functions;
adaptation of the existing computer system to the requirements of the
model software. The usual requirement for computing equipment con-
sidering different process models is given in Table 6.8.
A static model based on algebraic relations between the process variables may
suffice for the description of a stationary state of the process and the optim-
ization of plant operation, when plant response times are negligible versus the
minimum period of significant external perturbations. The model can be suitably
applied for the optimization of electroenergetic and water supply systems, and
partly so for natural gas transportation systems, for thermoelectric power plants
Table 6.8 The usual requirements for computing equipment
Type of model Mathematical Computing Use

formulation equipment
Static Algebraic equations Microcomputer or Long-term

linear and/or small to medium optimization of
nonlinear minicomputer process or plant
operation
Dynamic linear Linearized Medium to big Stability analysis
differential minicomputer and control system
equations design
Dynamic nonlinear Nonlinear Big minicomputer Implementation of
differential Mainframe complete control
equations and logic computer system and operating
operations procedures; operator's
training
and most chemical processes. However, to obtain significant optimization results,

a very accurate static model is required both in the structure and in the quan-
titative definition of process model parameters.
A dynamic model of the process or plant is based on algebraic and differ-
ential equations of plant variables versus time that are almost always nonlinear.
For the purpose of stability analysis, these equations may be linearized around
the process steady-state point obtaining a linear dynamic model of the pro-
cess. A dynamic nonlinear model of the process or plant may include the
simulation equations of, for instance, steady-state operation, start-up, shutdown
and safety actions responsible to cover all the plant operational situations
continuously and to interact easily with real-time data and operators.
Extended real-time software is sometimes the only possible way for devel-
opment of a new production process or the control system implementation
on already-installed processes. The results of preliminary process models can
predict some features of planned production systems and the effects of planned
control systems. Process simulation facilitates:
an aid to basic process research;
designing the pilot-plant;
designing the process control system;
defining operating procedures;
testing the control system prototype;
operator training;
experimental tests on a new plant;
rough calculation of the economy and availability of the plant.
The model of a field F of certain properties, like temperature, chemical density,

neutron density, of a moving fluid is described by the following second order
partial differential equation
aF/at + v grad F - D AF = f(F, x, y, Z, t) (6.21)
where v is fluid velocity, D is the diffusion coefficient for a property con-
sidered, f is an external forcing function. For a fluid with a very low rate of
diffusion so that the diffusion term in equation (6.21) can be neglected with
respect to the transport term, and for a case when the fluid motion occurs
mainly in one spatial direction x, and when the fluid motion is of the turbulent
type as is the case of a channel with a high flow rate
D=O (6.22)
F = F(x, t)
and equation (6.21) becomes
aF/at + vaF/ax = f (F, x, t) (6.23)
Equation (6.23) is a first order partial differential equation with two independent
variables and, although simplified, this equation is usually adequate to describe
transport phenomena in most plant and process components. Solving equation
(6.23) usually requires the use of a computer, but a typical solution, with the
given initial and boundary conditions, can be represented, for example, when
using a characteristic method as a series of calculating steps in discretized
time and space variables (see Figure 6.21). The corresponding equation for
the fluid mass flow is also given in Figure 6.2l.
Thus, the modelling of a process requires a large portion of computer power
for model arithmetic operations. Therefore, the speed of execution of arithmetic
operations is the most crucial parameter for the estimation of the feasibility of
computers for model calculation. According to the information available, the
number of calculating steps is about 500 for a single natural gas network node
(Zelenko, 1976). A total time requirement for ten pipes of a natural gas network
calcuhlted for one hour of gas flow dynamics requires about 3 x 108 calculating
steps. Since there are required integer operations on real numbers and double
precision operations over operands, the differences between the speed of ex-
ecution of arithmetic operations can be essential for the selection of a process
computer. Usually the back-up computer at the dispatching level is used for
on-line model calculations, since the working computer already has a high
workload. Off-line model calculations are performed on mainframe comput-
ers. The differences between mainframe computers and process control
computers, due to larger words for the floating point representation of num-
bers in mainframes, can result in a bigger round-off error in process control
computers. The error can be decreased by a carefully designed simulation
algorithm.
t = t max
I-:r--
ti +! t
ti
U
-r
I
F{x.1 tt~-
i) I
:i
jlF (xi+!' t i+1)
-·--------+F-'R-l-.G~\=-------F--="..:..R.:...'-'G..:...,--=,: . . - - - - - - -
-
- ' - -
it L-----'
- x
------"----------1 x=l
F = (F'+F")/2+[v.(F'-F")+(C~-v~)(R'-R")-(C.-v.)G"~t-(C.-v.)G'~t]
1 1 1 1 1 1 1
2C.1
where Ci - sound speed in a fluid
vi - fluid velocity
Figure 6.21 The series of calculation steps for gas transportation model: R fluid
density; G, friction coefficient; F, fluid mass flow.
The communication between a simulation model in extended real-time work

and other real-time programs requires the following:
compatibility of connections between the model and the real-time system
data;
topological correspondence of the model and the modelled process;
modifications of model structure and parameters to facilitate model de-
velopment and parameter convergence.
A simple communication between real-time and extended real-time software
REFERENCES
ASEA Information (1978) YLDA 78-08.

Bailey, S.J. (1988) Control Eng., July, 75.
Bodlovic, P. (1977) Koncar Informations, 24, 1, 9.
process
process data
system
operator
process
,
•
acquisition topology
determination
real-time
process state software
estimation extended
real-time
software
extended real-time
database
process calculation determination

of process of optimum
security process
control
analysis alternatives parameters
operator
communication
programs ,• process
operator
Figure 6,22 A simplified communication between real-time and extended real-time

software.
Borevkovic, B. and Cvitas, Lj., (1983) Koncar Strucne Informacije, 30, 1-2, 47.
Debeljuh, V. and Vranic, P. (1982) Koncar Strucne Informacije, 29 1-2,4.
Heninger, K. (1980) IEEE Trans. Software Eng., SE-6, 1, 2.
Jovic, F. (1989) Proc. MIPRO '89 Conference, Opatija, PU Section, 6-1.
Koncar, R (1980) Documentation of the Process Control System DS-80.
Laduzinsky, A. (1982) Control Engineering, Nov., 64.
Leinbaugh, D. (1980) IEEE Trans. Software Eng., SE-6, 1, 85.
Lichtberger, Z, (1990) Personal communication.
Lundberg, M., Goldkuhl, G. and Nilsson, A. (1979) Inf. Systems, 4, 1.
Martinovic, A. (1983) CEP, Feb., 72.
Skeen, D. and Stonebraker, M. (1983) IEEE Trans. Software Eng., SE-9, 3, 219.
Tanenbaum A.S. Operating Systems, Prentice-Hall, Englewood Cliffs, NJ.
Tunon, J. (1982) Control Engineering, Oct., 110.
Wolverton, RW. (1974) IEEE Trans. Comput., C-23, 615.
Zelenko, B. (1976) Informatica Proceedings, Bled, 5, 111.
Part Three
The Man-Machine Interface
7
Reduction and visualization of
data and procedures
7.1 INTRODUCTION
Process control systems are developed, designed, used, maintained, repaired

and abandoned, or redeveloped, by and for humans. Production processes are
therefore a series of interactions: process--control-system-human--control-
system-process. The integration of humans into the production process should
be planned, adequately investigated and developed. To this end, a detailed
study of human use of a system must involve the aspects and characteristics
outlined in Table 7.1.
Certain characteristic human roles are recognized in system usage: system
operator(s), process operator(s), maintenance staff, training staff. Part Three
emphasizes the role of humans in process control functions in production (i.e.
on process operator or dispatcher functions and facilities).
The function of a process operator or dispatcher is to manage the system,
keeping it in balance, and produce the product at the right time in the correct
quantity and in an appropriate and safe environment for both humans and
equipment. The operator or dispatcher has numerous subsidiary managerial
roles: communicating with other humans involved in a process, controlling
various process states and occasionally reporting or consulting other services
and teams. The basic task is to make decisions on production that can be
supported by:
sufficient information on everything operating normally, or on places where
abnormalities are present;
certain records of events that have occurred, knowing where the equip-
ment is installed, what devices have obtained commands, what part of the
process is supposed to be in a certain state;
information on the process variable states and trends to determine whether
the process is working properly;
facilities to issue necessary commands and information on the results of
command execution.
Additional features, as listed below, can help the operator or dispatcher perform
his/her function in the best way.
231
Table 7.1 Aspects and characteristics of human use of a process control system
Human use aspects Characteristic aspects
Estimation Psychology Anthropology Mechanics Industrial Education

Acoustics design
Optics
System user - basic description + + +

Engineering of human use of the + + + + +
equipment
Requirements on human knowledge in + + +
system usage
Training concepts + + + + + +
Technical manuals + + +
Testing of system - operator work + + + +
Operator-system interfaces and facilities + + + + +
Maintenance aspects + + + + + +
Operator-system communication 233
1. Fast reaction of control equipment, and a real feeling of being present at

the process site.
2. Minimum noise in the control room.
3. Appropriate seating accommodation.
4. Proper lighting.
5. Minimum outside interferences, especially by telephone.
6. Moderate or light clerical work (e.g. reading or preparing reports).
An operator's work demands sensory perception, selection of signals, long-
term and short-tenn memory, control of data, delivery of conclusions and
decisions, and types of manipulation. This type of intellectual work includes
the use of research techniques to detect disturbance sources and to make
decisions and intervene if necessary. The amount of infonnation must be adapted
according to the operator's ability.
Visual perception is the most effective human communication channel. Thus,
communication between the control system and the operator or dispatcher is
established by means of a rationally reduced number of process variables.
7 .2 OPERATOR-SYSTEM COMMUNICATION
The problem here is one of data reduction and it is useful to consider the
following systems on the basis of their size, since this affects the amount of
information to be presented to the operator.
1. Small control systems, where the number of status and counter state data
is in the order of 200 or less, and the number of analog data is 100 or
less.
2. Medium control systems, where the number of status and counter state
data is between 100 and 2000 and the number of analog data between 50
and 1000.
3. Large control systems, where the number of status and counter state
data exceeds 2000 and the number of analog data exceeds 1000.
Large control systems demand a special communication between operator and
control system, and are often used, for efficient process control on the operator's
side, as a set of medium control systems tenned 'functional systems'. Small
control systems usually do not present any special problem to the operator-
system communication, i.e. the number of data presented or the number of
commands issued. The basic characteristic hardware functions of medium control
systems concerning the reduction of data number, visualization of data and
the issuing of commands are:
exchange of infonnation between operator and process by the central
computer and peripheral equipment;
logging of process data and historical process data performed by the central
computer and peripheral equipment;
234 Reduction and visualization of data and procedures
print-out of the process data and the documentation of process states by

the central computer and peripheral equipment;
operator commands issued by peripheral units of the process control system;
automatic protection actions performed without the operator's interven-
tion, and the information on these actions presented to the operator.
Figure 7.1 shows the scheme of a simple configuration that enables the
implementation of these hardware functions.
Monitoring and control of production processes involve the following
operator's actions.
1. Sensory perception.
2. Signal discrimination.
3. Short-term and long-term memory.
4. Mental data processing.
5. Decision making on the basis of data processed and on short-term and
long-term memory.
6. Some sort of manipulative action to implement certain decisions.
A simple reaction time of the operator is the delay between the occurrence of
a single fixed stimulus and the initiation of a response assigned to it. The
reaction time Tr to the information presented to the operator is given as
(Dallimonti, 1976):
Tr = a + bH (7.1)
where a is typically 0.25 s, b is typically between 0.25 s2/bit and 0.33 s2/bit
and
H =W ld A2/E2 (7.2)
where H is the input information rate (bit/s), W is the input information signal
bandwidth (Hz), A is the mean square root of the change of input signal
amplitude and E is the mean square root of the reading error. Thus, equation
(7.1) gives the simple reaction time of a human to a given amount of input
information H. Equation (7.2) gives the amount of information generated by
an instrument or by a display.
The complex reaction time of the operator is given by the data in Table 7.2.
The estimate of perception, recognition, decision and action time, T PA (ms),
can be reached from the relation
(7.3)
where n represents the number of logic discriminations required before reacting
(Woodson and Conover, 1966).
When a number of instruments and displays are continuously scanned by
an operator, the short-term memory of the operator may present serious
problems. The short-term memory is the ability of a human to recall information
)1•
system
( operator
central
process (~,)
me~ory
umts
Loperator
system
cons0!s--
unit
A I controller
system bus
I 1
"
fast
I I
semigraphic
memory colour
unit console
bus/bus •
• ( productio
converter
II planning a
supervisio
process unit bus
I
data
I
data
I
process
input output input!
units units output units
f•
•
I ~
process process
command signalization
,
console console I
• to /from process
devices/sensors
process
operator
Figure 7.1 A scheme showing the basic characteristic hardware functions.
that has just been acquired but has not yet become part of the long-term
memory. The work by Peterson and Peterson (1959) indicates how simple
words and three words are recalled after short time intervals (Figure 7.2).
This ability of humans emphasizes the need for simplification and reduction
of data presented to the operator. The reduction of process data requires more
Table 7.2 Operator reaction times
Brain perception of what the eye sees D.1 s

Brain recognition D.4 s
Decision making 4-5 s
Action D.Q1-1 s
correct recall
(percentage)
100
'\\
\ one word
\ --____~~73~o/.:o_____71% 71%
'q68% 0------'0
\,
50 \\
~%
""""o.! 5% three words
___ 20% 18%
-<>-----c:>-_ 16%
----0
3 6 9 12 15 18
retention time (seconds)
Figure 7.2 Correct recall of one or three words after short time intervals.
efficient coding of useful process information with the representation of graphic

patterns that relate to acceptable and unacceptable process states. Graphic
pattern and graphic presentation must reflect the visual and audio perception
of the operator. Data on human visual and audio perception, together with
relevant data processing functions, are given in Table 7.3.
This ability to work also depends on the physical condition of the operator
(see Figure 7.3). The reaction time of an operator under the influence of
alcohol is slower if the operator is also alone and on duty during a holiday
period.
Table 7.3 Visual and audio paception of the operator
Visual perceptiuon Audio perception
Range of signals accepted Light Sound

380 nm - 780 nm 10 Hz - 15 kHz
red-violet
_10" - +50" from optical
centre
Peak sensitivity 540 nm yellow-green 1 kHz - 3 kHz
Resolution angle minute* 10- 16 W/cm2 - 0 dB
1-1.5 at ± from centre at 1000 Hz
10-3 W/cm2 - 140 dB
Information flow through 2.10 8 bit/s 4.10 4 bit/s
sensors
Conscious data processing 50 bit/s
Reception sensitivity 570 relative different 325 relative different
intensities at white light intensities at 2000 Hz
* corresponding to a resolution of 0.5 mm at 1 m
The basic pattern of human data processing can be schematically given as

presented in Figure 7.4, where short-tenn and long-tenn memory are included.
In process control application, operator-computer communication can be
divided into three distinct levels:
1. Process control operations.
2. Process engineering.
3. Computer programming.
The differences between these applications lie in the speed and frequency of
communication, in the method and amount of data presented, and in the level
of the user's knowledge, responsibility and education. Typical communication
functions and communication channels for these three levels are outlined in
Table 7.4 which also refers to their hardware facilities.
The type and relative locations of infonnational devices for the control of
processes usually maintain the basic objective to provide the capability to
manage the plant by a single operator, or crew, in all nonnal and abnonnal
situations. In large and medium control systems, two principles are involved
to provide this objective:
1. Split-up of process control functions and all other noncontrol functions
into subsets that can be separately observed at any time.
2. Hierarchical distribution of discrete control functions into a multilevel
organizational structure.
percentage difference to
physiological average value of
working efficiency
%
60+----------r--------~----------+_------~
40+-~~--~~--------~----------+_---------1
40+---------~--------_+----------~~~~~1
60T----------r---------+--------~~~~~~
\ ,
, ....
80----------~--------~--------~------~
6am 12 6pm o 6am
time of day
Figure 7.3 Physical condition of the operator:-, physiological; ----,

effective (motivation ± 10%); ----, alcoholic (0.02% alcohol).
The control levels or priority levels of a power station are illustrated in Figure
7.5. For instance, the control of the power station is handled at the supervisory
level enabling the coordination of power plants (e.g. monitoring the overall
station efficiency, maintenance, staffing) and the control of separate power
plants involves the control of the entire plant as a system such as plant start-
up, plant load change policy. The control of plant subsystems enables the
efficient and quick operation on separate plant's subsystems (e.g. burners,
cooling system, heat exchangers) and system component controls enable the
separate operation of each plant component (e.g. pump, motor, valve).
Organization of hierarchical levels of data presentation usually varies between
two and four levels of data presentation.
correct repeating of data associative memory
direct after reception often refreshed
Figure 7.4 The basic pattern of human data processing.
Table 7.4 Communication functions and channels for the three control levels
Process control Process Programming level

operations level engineering level
Functions Process control by Gathering of data Input of system

exception for process programs
Determination of evaluation Trouble-shooting
the reason for Entering constants Maintenance
process deviation and control Systems expansion
Immediate process equations Development of
action Implementation specific process
Acknowledgement of application control algorithm
for the process programs and Development of
action control systems specific supervisory
Data logging programs
Event recording Program dumping
Bulk reloads of the
system
Communication Display read-outs Line printer I/O type unit
channels (units) CRT consoles Printer Diskette unit
Special purpose Diskette units Printer
keyboards Special purpose
Printers keyboards
CRT consoles
power
station
power station control
functions
other functions
power
power plant 1 power plant 2 plant
control
functions
system
plant plant plant ••• control
subsystem 1 subsystem 2 subsystem 3 functions
component
component 1 component 2 component 3 control
functions
Figure 7.5 Hierarchical levels of control functions.
First level This keeps the operator informed that the system is working without
perturbation and what has to be searched for with the highest priority in the
case of perturbation.
Second level This exhibits the structure of the process, or plant, that is
controlled, indicating the principal energy and material flow and processing
states and including an alarm summary with the last received alarm.
Third level This enables the operator to see those process details that actually
exist and those process and alarm states that have actually happened.
Fourth level Additional data are shown for the supervision and control of
specific process components, e.g. analog values, limit values and zone values
of a specific process variable.
Table 7.5 details a typical four level process data presentation for normal
and alarm process conditions.
An important view of operator-system communication is established by
report issuance that serves the process state documentation, and the initiation
of the necessary process state calculations. The report issuance can be divided
Visualization and data logging 241
into process event documentation and collective process data issuance (Table
7.6).
7.3 VISUALIZATION AND DATA LOGGING
The visualization of all pertinent process control data is performed by alphan-

umeric, semigraphic and graphic display units, and purpose-built operator
consoles and mimics.
A survey of non intelligent alphanumeric display units is given in Table 7.7,
and the main features of semigraphic and graphic display units are outlined in
Table 7.8.
The print-out of all pertinent process control data and the logging of data
is performed by a printer, line printer or other hard copy unit.
A survey of printers and line printers, according to their speed, is given in
Table 7.9. The main features of low speed printers are as follows.
1. Printing speed from 15 characters per second to 600 lines per minute.
2. APL, ASCII, Baudot, EBCDIC, Katakana and OCR character code sets.
3. Switch selectable printing speed.
4. Incremental and automatic bidirectional printing.
5. Program control multiple fonts.
6. Half-sp'ace matrix fonts and condensed printing.
7. Uppercase and lowercase printing.
8. Double-width printing.
9. Multicolour printing.
10. Full and multiline buffers.
11. Double-strike printing for camera-ready copy.
12. Wider platens and extended carriages.
13. Power stacker and power hood.
14. Graphics plotting.
Thus, visualization and data logging can be broken down into two areas:
hardware equipment for visualization and data logging;

protocols for process data visualization and data logging.
Since the same protocols can be implemented on different equipment, the
main issues in visualization and data logging are protocols for man-machine
communication. These protocols are implemented differently for various
processes, since the main task of process data visualization and data logging
is to enable the best direct communication of the operator with the process,
and the fastest possible decision making by the operator.
The main point in direct communication to the operator and the process is
to enable the stratified approach to data presentation and print-out. The most
Table 7.5 Four level data presentation for normal and alarm conditions
Information type Normal process conditions Alarm conditions
Information level Type of display Display hardware Information level Type of display Display hardware
Status of process Permanent display Alarm is always Acoustic indicator,

connection and spontaneously Gong, CRT display,
configuration state: indicated in the Operator's table
hardware Elementary Operator's table form of acoustic
connecting CRT display signal, general
elements alarm indication,
(position of valves, flashing light of
breakers, etc.) the object on a
CRT screen
bulk process 2 Compressed CRT display
hardware
elements
(transformer
sections,
manifolds,
vessels, etc.)
indicators (state 3 Complete on call CRT display 3 Clear text on CRT display
sensors, switches, alarm field of
etc.) CRT screen
Analog values:
process primary 1 Alarm indication CRT display
values such as 2 Not displayed when passing of Operator's console
voltage, current, 3 limit value
flow, level, etc.
process globai 4 Completely CRT display 3 Text indicating CRT display
values such as net displayed Measuring the cause of
net frequency, Permanently instrument perturbation on
valve position, etc. displayed Analog recorder CRT alarm field
Possible selection Numeric indicator
XY plotter
Counter state, 4 Possible selection Measuring Usually not
Increments instruments displayed
Counter Display
Table 7.6 Types of report
Information type Content Availability
1. Events All operator actions such Spontaneous at the moment

Exploitation events as commands, alarm of appearance
quittings, all process
spontaneous changes,
signalizations alarms,
passage through limit
values
Alarm events All alarm conditions On demand
Process configuration Overview of all process On demand
connection in a real
configuration
Event recording Print-out of perturbations Automatic when provided
in chronological order with event-recording
On demand when event-
recording in a mass memory
2. Measuring values
Communication of Issuing of measuring and Periodically or on demand
measuring values limit values
Supervision of Issuing of selected Periodically or on demand
measuring values measuring values
Statistics Recording of mean and On demand and often
maximum values graphically on CRT screen
or on analog recorders
3. Counter state
State of process All counter states Periodically or on demand
variable counters
important fact in fast decision making by the operator is to provide exact

information about the useful information on the process state. Thus, if the
operator is provided with a better protocol for the presentation of the useful
information on the process state, the overall process control exhibits a higher
information efficiency Ie that can be presented as
(7.4)
where ST is the total number of possible different process states to be presented
to the operator that completely describes the process, and Su is the number of
different process states that are actually presented to the operator and that
differ from normal state; the operator then decides what is useful information
Table 7.7 A survey of nonintelligent alphanumeric display units
Terminal features Basic characteristics Quantitative relations of basic

characteristics
Control unit maximum number of displays 1 (66%),32 (32%), other (21%)

standalone 77%
configuration
} 7% (both)
clustered
configuration 16%
Display unit screen size (char) 1920 (43%), 960 (16%),480
(10%), other (31 %)
characters per line 80 (76%), 40 (12%), other (12%)
lines per display 24 (50%), 12 (20%), 10 (8%),
other (12%)
retention and recall of display no (75%), yes (19%), other (6%)
content
character set (char) 64 (43%), 96 (28%), 128 (22%),
other (7%)
character generation 5 x 7 (53%), 7 x 9 (20%),
other (27%)
Keyboard control character typeover yes (86%), no (14%)
functions character insert/delete yes (57%), no (35%),
optional (8%)
line insert/delete no (48%), yes (42%),
optional (10%)
screen segmentation no (66%), yes (30%),
optional (10%)
text scrolling yes (61 %), no (34%),
optional (5%)
Data transmission communication line voiceband 32%
voicegrade 32%, other 23%
narrow band 13%
line speed 110 bps - 960 bps 33%;
different 67%
line mode HDX/FDX 64%, HDX 22%,
other 14%
line code transmission timing ASCII 77%, other 23%
asynchronous 60%, asynch/synch
18%, other 14%
Error control error detection parity 50%, VRC, VRC/LRC

22%, CRC 8%, other 20%
error correction autotransm. 29%, manretrans.
24%, manual 21 %, no 12%,
other 14%
The numbers and percentages are based on the US and European market. VRC, vertical
retransmission control; LRC, line retransmission control; CRC, character retransmission
control; HDX, half duplex transmission mode; FDX, full duplex transmission mode;
SYN, synchronous transmission timing; ASYN, asynchronous transmission timing
Table 7..8 The main features of semigraphic and graphic display units
Features Audio response Semigraphic display Graphic display unit

systems unit
Assisted computer minicomputer microcomputer graphic processor

system (ASC) or ~ 64 kbytes refresh ~ 2-4 kwords
controller
Connection to ACS parallel interface integrated into the integrated into the
or controller system system
serial interface serial interface
RS232C
I/O bus connection (HDX, FOX)
300-9600
Minimum-maximum 10-10000 words 64 characters 95 ASCII + 27
number of unlimited 64 special generated symbols
characters, points or < 300 phrases symbols
words expressed 20--140 lines per unit
60-80 char/line
graphic matrix
1024 x 512 dots
512 x 256 dots
Selection of words/phrases 4-64 intensity levels 16-24 intensity
expression selectable 7-9 colours levels
foreground 7-9 colours
(7-9 colours 240 character sizes
background)
Expression timing word length 0.5 s 60 refresh/s 5 ~l.s/character
phrase length 1.6 s < 40 km/s vector
moving rate
Expression digital compressed dot matrix vector drawing
generation technique analog, recorded on
film strips, speech
synthesis
Options and control selection of words zoom typewriter, cursor
pad
selection male/ selective erase 128: 1 zoom
female
any language reverse video depth cueing
typewriter ~ 128 additional
user-defined
characters
cursor pad
automatic plotting hardware-generated
circles and arcs,
console devices like
control dials,
joy-stick, digital
tablet additional
refresh buffer
Table 7.9 A survey of printers
Printer type Printing speed range
Serial matrix xxxxxx (30-300)

Distributed matrix xxxxxxx (100-600)
Impact 1 Serial character xxxxx (10-100)
Line printer character xxxxxxxxxxxxx (30-2000)
Laser xxxxxx (100-300)

Thermal xxxxxx (10-100)
{
Nonimpact Electrostatic xxxxxxxx (1000-7000)
Xerographic xxxxxx (8 000-20 000)
Ink jet xxxxxxxxxx (30-500)
Mini and microcomputer

applications xxxxxxxxxxxx (30--1000)
Speed (cps) 10 30 100 300 1000 3000 10 000 30 000
LS
VC
FS
~ TT
---
-- --
- -- --- -
FT
Figure 7.6 A simple hypothetical basic process unit: VC, valve command; FT, flow
transmitter; IT, temperature transmitter; FS, flow switch; LS, level switch.
Table 7.10 Total number of process states
Process Number of Number of Total number Total number of

variable normal process alarm states of variable possible new states
states states (deviation from
normal state)
Valve 2 2
command
Flow (high) (nonnal to high)
transmitter 2 3 2
(low) (nonnal to low)
Temperature (high) (nonnal to high)
transmitter 2 3 2
(low) (nonnal to low)
Flow switch 2 2
Level switch 2 2
Total number of process states: ST = 2 . 3. 3. 2. 2. = 72

total number of possible new states differing from nonnal states: Su = 7
on the process state, in order to make a decision on the presented abnormal

state. The example of the calculation of Ie for a particular simple, hypothetical
basic process unit can be based on the data given in Figure 7.6. The total
number of possible process states given in Table 7.10 equals the product of
all process variable states. For instance, suppose that the presentation of the
change of state is sufficient for the decision on actual process state. Since
seven possible change of state events can be expected, supposing the process
is in normal state concerning the analog signal values, the information effi-
ciency Ie for this example equals 72/7 = 10. The transitions to alarm process
No Name 01 Measurement Place Measured Planned Dill· Date
1 Output Flow 1 3100 5200 2100 08:05

2 Output Flow 2 5500 9500 4000 07:04
30 Output Flow 30 0 0 0 15:36
31 Output Flow 31 2500 2500 0 08:21
60 Output Flow 60 50 100 50 08:19
Flows in natural-gas network

State onJuly 03-81. Distribution Plan 1
Figure 7.7 An example of a process data print-out.
No Name Pressure Time
60 Output Pressure 1 47.0 07:55

61 Pressure Production Unit 1 47.0 07:55
Pressures in natural-gas network

State onJuly-3-81
Figure 7.8 The usual format of the alphanumeric video display data.
states can be supported by the time tag, so that the obtained print-out can be
used for official supervision and arbitrations. Fast, complex and mutually in-
terconnected processes (e.g. electrical energy generation, transmission and
distribution systems) are then provided with such functional recording systems
usually termed chronological event recorders, that are primarily used for the
time detection and the time-of-event discrimination. Such chronological event
recorders are usually installed for a particular process control. For the large
interconnected systems exact time distributions have to be provided to satisfy
the demand on precise time-of-event control in a process.
The data on the process state and on the change of process state appears on
the screen of the operator's console usually iri order of appearance to enable
faster decisions by the operator and production supervisor. An example of the
process data print-out is given in Figure 7.7. The visualization of process data
on alphanumeric video displays greatly resembles the print-out formats (Figure
7.8), but it can be modified to present data in different forms like those given
in Figure 7.9.
The most efficient data visualizations are obtained by using semigraphic
and graphic video terminals where process mimic diagrams can be shown and
NETWORK INPUTS 03 - JULY 81
CONS 1 PROD 1 PROD 2 CONS 2
-------0 -------o--------o---------o----------q------------o
I P = 43.0 P = 0.1: I
l F = 9820 F =0 I : PROD 5
I I
PROD 4 I I
I
r---9 PF == 43.0
53928
, I
lI P = 42.0
I I
PROD 3 I r-----o-- F = 2110 I
---0--------- 0 I :
P = 28.2 I I I
F = 67514 I <? ----------------- <? CONS"3---------- ~
I
I
PROD 6:
P = 43.0 O----~
F = 5450 I
I
II
, I
I I
I
I • I 1
b- ---0----------------- 0 ----Q----Q---0---.L.---.Q PROD 10
CONS 4 PROD 7 PROD 9 P = 16.0
P = 42.0 P = 41.0 F = 2269
F = 11412 F = 774
Figure 7.9 The pseudo mimic formatting of data on alphanumeric video display:
PROD 1, natural-gas production point 1; CONS 1, natural-gas consumption point
1; P, pressure (bars); F, flow (cubic metres per second).
dioptres
t
accommodation
28 year distance
8T-----------------~C------------------------- 12.5cm
6
16cm
41 year
4
25cm
2
50cm
1 2 3 4 5 time
(s)
Figure 7.10 The accommodation step response with time and age of operator for
yellow-green colour (Krueger and Muller-Limmroth, 1981).
provided with up-to-date process data. Such an interface enables the presentation
of the process to the operators in a way that they can:
see the actual structure of the process and the interactions between specific
basic process units mainly by means of the static part of the mimic
diagram;
estimate the quantitative relations between process variables mainly by
means of a variable part of the mimic diagram that presents actual and
historic process model data.
The ability of a mimic diagram to enable the fast magnification of process
details by calling mimic diagrams of specific process parts of complex processes
increases the speed of decision making by the operator in the case of vast
numbers of process data. The efficiency of visual presentation can be increased
by approximately 40% in the information content (Smrkic, z. (1962) personal
communication) by a colour video display unit. The human eye has different
reaction times, depending on colour. The cause is possible due to chromatic
aberration where corrections of about +2 dioptres is needed for red, and -2 for
blue (Figure 7.10). Thus, yellow-green video display screens are preferred for
processes where fast operator's reaction is expected.
Table 7.11 Some common colour codes
Features Colour
Red Yellow Green
Positive associations Life, warmth, Sun, light, Vegetation, nature,

passion, valour, intuition, intellect sympathy,
sentiment prosperity, hope
Negative Spilled blood, Treachery, Death, lividness,
associations burning, wounds, cowardice envy, disgrace
war
Most common Colour of joy and Quarantine colour Feminine (American
cultural meaning festive occasions (medicine) Indians)
Fighting, anger, Emperor colour Fertility, vegetation
danger (China) (Egypt)
Marriage colour Happiness (Egypt) Youth, energy
(folklore) Gay (Japan) (Japan)
Protection from Sensationalism Jealousy, envy
death (folklore) (journalism) (language)
Fever or protection Sterile (medicine)
from disease
(folklore)
Holidays
Most common Flammable gas Transistor emitters Control grids wiring
technical meaning or liquid (electr. engin.) chassis (electr.
Stop (traffic) Heating circuits engin.)
Plus of power Oxidizing agent Nitrogen,
supply (electr. Radioactive matter compressed gas
engin.) Caution Clear or go signals,
Roads, telegraph (autoracing) permission (traffic
lines (geography) Waming of danger and safety)
(safety) Oil
General warning
Gas
Colour preference Second (adult) Eighth (adult) Third (adult)
Fourth (child) First (child) Seventh (child)
Colours possess a strong influence on human behaviour and therefore colour

video displays are preferred for many process control applications when
appropriate colours are used. Some common colour codes and their meanings
are given in Table 7.11. The use of colour for actual process control purposes
is given in Table 7.12.
7.4 RECOGNITION OF PROCESS STATES

Operator-system interaction is achieved by:
acquisition of initial data for data processing and manipulation;
Recognition of process states 253
Colour
Blue Purple White Grey Black
Sky, day, sea, Power, Day, Maturity, Mighty, stark,

thinking, spirituality, innocence, discretion, dignified, night
devotion, truth royalty, empire purity humility
perfection
Dark blue Sublimation, Spectral, Neutralism, Morbidity,
denotes night martyrdom, ghostly, cold, egoism, despair, night,
and stormy sea, regret void depression, evil, sin
doubt indifference
Virtue, faith, Worn by Virtue and Old, mature Death, winter
truth (Egypt) figures (China) purity (fashion) (psychology) (China)
Ghost, fiend Virtue and Death (with Wisdom Water (China)
(Japanese faith (Egypt) black) (Judaism) Black cat
theatre) Rage White flag (supersition)
(language) (surrender, Evil, despair
Winner (prizes) peace) (language)
Morbid
(psychology)
Transistor Power supply Bias supply AC power Grounds on

collectors minus (electr. wiring chassis lines (electr. wiring chassis
(electr. engin.) engin.) (electr. ~ngin.) engin.) (electr. engin.)
Caution on Radiation Regulation Steam (piping) Contours
repair (safety) hazards (traffic) (geography)
Protective Valuable Corrosive
materials materials material
(piping) (safety) (handling of
Sea (geography) goods)
First (adult) Sixth (adult) Fourth (adult)

Sixth (child) Eighth (child) Second (child)
manipulation of data as a real-time dialog between the operator and computer

systems;
presentation of data and results.
Most process data are acquired by computer process interface. These are data
on process variables and states such as temperatures, pressures, flows, voltages.
The lowest level process data are transferred to the computer from different
existing documents in computer-readable form, mainly by key-in operations
by the operator. These are data on changing conversion units, limit values,
timing restrictions and various comments on process behaviour.
The dialog process consists principally of sending and receiving messages
Table 7.12 The use of colour in process control
Data priority Process status Colour
Highest priority the part of process chosen for white

operator's control
refreshment of data blocked violet
process data not yet accepted light blue
unaccepted alarm red flashing
equipment failure red
process control function yellow
blocked
Lowest priority normal process state green or dark blue
between the operator and the computer via an interface. The quantitative measure
of merit of the dialog can be calculated when the content of information sent
by each of them is compared to the content of information received by any of
them. There are crucial points in this dialog design:
1. Creation of environment for the user to use the computer properly.
2. Creation of safe and comfortable dialogs. This implies, for example,
standard keyboard style of terminals, the same special characters all over
the plant control system, dangerous commands to be typed character by
character.
3. Involvement of the user into system and dialog design process.
The goal of a dialog is the recognition of process states. The recognition of
process states is based on previous knowledge of the process state
interdependencies, as opposed to the recall process by an operator. The result
is a higher percentage of retention of the operator's knowledge of process
states (Figure 7.11). The success of dialog therefore depends on a menu-
driven dialog, user involvement in the system and dialog design process and
on available computer graphics. The design of an operator system dialog based
on Pascal, such as CCITT language, named MML (man-machine language)
(Hombach, 1982), facilitates the operation and maintenance of computerized
A fully successful operator-system interaction can be designed for coarse
processes that change slowly with plenty of informational feedback. There are
environment critical processes that degenerate fast and with much interdependent
complexity where operator's control is impossible, as shown in Figure 7.12.
To enable operator control, an adequate design of basic process units, computer
hardware and software must be made (ASEA-ATOM, 1982) and expert system
engineering included (ABB, 1989). Using knowledge base engineering methods
a further reduction in data volumes is achieved by pointing from process facts
retention(%)
100
90 .--~
80
70
60
50
40
30
20
10
20 min 1 h 2h 24 h 48 h
interval following learning
Figure 7.11 Retention of knowledge versus the interval following learning.
y expert
system
,
support
all manual ; ~
control I ~ control concept
limit 1// change
high
I
I
I
I
I
I
I
I
I
I I
increasing control
difficulty
low
~--------------L------------------------- __~x
slow-stable-simple fast unstable
complex
Figure 7.12 Degree of operator control (Livingston, 1982): y, requirements for

operator-system engineering; x, process dynamics.
to most probable process hypotheses states, eliminating the whole burden of

the exact processing of inexact process data; see Chapter 9 for details.
REFERENCES
ABB (1989) GRADIENT - Graphic Dialogue Environment, Publ. No. D CRH 133
89E.
ASEA-ATOM (1982) Special Features of the Control Equipment for the ASEA-ATOM
BWR Appendix to IAEA Guide Book on I & C. ASEA-ATOM, Feb.
Dallimonti, R. (1976) Instrumentation Technology, May, 39.
Hornbach, B. (1982) IEEE Trans. Commun. COM-30, No.6, 1329.
Krueger, H. and Muller-Limmroth, W. (1981) Arbeiten mit dem Bildschirm-aber richtig,
Bayersche Staatsministerium fur Arbeit und Sozialordnung, Munchen.
Livingston, W. (1982) IEEE Trans. Power Appar. Syst., PAS-lOl, No.8, 2543.
Peterson, L. and Peterson, M. (1959) Journal of Experimental Psychology, 75,
193-8.
Woodson, W.E. and Conover, D.W. (1966) Human Engineering Guide for Equipment
Designers, University of California Press, Los Angeles.
8
Process management and control
8.1 INTRODUCTION
Once the problem of hardware for process control had been solved the next
move, in the 1970s, was to provide monitoring and simple control systems.
Inevitably, things have advanced since then and one definite trend is towards
the integrated automation of specific processes. Such automation is absolutely
necessary for fast and complex processes (e.g. electric power generation,
transmission and distribution systems or natural gas distribution systems) and
consists principally of two possible system behaviour steps:
1. Recognition of situation for automatic small-size system adaptations
and the small-size control action and automatic reporting on the action
to the operator.
2. Automatic on/off control action on recognition of specific process state,
such as alarm state or danger, and reporting to the operator on actual on/
off control actions performed.
The second step is performed on the basis of process control algorithms simpler
than the first action that requires fast, accurate and updated action, model
calculation and model verification algorithms because it usually represents the
fast and complex action on process behaviour. A relatively simpler case of the
first step can be executed for slow processes where the recognition of process
action is modelled and reported to the operator who then decides and/or un-
dertakes the appropriate action.
Four cases of process control actions considering the role of the operator
are given schematically in Figure 8.1.
8.2 PROCESS STATES
The working state of the process is usually characterized by its regular

production activity. The working state can be detected by the supervision of
the states of process variables. The working state can be disturbed by changes
257
C,
process
operator control process
• system
control
action
(a)
simple
protection
(b)
•
-----------
• control
action
(c)
II process
model
•
dialog t-~~~--c-o-n-tr-o~l--:;~------~------~~
, II
I
I action
I (d)
I
expert
system
,,
\
process
__ A
model .JI
_______
'-
Figure 8.1 Four cases of process control action: (a) control action by operator's
decision; (b) automatic, by control equipment; (c) control action by operator's
decision based on process modelling; (d) automatic, by control equipment with
feedback of information to the operator or automatic through expert system or
through the dialog between operator and expert system.
Process states 259
total number of
process states
controlling
solvable
process states process states
safe states dangerous allowable unallowable

states states states
working nonproductive
states states
Figure 8.2 The main process states.
in the process hardware, input or output, or by changes in the control hard-

ware or software. The change of process state can be influenced with and
usually limited by technology actions, or by the control actions. The tech-
nology actions are apparent to each process and their fulfilment can lead the
process either into safe or dangerous process states. Safety is measured and
estimated as being related to the operator's life, surrounding people and bio-
logical balance of the environment. Technological actions are not influenced
by the operator's actions and thus form a part of a passive protection means.
A new and more preferred process state can be reached by detection of
actual process variables and adequate actions in control equipment. Such control
actions can lead the process into allowable or unallowable process states
(see Figure 8.2).
For some processes the actual process state can be detected by following
the values and states of the particular process variables, where two different
groups of data can be distinguished according to the type of signal registered
(Table 8.1). The analog signal data for particular data processing functions
can be used as given in Figure 8.3.
The states of status signals and on/off commands can be combined to obtain
Table 8.1 Types of signal registered
Type of Content of Time/space function relation

registered signal registered signal of process variable
Status signal data Status information Not so urgent

Not always available
Not always up-to-date
Warning signal Almost always available data
Prompt information
Alarm signal Always available
Prompt information
Usually requires controller
action
Analog signal data Normal-to-alarm range Not so urgent data processing
Not always available
Not always up-to-date
Alarm-to-out of range Always available data
Prompt data processing
Out of range signal Always available
Prompt data processing
Usually requires controller
action
analog signal amount (%)

t out of
100% 1--------::;3iI!!!.~~!!E_------------_a_ range high
alarm to out
of range high
75%1----~~----------~~-----------------.~
normal-to-alarm
50% 1-4~--------~E:---------:3~+range
25%
10%~--+-~~~~=--=~==~~~~---------L.1
t alarm to out
of range low
tout of
range low
dead zone of data processing,
for example 2% of rated
signal amount
Figure 8.3 An example of the data processing function for analog signals.
Process states 261
VCI
VC3
V2
- - - - - - - - - - -'---+----, PS
FSI FS2 PS supposed process state actions to be taken
0 0 0 safe state none
0 0 I dangerous state VC2 = 'i'
0 I 0 safe state none
0 I 1 safe state none
0 0 safe state VCI = 0; VC2 = i
I 0 I dangerous state VCl=0;VC2=i
I I 0 safe state none
I I dangerous state VCl=Q
Figure 8.4 Safe and dangerous process states and corresponding actions to be
taken: FSI, inputfiow switch ('1' ~ FLOW); FS2, outputfiow switch ('1'~
FLOW); PS, pressure switch ('1' ~ HIGH PRESSURE); VI, input valve; V2 output
valve; VCI, input valve command ('i' = OPEN, '0' = CLOSE); VC2, output valve
command ('i' = OPEN, '0' = CLOSE).
a complete picture of safe or dangerous process states (Figure 8.4), and to

decide on the action to be executed to obtain a safe state.
Technology actions, supposed state of process and control, and the opera-
tor's actions for the case shown in Figure 8.4 are detailed in Table 8.2. As can
be seen, a large amount of preparatory work should be done before entering
safely into automatic control of a very simple basic process unit of the type
given in Figure 8.4.
Table 8.2 Control of a simple basic process unit
Actions Process state Status signals Control and operator actions Supposed process state
number FSl FS2 PS
None 0 0 0 None Tank is empty

Pressure on tank above 2 0 0 Open valve 2 in order to Tank is full above safety
safety level; decrease pressure level
Upon opening of valve If after given time interval
V2 the tank is being (L\ Tl) the pressure should not
emptied which should be be decreased;
checked through flow check pressure switch
switch FS2
Emptying of the tank is 3 0 0 No control action Tank is being emptied and
taking place If flow switch state equals zero the pressure low
then the tank is supposed to be
empty
Emptying of the tank is 4 0 None Tank is being emptied and
taking place If state 4 lasts more than given the pressure high
time interval L\ T2 then check
pressure switch and flow
switch
Tank is being filled 5 0 0 If tank is being filled more Tank is being filled above
than given time interval ~ T2 safety level
and pressure is still low then
close valve VI, open valve V2
and check pressure switch PS
Tank is being filled 6 0 Close valve VI and open Tank is being filled above
valve V2 in order to decrease safety level
pressure. If after time interval
~ T3 the pressure is still high
then check pressure switch PS
and flow switch FS2
Tank is being filled and 7 0 None Tank is being filled and
emptied at the same time emptied at the same time
below safety limit
Tank is being filled and 8 Close valve VI in order to Tank is being filled and
emptied at the same time decrease pressure. If after time emptied at the same time
interval ~ T4 the pressure is above safety limit
still high then check the
pressure switch PS and the
state of valve VI
264 Process management and control
The basic reason for this complexity is that given N status process signals
of a basic process unit that independently describe independent process variables
and process states, then a total of ST process states is
ST = 2N (8.1)
All possible ST states should be analysed and adequate decisions for each of
them prepared. These decisions, executed by the operator, range from very
simple automatic actions (e.g. automatic action on a button after receiving
a signal on the video terminal) to sophisticated operator-system dialogs that
involve process modelling and emulation.
Even more process states than given with equation (8.1) appear when analog
and counter state signals are taken into account.
Distinguishing k groups of signals, each group with an equal number of
states, the total amount of states ST is
k
ST = I1 Nfi (8.2)
i=l
where Ni is the number of signals in the ith group each with Ii states.
Thus, a recognition of process states should be preferred since a very large
amount of process states is expected.
8.3 OPERATOR/PROCESS INTERACTION - ACTIONS,

COMMANDS AND TIMING
The interaction of an operator with the process can be caused by three main
reasons, as viewed from time requirements (Table 8.3).
1. Fast interaction (e.g. start/stop or on/off) operations on final control ele-
ments, named control operation; the time for these operations ranges from
a few seconds to a few hours.
2. Slow interaction named planning operation (e.g. load and demand pre-
diction, process simulation for optimization purposes, economic produc-
tion, distribution and process security); the time r<:mge for these operations
extends from minutes to a few days.
3. Very slow interactions named following up (e.g. operations caused by daily,
weekly or monthly change in production, demand, load, power and ma-
terial exchange); these operations range from a day to a few months.
The main tasks of various computer control systems at different hierarchical
levels of control are given in Table 8.4. The control levels below substation
or remote station are usually equipped with synoptic panels with dedicated
operator-system communication and used both for synoptics of the basic process
unit and the process controller.
The operator-system dialog is performed primarily to satisfy the function
Operator/process interaction - actions, commands and timing 265
Table 8.3 Data and functions for different types of operation
Type of Content of operation Data and functions needed

operation for operation
Control operation on/off operations on final control process status signals

ranging from few elements process command signals
seconds to few
hours start/stop operation on BPU* process status analog and count
signals
process command signals
BPU local console
auto/manual operation on BPU process data acquisition and
presentation
limit values monitoring
auto/manual facilities
increase/decrease operation on analog and status data
regulators monitoring
limit values monitoring
set-point issuance
Planning operation scheduling process data acquisition
operation ranging load and demand prediction monitoring of limit values
from few economic planning process modelling
minutes to few operation simulation process security modelling
days local process optimization load and demand modelling
and stratification
process economy modelling
Following up daily, weekly and monthly production, load, demand and
ranging from day production logs event data
to few months load and demand reports process environment data
event reports
process environment
reports
* BPU - Basic Process Unit
of process state supervision. The division of global functions of an operator-

system dialog is given in Figure 8.5 for the remote control system of the grid
substation and for process control stations of the plant.
The whole operator-system dialog and operator interaction with the process
is supported and designed on the following basis:
system and application software, including database/knowledge base sys-
tem;
hardware data acquisition, processing, presentation and hardware com-
mand execution;
operator-system dialog elements as mutually agreed upon.
Table 8.4 The main tasks of different computer control systems
Level of Operation Planning Following up

hierarchy
National Production
grid Supervision of: Consumer prediction Reporting and accounting
consumers production schedules statistics
Inter- production balance planning following up of efficiency
connected reserves planning of reserves fault analysis
utilities process network coordination of
Operation and overhauls
control of plants
Utility Production
Supervision of: Consumer prediction Reporting unit accounting
consumers production schedules statistics
Regional production balance planning following up of efficiency
grid reserves planning of reserves fault analysis
process network coordination of
Operation and overhauls
control of utility
Group Production
of plants Supervision of: Short-term planning Reports on production
consumers according to accounting data
District production directives statistics
grid process state
plant components
Operation and
control of
substation
Plant Control of Work planning Sequential event
variables, recording
sequential start/
stop functions,
automatic system
restoration,
protective
functions,
supervision of
process variables,
auto/manual for
local equipment
Common functions
Selection of the computer
Keybord and CRT selection
Basic mimic diagram
Dialog and system selection
Acoustic alarm quitting
Copying
Interrupt
Process Operation
planning
Selection of mimic diagram
Deletion of time controlled from a selected system
reports Alarm quitting
Memorandum issuance
.l Process shutdown
RCS . PCS I
RCS" .. PCS
CRT Control Manual/computer switchover Control
Functions blockage Control of remote stations On-off control
Process data limits read-in Printer and print-out selection Set-points control
Control of stations status System time Status control
Control of CRT Data read-in Data read-in
Figure 8.S The division of functions according to type of operator-system dialog:

RCS, remote control station; PCS, process control station.
The basic components of a programming system supporting operator-system

interaction are system functions executed in real-time and extended real-time,
and user's application functions, which are all usually supported by the fol-
lowing program functions in real-time:
1. Data acquisition, data processing and visualization of process data.
2. Data acquisition, data processing and visualization of control system data.
3. Process control, including modelling and real-time expert system.
4. Control of the control system.
Software support for extended real-time functions enable process planning
(i.e. the change of system parameters such as the change of process, process
control and visualization databases), reporting (i.e. the acquisition of process
data and their systematic presentation in given time intervals or at operator's
demand) and extended real-time processing (i.e. all data processing functions
longer than the given system refreshing interval).
The application programs specially developed by the system user enable
the fulfilment of user-specific functions appropriate, for example, for their
process hardware and technology.
Application programs use and share the same computer resources and are
therefore limited by the integrity and speed requirements imposed on real-
time and extended real-time software models.
The appearance of a process event indicates the change in a database and
is processed according to the content attributed to its appearance.
Events demand different simple data processing, for example, a list of events,
alarm and failure recording, as well as chronological event recording.
A composite data processing of events results in display refreshment, which
supplies information to the operator on events such as display of real-time
mimic-elements, grouping of real-time mimic elements, and selection and
authorization of work of operators.
A list of events, alarms, failures and chronological events is issued on
printers for operators' information and control.
Operator--<:omputer communication can impose special requirements in real-
time system operations. Here, the overall system reaction time is essential for
system functioning (Figure 8.6). Table 8.5 outlines individual reaction times
for trained operators, as depicted in Figure 8.6.
To assist the operator in undertaking the proper action at the correct time,
some rules, elements and propositions for operator-system dialog have been
standardized (DIN, 1979), and given in Tables 8.6, 8.7 and 8.8 for optical
signalization, and in Table 8.9 for acoustical signalization in process control
systems.
The interaction of an operator with the process is done by changing the
process control functions; but there are many factors that influence these
functions (Wilhelm, 1979), which are outlined in Table 8.10. The usual main
process control commands for operator-system communication are:
expert
system
T4
process
Figure 8.6 Overall system reaction time.
1. Start command that requires the transition of the process from OFF to
ON state.
2. Stop command that requires the transition of the process from ON to OFF
state.
3. Activate command that requires the transition of the process from inac-
tive to active state.
4. Deactivate command that requires the transition of the process from
active to inactive state.
5. Unlock command that requires the transition of the process from un-
available to available state.
Table 8.5 Data for Figure 8.6 for trained operators
Reaction Approximate Brief description

time time range
Tl 0.4 s - 10 s Execution of demanded control or dialog

actions of the operator on control system
(on keyboard, functional keyboard or panel
command pushbuttons)
T2 0.1 ms - 30 s Execution of accepted command towards a
final control element; the main time delay
can be caused by communication facilities
T3 0.1 ms - 100 s Execution of command at the final control
element; the main time delay can be
deliberately set high enough to enable
system stability
T4 up to 1000 s Process reaction; ranges from few to tenths
of minutes depending on process type
TS 0.1 ms - 10 s Process state detection time; depends on
sensor reaction time
T6 0.1 ms - 10 s Control system event processing; depends
largely on I/O filter constants and data
processing algorithms
T7 10 ms - 10 s Visualization of process event to operator
TS (individually) Operator reaction time to process event
o.s s - 30 s and decision to take immediate action in
real-time control environment
TS' (individually) Operator reaction time to process event
O.S s - 60 s and decision to take immediate action in
real-time environment supported by an
expert system software
6. Lock command that requires the transition of the process from available
state to unavailable state.
7. Notice command that requires the process control system to recognize a
state transition and to take any necessary action.
The basic process states transitions and process commands are given in Figure
8.7.
Particular process states (e.g. an unavailable state) can be reached by issu-
ing commands and by changes in process equipment and software behaviour
due to errors and failures that were not previously predicated. In such cases,
after the system has been repaired, particular process states are put into the
Table 8.6 Some standards for operator-system dialog (optical signalization)
Main Group Corresponding Recommended max. no. of Comment Examples of elements

characteristics characteristics alphabets elements
alphanumeric 64 (95) characters language 1, A, a,

character dependent
abstract < 20 at same time language
figures independent DOO+
Figure ~
pictograms < 20 at same time language
> 20 for known independent $ bJQ © =+=
pictogram
lines 4 different types _ JlJL IV IV
Shape ( the type of 2 different types 1, A, a,

writing 1, A, a,
linear 2 different types A,B,C,D,A,B,C,D
distortions
Form
1,
< figure size 3 different types logarithmic 1, 1,
rates
recommended
line width 3 different types logarithmic
rates
recommended
3 different type!> III
{ type oft"",
Texture type of two- 3 different types
dimensional
hachure <.,',".
0 ~ ii
Main Group Corresponding Recommended max. Comment Examples of elements

characteristics characteristics alphabets no. of elements
type of colour specific colours 6 + black and white only for characters purple, blue, blue green,
from 20' of visual green, red, yellow
angle; for smaller
angles only red, green,
blue and purple
saturation saturation not for coding
light light 2 different types alternative or redundant green-light green
Colour coding
contrast colour contrast 2 different types colour on coloured positive (colour on neutral
background not background)
recommended negative (neutral on colour
background)
neutral 2 different types positive (white or grey on
contrast black)
negative (black on white
or grey)
system by operators or service staff issuing auxiliary operations and manual

commands. The procedure whereby an operator intervenes after such a failure
to put the system into a particular process state should be well-defined and
carefully studied; it demands the manual preparation of all basic process units
so that the change over can be successfully made to computer controls operation.
This presents special difficulties for some plants, especially where conditions
are hazardous for humans or the units are not easily serviced.
The corresponding commands and process signalizations can be issued on
different process control equipment. The specific features of different process
control equipment for command and process signalizations are given in Table
8.11. The indicated time intervals are heuristic and correspond to expected
operator's and computer's reaction times.
The control of process variables is performed either as an on/off control, or
as a regulation or governing function. On/off control is a procedure that be-
gins with the selection of the mimic element. The computer checks whether
the operator might control the selected process element controlling the fol-
lowing suppositions:
1. Control competence must be attached to the operator.
2. Process element data must be refreshed.
3. On/off control function must be allowed.
4. Only this control place is in the control procedure for the selected process
element.
5. Logic control function must be fulfilled.
The operator is informed on the result of a computer check procedure and, if
the computer permits on/off control, the appropriate keyboard action may be
executed. The keyboard is blocked until the time supposed to execute this
process command elapses, or until acknowledgement of the command execu-
tion. The change of process element state is usually registered and, if it does
not occur in the prescribed time interval, an alarm is issued to the operator.
A regulation or governing function is a procedure that must fulfil the same
suppositions as for on/off control. The computer issues the allowance to the
operator for the execution of the regulation, when two procedures become
possible:
incremental regulation, which is executed by functional keys that enable
positive or negative increments of the process final control devices;
set-point regulation, which is executed by the operator typing-in the set-
point value.
The new process element value is checked against the given limit values.
Both instantaneous and present values are displayed.
Extended real-time functions usually enable some of the following features:
1. Control of measurands' plausibility.
2. Display of actual process topology.

position absolute 9 different positions upper right

position bottom left
relative 2 different positions indices
position
orientation orientation 8 different
of figures orientations
without
Place reference
*~
orientation 24 different more orientations (\
of figures orientations also possible
with reference
time variation time variation 3 different types applicable to slow, fast, pulsed,
of shape of figure process control stepwise, continuous
purposes
time variation frequency of 3 different application for danger 0.7 Hz, 2.2 Hz
of colour by colour frequencies and attention
Time pulsation pulsation attraction
duty cycle 3 different types 1 : 1, 1 : 3, 1 : 10
speed of 2 different types applicable to continuous, stepwise
change process control
purposes
time change translation specific for process continuous, stepwise
of place control purposes
rotation specific for process continuous, stepwise
control purposes
Table 8.9 Some standards for operator-system dialog (acoustical signalization)

tone one tone sinus tone

many tones ground tone with
upper tones
Type of accord 2 harmonic accord
sound disharmonic accord
~ disturbances noise
complex 3 buzzing, ringing
sounds
speech call 10 all necessary warning
and calling
words, text very articulate
Frequency chosen chosen 3 not to apply octave
frequency frequency frequencies
Level acoustical acoustical 3 minimum 10 dB
sound level sound level above noise
Time duration duration 2 long, short
time variation series of tones 2 melody
of frequency
speed of 2 continuous, stepwise
and tones
change
frequency 3 slow, fast
change
time speed of 2 continuous, stepwise
variation change
of level
frequency 3 slow, fast
of change
pulsation 3 pulses, regular pulsation
Table 8.10 Factors that influence process control functions
Factor Factor's influence Comment
Operator operator inputs in auto/manual systems

Final control element status of final control element in dynamic or erroneous
situations
Sensor status of sensor in dynamic or erroneous
situations
Control functions status of related control functions depending on mutual
connections of basic
process units
Stages of operation status of production in batch or
sequential processes
Process state change of process control some regions of operation
function may require changes in
control functions
3. State estimation of process variables.

4. Network flows (material, energy, information).
5. Analysis of work safety.
6. Calculation of optimum process control.
7. Calculation of state indicators for specific process parts.
8. Follow-up and reporting of process state indicators.
9. Analysis of statistic features of historic data.
10. Short-term system state prognosis.
The communication of extended real-time programs with the operator is per-
formed by an interactive dialog using the whole CRT display, or by real-time
mimic displays with specially designated mimic parts (or mimic elements)
calculated according to extended real-time programs. The interaction of the
operator with the process, according to knowledge obtained from these calcu-
lations, is done in the same way as for real-time interventions. An example of
the division of the CRT display for operator-system communication is given
in Figure 8.8.
8.4 PROCESS PROTECTION AND AUTOMATIC ACTIONS
The working body (system) is the process body observed separately from all
the other mutually interrelated bodies (systems) and where they are all treated
as the system environment. The process is a macroscopic reflex of outer,
variable and microscopic changes in the working body. The process states can
Process protection and automatic actions 279
active command
process
process
active inactive
state
state
process process
on off
state state
(a)
lock command
process process
available unavailable
state state
unlock command
Figure 8.7 The basic process states transitions and process commands (a), (b) -
optional transitions.
be distinguished as stable and unstable states, according to their character. By

isolating the working body from the system environment, the stable states are
obtained. The interaction of the environment inevitably involves the instability
of the process. Thus, there are two opposite tendencies during the process
life; the outer tendency from the environment works towards the process in-
stability, and the inner tendency inside the system works towards the process
stability. Since almost all systems are bound by many subsystems with their
environment, most systems are inevitably unstable.
To prevent the effects of process state instability, the protection of particu-
lar processes is organized according to best known experiences and calcula-
tions for each separate process. It is useful to distinguish between active and
passive protection, and between static and dynamic protection. Figure 8.9
shows the interrelations of three systems. Consider, for example, system 1 in
Table 8.11 Specific features of process control equipment for command and
process signalizations
Input or output Communication Operator-system

device action mode input/output expected
response times
Operator ~ QWERTY Text input - 0.2 s/character

keyboard mnemonics
Operator ~ QWERTY Coded input 0.2 s/character
keyboard
Operator ~ Functional Functional keys 0.2 s/key
keyboard
Mimic display ~ operator Question Is/question
Mimic display ~ operator Warning on false 0.5 s/warning
response
Computer action Computer action 0.5 s - 10 s
Mimic display ~ operator Information on Is-lOs
action result
Mimic display ~ operator Information on 0.1 s - 5 s
spontaneous
events
Figure 8.9 as a boiling water system, system 2 as a saturated steam system

and system 3 as part of the water circulating tubes and armatures; then the
following applies:
passive protection of system 1 against the action of system 3 on system 1.
For example, better isolation and screening of tubes to obtain improved
mechanical performances when conducting boiling water;
active protection of system 1 against the action of system 2. For example,
the inclusion of the facility for forced cooling of the saturated steam.
The active protection facility can be of either a static or a dynamic nature
depending on the type of process or process behaviour it covers. A static
protection system is used for the protection of process steady state work and
operates on the basis of stationary process protection variables. A dynamic
protection system is used for the protection of process dynamics and operates
on the basis of dynamic process protection variables.
Usually some simple processes only start up with manual intervention by
the operator. More complex processes have to be automatically started, re-
started, put into work and stopped. Both manual and automatic operations can
o 2930 42434647 63
o A I B Ic I D
1 E
47 GI H I I I K
I L 1M
01 2 13 42 43 6061 63
Figure 8.8 An example of the division of the CRT display for operator-system
communication: A, mimic identification number; B, type of dialog; C, reserved for
operator's reminder system; D, time of day; E, last unaccepted alarm; F,
individual mimic display; G, type of computer and display unit; H, key-in of
alphanumeric characters; I, procedure error; K, text of the selected process value;
L, computer message; M, function designation.
be, or sometimes have to be, supported by dynamic protection systems. The

reason for this lies in the complexity and number of decisions that can other-
wise be expected from the operator, thus saving time, energy and materials, as
well as avoiding dangerous situations for people and process equipment alike.
Automatic actions for process protection in the dynamic range are' supported
by those actions that serve for process protection in the steady state. These
two types of protection are usually separated in their actions because, when
used, the variables and process conditions differ considerably. The typical
protection system of either a static or dynamic type consists principally of the
same type of hardware and software as given in Figure 8.10 for a common
protection channel. The process protection variable is detected by a sensor
and then transmitted as a protective action signal to a limit signal sensor. The
cause of an appearance of a process protection variable in the range that is of
interest for the protection action can be:
stochastic change in the process variable;
systematic change in the process variable;
Figure 8.9 Active and passive protection measures: a j • action of system 3 on

system 1; a2 • action of system 2 on system 1; Rj, passive protection measure of
system 1 against action a j ; R2 • active protection measure of system 1 against
action a z .
change in the process variable caused by the change in another part of the
process;
change in the process variable during repair, damage or unit shutdown.
The protective limit signal generated at the output of the signal limit sensor is
fed to the logic operation circuit where it is combined with other protective
limit signals to generate (1) appropriate signalization to the operator and (2)
signal of protective action condition. The signal of the protective action condition
is evaluated by redundant logic circuits (e.g. 'two out of three' circuits, or
similar) and an output protective signal is generated on this basis to the control
system. The control system generates the appropriate output protection signal
with the required power and time requirements to the final control element.
The protection systems are designed for various processes. The criteria for
their design include the following minimum requirements.
1. The existence of process conditions that require protection.
2. The existence of process variable to be monitored to provide protective
action.
3. Exact limit values and levels for each process variable that require pro-
tective action.
4. The margin between each protective limit and the level considered to
designate unsafe operation.
5. Protection performance requirements such as protection system response
time, system accuracy and ranges of process variable conditions (normal,
abnormal, accidental) both in magnitude or/and rate (Figure 8.11).
--- - - - - - --- --- -- -- ------- - - -A

process process protection variable T
I
I
I
I
exciting
level
I
I
I
I
I
I
I
I
limit signal I
-------------------------~
I
signalization
•
logic
level
II
II I
I
_ ___ ~~tp~:.~r~~~~~e_s!~~ ____ __ ____;_
~~
control
level
I
final control element
---___ ___---__-_-__-_-..::~:, -_-__-_.,.-__- _-__ ____ __________ t __
I
Figure 8.10 A typical protection system.
Protection systems are usually implemented by separate hardware that ensures

process protection as specified in the process protection requirements and
they are an integral part of a process technology. Since there are different
approaches to the role of protection systems, an example of static protection
system equipment for small generator units is shown in Figure 8.12 where the
following can be inferred:
1. Even the simplest actions of the generator (e.g. overvoltage relay action
on generator breakers) is perfonned through the protective unit.
2. Complicated actions of load-shedding can be perfonned by appropriate
actions of overload and reverse-power relay (OLR), as shown in Figure
8.13, where the load-shedding by the first stage of relay OLR is perfonned
at the time instance t 2 .
process variable v al ue amount, rate
accidental
range ac cidentallevel
I........
unsafe
operation
range safe operation level
- -- - --- ----protective----- - --- - - -- ~
r
limit range 1?!otective limit level
abnormal operation range
normal operation range

start-up
operation -<
range
abnormal operation range
Figure 8.11 The ranges of process variable conditions.

N---...---
ENL
,
PU
•
Figure 8.12 An example of a static protection system for a small generator
unit: ~. three phase measurement transformer; ::t>-.OR circuit; CD. voltage
measurement sensor; -cD. current measurement sensor.
p
Pr -+----~------------_T------------~-----------------
Figure 8.13 Load-shedding action of an overload power relay (OLR): P, output

power; Pr, rated output power; P max' maximum output power.
REFERENCES
DIN Coding of Information (1979) CRT Display Working Place, DIN 66234.
Wilhelm, R. (1979) IEEE Trans. Autom. Control, AC-24, 1, 27.
9
The role of the operator in process
control systems
9.1 INTRODUCTION
There are two trends in the role of an operator in process control systems:
to use automated process control systems as fully as possible and to de-
crease the number of operators and the skill required of them;
to collect a broad number of fairly skilled, trained and schooled operators
from a 'field' operator to a general system operator and process planner,
and to support their work by the appropriate process control systems.
These trends, to possess a fully automated process control system and fairly
skilled process operators, are interrelated since only on the basis of efficient
automation can there be a full use of the operator's skill. This interrelation
materializes when microcomputers are used in process control in such a sense
that there appears a fully automated process which requires less labour.
Production processes are becoming increasingly specialized but there is a trend
to group them together. The variability of production conditions and technologies
with their fast development and competition for energy, raw materials and
human resources has become so powerful that it limits further expansion of
already-installed processes and forces new processes to appear and take over.
The knowledge of a particular type of production seems to depend on the
experience and skill of workers in the process, whose gift, speed of reaction
and creativity help, by using the same amount of raw material, energy or
efforts in the production, to increase the quality of products. Such workers
can be supported by computerized process control systems, and their effec-
tiveness is usually increased by collecting a set of relevant process data and
relations. An operator's creativity can be supported by adequate process control
devices that increase insight into process states and behaviour, and enable
fulfilment of process commands, as well as increased productivity made poss-
ible by fast calculations and personal experiment. The most promising tool for
operator support today is expert system software since it uses artificial
inferencing similar to human reasoning.
The operator's role in a process control system may be presented as four
groups (Table 9.1), and are as follows.
287
Table 9.1 The operator's role in the process control system - functional subsystems
Types of Measurement Regulation Governing Supervisory Control Special production

operator's work system system system system system machines
Process + + +
manipulation
Process + + +
supervision
Process + + + +
execution
Process + + + + +
quality control
+ applicable
not applicable
Artificial inferencing tools 289
1. Process manipulative role (e.g. workers in the textile industry).

2. Process supervisory role (e.g. operator in natural-gas network).
3. Process executing role (e.g. operator controlling a numerically controlled
machine tool).
4. Product quality control role (e.g. worker controlling and inspecting the
final products in the food industry).
9.2 FORMALIZATION OF THE OPERATOR'S ROLE
The operator's role in process control is often very formalized for the follow-
ing reasons:
preservation of the unified manipulation with process control and produc-
tiOl~system;
simplification of the design of process control and production system;
acceptance of specific human reactions and physical and mental features
and limitations;
simplification of the training of operators and their prequalification, if
necessary.
Table 9.2 details the formalization of the operator's role. Three main char-
acteristics seem to exist in operators observed as users of process control
systems:
1. Operators prefer analog to digital data and process presentations.
2. Operators prefer traditional process control procedures, geometry and
colours to new solutions on either one of them.
3. Operators are selective in their choice of tools and means of control,
choosing only the most efficient.
The first characteristic, regarding the design of control panels and displays,
enables an overview of a process state by bar graphs and process mimics.
Traditional solutions require a specific design of process controller panels,
regulator front panels and emergency and alarm control panels.
The formalization of the operator's work can be done effectively by using
formal operator-system dialog such as the CCITT man-machine language.
In such a way, all four reasons for formalization of the operator's work in
process control systems can be solved.
9.3 ARTIFICIAL INFERENCING TOOLS - EXPERT SYSTEMS

FOR PROCESS SUPERVISION AND DIAGNOSTICS
Operators' decisions are based on inferencing and execution/mechanisms.

Inferencing is an intriguing task. Sometimes it is connected with exact rela-
tions of the type IF-THEN-ELSE; logic judgements examples are in pro-
Table 9.2 Formalization of the operator's role
Aspects System requirements: Categories of personnel equipment data

maintenance and
operations Human Personnel Training Training Technical
engineering requirements concepts equipment manuals
and plans
Unified System accuracy Maintainability Personnel policy Job Training Operator

process Information handling design procedures equipment manuals
control and Narrative and pictorial Job skills Training materials
production descriptions
manipulation Operation conditions
Simplification System reliability Equipment Training Training Design
of the design System vulnerability design methods equipment design procedures
of the process Descriptions Layouts
control and Theory of operation Design
production trade-offs
system
Human System reaction time Man-machine Illumination Training
physical and Environment data function Manual control data
mental Manning data allocations Visual displays Special
features Life support Auditory signals training
functions Sensory signals requirements
Safety Conformity
(biomonitoring,
emergency)
Operator's Maintenance support Crew performance Training Maintenance
training Maintenance goals and curricula, materials
simplification procedures techniques and Parts lists
Maintainability scheduling Duty lists
grammable logic controllers implemented for process control. More complex

and less exact inferencing is an attribute of human-like inferencing mechanisms.
Logic judgements are used for stochastically unconnected events. Inexact
inferencing is used for mutually connected appearances - facts. Exact inferencing
is both well defined and easily followed. By contrast, inexact inferencing is
something all of us do daily - but without being able to explain just how we
do it.
Most technical inexact inferencing mechanisms are based on Bayes' statis-
tics and entropy concept (Brillouin, 1956). Bayes' theorem states that even
for the most improbable event there exists on a priori probability of appear-
ance and the fact of a detected event is connected with the post priori probability
in the following way
P(H) + P(E) => P(H:E) (9.1)
where P(H) is the a priori probability, P(E) is the probability of appearance
of the event in the concrete case and P(H:E) is the [post priori] probability of
the appearance of the event. We can use the relations
P(H:E) = P(E:H) . P(H)/P(E)
(9.2)
P(E) = P(E:H) . P(H) + P(E:not H) . P(not H)
where P(E:H), written as Py for short, means the probability of event appear-
ance (fact) in the case of the hypothesis being true, and P(E:not H), written as
Pn' means the probability of event appearance in the case of the hypothesis
not being true. By using the relations (9.2) we can obtain the formats used in
a knowledge base
P(H:E) = Py . p/[p y . p + Pn . (1 - p)]
(9.3)
P(H:not E) = (1 - Py) . p/(1 - Py) . P + (1 - Pn) . (1 - p)
where p stands for p(H).
Having calculated a new P(H:E) from newly acquired facts one can forget
the original P(H) and instead use this new P(H:E) as a new P(H). The same
applies also for P(not H). So the whole process can be repeated time and time
again, with a different probability derived from the last post priori probability
being used each time.
When hypotheses form a complete mutually exclusive family, meaning that
n
(n: universal set), and

n
i=!
and for E being any event holds there (see Figure 9.1), then
292 The role of the operator in process control systems
Figure 9.1 Hypotheses and events.
(9.4)
The supposition that events E!l HI ( i = i, ... , n) are mutually exclusive means
that
n
I. P(E n Hi) = P(E) (9.5)

i=1
Applying the relation P(E !l HI) = P(E) P(HI:E) = P(H I) . P(E:H I) we
obtain
n
(9.6)
i=1
The probability of event E can be calculated if we know the probability of HI
(usually named hypotheses) and conditional probabilities of events E in relation
to hypotheses HI (i = 1, ... , n). By using the relations (9.2) and (9.6) we can
obtain
(9.7)
which is the Bayes formula applicable to a range of practical tasks where the
probability of the conditional hypotheses Bi is sought, while the event E has
happened which usually appears with one of the hypotheses Bj (j = 1, ... , n).
Example
The case of a masoot warmer of a thermoelectric power plant is given in
Figure 9.2 where four different events are taken into account and put into
probabilistic relation with four hypotheses. This is a case for a binary expert
diagnostic system for which the corresponding data in the knowledge base are
given in Table 9.3 as well.
masoot
8 - filter
wanner ~
@- pump
water
180°C
masoot 7 bar
Figure 9.2 The functional scheme of a masoot warmer in a power plant.
Table 9.3 Knowledge database of the binary expert system
Hypotheses Events - facts

p(H)
A4asoottemperature A4asoottemperature Flow Pressure
high low OK OK
Failure of the Yes 0.2 0.8 0.5 0.4

masoot warmer No 0.9 0.1 0.9 0.5
pHI = 0.05
Failure of hot Yes 0.05 1
water supply No 0.999 0.05
pH2 = 0.01
Masoot filter Yes 0.4 0.7
failure No 0.2 1
pH3 = 0.2
Masoot pump Yes 0 0
failure
pH4 = om
No
rg~ll 0.99 0.95
p(H4:E) P(H4: E)
Table 9.4 The calculation of rule value (sideways chaining) for the case of the boiler
feed preheater from Figure 9.2
RV pACTl = 2.335 01 Masoot temperature high

RV pACf2 = 0.452 79 Masoot temperature low
RVpACT3 = 1.52778 Flow OK
RVpACT4 = 0.187 03 Pressure OK
Question: masoot temperature high?
If answer: no, calculation of new probabilities
PHI NEW = p(H,:E) = 0.29629
PH2NEW = P(H2:E) = 0.905 62
PH3NEW = P(H3: E ) = 0.15784
PH4NEW = P(H4:E) = 0.002 04
And new values of the rule values follow:
RV pACf , = 0 Masoot temperature high
RV pACTI = 1.67363 Masoot temperature low
RV PACT3 = 1.113 79 Flow OK
RVpACT4 = 0.122 83 Pressure OK
Thus one has to ask for the fact of low temperature of masoot, etc.
The hypotheses have to make a complete family of events in order to apply

the formula (9.7), i.e. an event outside the hypotheses cannot occur under any
known conditions.
The calculation of each hypothesis under certain facts gives a change in a
priori probabilities according to formula (9.2). There appear to be two prob-
lems for the expert systems in process control:
the number of rules connecting facts and hypotheses;
evaluation criteria.
A probabilistic conclusion needs the elaboration of evaluation criteria.
A simple evaluation criterion is given by a rule value (Naylor, 1987):
n
RV = I.!P(Hi:E) -
P(Hi:notE)1 (9.8)
i=i
Calculating rule values according to formula (9.8) from the given example in
Table 9.3 data can be obtained as given in Table 9.4. Thus the indication of
the most probable hypothesis is given. By scanning the facts for activity or by
asking about their state it can be found that FACTI = '0' so that new a priori
Bayes probabilities can be calculated, as given in Table 9.4, new rule values
obtained, etc.
P(H:E)
P(H:R)act
P(H)
P(H:E) Relative
O-------------~r-------------~r_--------~O_~
-5 does not actual response
+5
respond process
response
Figure 9.3 Uncertainty in process response.
The activation of each fact drastically changes probabilities and evaluation

results. Therefore when two or more facts are activated the rule with the
strongest connection to these facts will give practically the most dominant
evaluation.
The sequence of search should be directed towards the item of evidence
(fact) which can induce the greatest probability shift in all hypotheses under
consideration. Questions stored in the knowledge base should be answered
either by the operator or automatically by the system, for example by using
different models for specific process situations (Naylor, 1989), and involving
the finer measurement tunings and control procedures in order to find the
most probable relevant hypothesis.
The response variable R can be scaled, for instance between -5 and +5, and
all hypotheses which are referenced to that item of evidence in the knowledge
base have to be corrected to find P(H:R) as
P(H:R) = P(H:E) . P(E:R) + (H:not E) . P(not E:R) (9.9)
(see Figure 9.3).
P(H: all remaining facts supportive) = P max

Ml
change
in P(H) ' \ .
f
final P(H)
current
P(H)
M2
P(H: all remaining facts against) = P min
o
Figure 9.4 The five probabilities used for a given hypothesis.
The rule values for all items of evidence should be recalculated to allow for
the change in probabilities that have taken place given the last process response.
Then the minimum and maximum values which each hypothesis may yet
attain should be calculated (see Figure 9.4). Essentially there are five quantities
for each hypothesis; each of these quantities is a probability, namely:
P(H) is the current estimated probability of that hypothesis being true;
P max is the current maximum probability that this particular hypothesis
could attain if all the remaining facts went in its favour.
P min is the current minimum probability which a particular hypothesis
could attain if all the remaining facts worked against it.
Ml is the upper threshold criterion for accepting a particular
hypothesis calculated as a proportion of P max before any fact has
been acquired at all.
M2 is the lower threshold criterion for rejecting a particular
hypothesis calculated as a proportion either of P max or P min before
any fact has been acquired at all.
knowledge base start knowledge base
N---
prior hypotheses
probabilities for each Hi
extract p(H.)
N-- HI
p(H) start
I
evidence-fact H"
P max ' Pmin
,...
-
1 Fj k
--- rt rule
p, V-- calculate RV values
and
~Vi.
pn
~cture
qt··· qt
check-ups
- find RVrnax find if any hypothesis

- check the most influencing has a maximum
fact (MIF) in process possible value which
or with operator exceeds this maxima
- establish the value of MIF of the minima
(formula 9.9)
- recalculate all hypotheses
using MIF value
- calculate min and max
values which each yes
hypothesis may yet attain found?
no
most likely
outcome found
W
report
I
Figure 9.5 Bayes decision procedure and knowledge base.
The most likely outcome has been found if there is some hypothesis for which
P min is greater than Pmax for any hypothesis. The likely conclusions are those
hypotheses for which P min is greater than MI.
Uncertain conclusions are those hypotheses for which P min is less than Ml
and Pmax is greater than M2. These items are potential for resolving the
uncertainty.
No inference is possible (or the hypotheses are false) for those hypotheses
where Pmax is lower than M2.
The flow chart of the proposed procedure is given in Figure 9.5.
9.4 CONTROLLING THE OPERATOR'S WORK USING

A PROCESS CONTROL SYSTEM
The ability of a control system to collect and process a large quantity of data
often obliges process control engineers, managers and even system engineers
and designers to implement some sort of programmed features that can be
used for the control of an operator's work. The usual requirements that can be
expected from the system to control the operator's work, without affecting
freedom, dignity or privacy, seem to be as follows.
1. Registration of all process events caused by direct operator's action on

command buttons and keyboards.
2. Registration and print out of time-of-day data of all process events caused
by either a process or by the operator's action.
3. Control of operator's identity when executing some control function and
issuance of permissions for a particular process control action.
4. Control of database approach by appropriate operator identifiers.
5. Control of system productivity as per shift and datum, without citing the
operator's name in the shift.
6. Control of raw material and energy consumption per shift without citing
the operator's name in the shift.
7. Print-out of system failures and cumulative system failures with datum
and brief description without citing the name or responsibilities of the
operator or maintenance staff.
8. Analysis of an operator's reaction time to specific process events without
connecting this with the operator's name.
9. Built-in features and programs to estimate the system tuning to actual
process situation, according to some ideal system behaviour, without
mentioning the operator's name.
To increase confidence in the control system, the operator has to be acquainted

with all the system features controlling the work and operation.
The most crucial operator's actions that have to be checked and controlled
are set out below.
1. Control of typed input of process data parameters, i.e. by checking all

five limits of analog data in the order of appearance to the computer.
2. Check-in of command execution by checking
authority of command centre since other centres involved can exist;

authority of control place since there can be other control places in-
volved in a command centre;
completeness of command code;
The limits of the operator's function 299
- logic states of command execution, local or remote state; feedback

information from final control elements;
actualization of the state of final control elements.
3. Logging of an operator's action on each terminal.
9.5 THE OPERATOR'S REACTION TO PROCESS CONTROL

SYSTEMS
Many aspects of an operator's reaction to a process control system can be

observed. The most important aspect, as viewed from a production standpoint,
is the period during system installation and work. This is:
honeymoon period, that is, the period of enthusiasm when using the new
control equipment. This period is exhibited at the beginning of the knowl-
edge of future process control and upon the description of its possibilities;
- refusal period, which follows immediately after the honeymoon period
and characterizes the period of criticism that is exhibited in comparison to
new and existing process control technology, and possibilities from the
standpoint of preference of known process control technology such as
relay controllers;
comparison or cooling period, which is exhibited by a rational approach to
computer technology listing the main aspects and differences between new
and known process control technology;
- training period, which is exhibited as a period of efforts used from the
acquiring of knowledge about the new system that will be installed and by
working efforts in the on-the-job training period with the new control
system;
- working period, which is exhibited by further knowledge about a system's
possibilities, by expanding or reducing its design possibilities and by a
critical knowledge of the efficiency of the present control system.
All these aspects can be easily surmounted when necessary preparatory work
is performed, which consists of using computer models, trainers and teachers
for the first four periods, and in using an open-ended approach in the fifth
period. Thus, a complex control system needs an extraordinary amount of
memory space and adequate software for further experiments (see Figure 9.6).
9.6 THE LIMITS OF THE OPERATOR'S FUNCTION
An operator's life is potentially full of surprises because virtually all large

processes have been designed and operated without a prior systems engineer-
engineering
1 Gbyte
tool (CAD)
expert
100 Mbyte systems
supervisory
10 Mbyle control and
data acqui ilion
systems
1 MbYle
process
controllers
andPLCs
100 kbyle
process
10 kbyte devices
1 kbyte
100 byte I I I I I I I I • year

70 73 76 79 82 85 88 91
Figure 9.6 Amount of software for operator-system interface and database support
in process control.
ing operational analysis (Livingstone, 1982). Process operational problems

already demonstrated by experience have been cured by an endless series of
solutions such as duplicated control loops, safety relief valves for pressure-
controlled circuits, bypass valves and backup manual sequencing. Some process
failures which have been foreseen or experienced can be supported by special
systems to respond and reduce the consequences of some unwanted process
functions. But there is always an irreducible residue of unforeseen process
functions. Figure 9.7 shows a typical frequency/severity characteristic curve
for off-normal operations. The reluctance to identify off-normals, which put
all designers into a defensive position, is precisely the reason why blind spots
occur and why the problem of process uncertainty exists. The application of
OSI concepts and the provision of diagnostic expert systems can prevent some
of these design failures.
References 301
G\
C \
II) ~ \
~ lS ,
E'" ,
II)
"
envelope of
operator-machine
ystem re pon ibility
severity
of
damage
t
- - ----
~
benefits
•
probability of occurrence
~:
even ts evaluated
as unlikely
to occur
Figure 9.7 The plant operations system.
REFERENCES
Naylor, C. (1989) ABB Publication No. 0 CRH 1333 89 E, GRADIENT Graphical
Dialogue Environment.
Livingstone, W.L. (1982) IEEE Trans. Power Apparatus and Systems, PAS-101, 8,2543.
Part Four
System Design
10
The feasibility study
10.1 INTRODUCTION
A feasibility study is carried out on some small, many medium and almost all
large process control systems. The process owner, user or designer may use
this study to formulate, state or clear up the main points of a future computerized
process. The results may be described as follows:
the ways and means of process computerization to obtain certain tangible
benefits from computerization;
the organization and coordination of the process that has to be changed to
allow computerization with given benefits;
application of other technical means to raise productivity and lor the quality
of products;
changes in the organization of the firm where the process is performed to
increase its productivity and/or the quality of products;
changes in the level of operator's and manager's knowledge of process
computerization in order to secure foreseen benefits;
data and data sources used for the calculation of the proposed computer-
ization of the same or similar processes;
the most effective computer configuration and plans for its implementa-
tion.
A feasibility study begins by issuance of the design task that states (1) the
process to be computerized, (2) possible tasks of process computerization, (3)
time and space margins for process computerization, and (4) service or the
person responsible for the coordination of work on the feasibility study.
The design task is usually prepared by the process user, owner or designer
but the preparatory work may be carried out independently by a preliminary
study.
The core of the feasibility study is the cost/benefit analysis and all efforts
have to be pointed towards ensuring its correct and unbiased calculation. A
proposal document usually contains the following data.
305
306 The feasibility study
1. Introductory note that describes the organization and type of work on the
feasibility study and the assessment of main study goals.
2. Proposed plan of work, including overall phasing and content of each
main phase of work.
3. Areas of investigation describing
contacts with customer experts;
listing of all potential benefits studied;
listing of all data sources, data transmissions, data control facilities
and procedures;
main content of the computer database, its access, updating and cor-
recting;
future process development;
hardware and software potentially available for the system imple-
mentation;
system implementation schedule;
process control and supervision and data communication;
process optimization; modelling and expert knowledge;
customer organization;
process control power supplies and emergency power supplies;
process control safety and reliability and maintenance;
process control security;
software requirements;
process control timing requirements;
training of personnel;
other data in the process control implementation.
4. Content of the feasibility study report specifying
functional specification of the required control system;
proposed computer control system and communication system for data
and speech transmission;
comparison of the proposed system with other possible solutions;
estimated costs for a phased development of the control system;
tangible and intangible benefits to be gained from the proposed control
system;
implementation schedule;
hardware and software required for the system implementation with
their possible sources;
performance of the proposed system according to functional specifica-
tion, particularly regarding its availability, reliability, security and safety.
The efficiency of a feasibility study technique depends on the type of process
studied, its computerization state, and on staff resources of the process owner,
user and/or designer. Usually there are five stages in a feasibility study (see
Table 10.1) and a simple flow chart of necessary tasks is given in Figure 10.1.
Table JO.1 Five stages of a feasibility study
Data collection Initial design Preparation of System design revision System finalization
requirements and and specification
questionnaires
Questionnaire Hardware Configuration Revision of system Incorporation of revision

preparation on data configuration discussion hardware, software and on presentation
volumes and rates communication
Discussion Software initial Operator-system Definition of the costs Agreement on final
design communication of hardware and specification
requirements software
Control system Process control Benefit discussion Implementation plan Definition of operator
schematic diagram modelling procedures
Plan of study Outline of final Process modelling Cost/benefit analysis Finalization of
preparation report discussion configuration
System expansion, Training support Finalization of system
maintenance, security specification
and safety discussion,
fall-back
Survey of available Installation schedule Final co stings
hardware and software
Data and speech System presentation Production of final
communication report
discussion
Training of personnel Report presentation
requirements
control system
and costs benefits
identification identification of
of requests existing and
and data acquisition future process problems
•
preparation of
functional
•
identification of
potential solutions
.-
specification and benefits
preparation of
preliminary
•
estimation of
main benefit values
design
l
preparation of
costs
•
quantification
of benefit amounts
1
calculation of
~ r--
•
research of
industrial support
costs
for real benefits
costs/benefits
analysis
•
estimation of
system feasibility,
preparation
of report
Figure 10.1 A simple flow chart of necessary tasks in a feasibility study.
10.2 DA T A VOLUMES AND FLOWS
The estimation of data volumes and flows is the first step in a system design.
The most practical way to do this is to split up the control system into
stratification levels and connect it to a specific functional system that will
Data volumes and flows 309
level 1
level 2
process
plant
Figure 10.2 An example of two-level process control stratification: ~,

continuous process plant; CPC, continuous process controller; BPC, batch process
controller;r+, signals and commands.
execute specific process control tasks at each level, as well as starting with
the estimation of data volumes processed at each functional unit. The tasks at
each stratification level may demand different functional units; it seems that
the process control functions for a specific process require the same or a
similar type of functional unit. Figure 10.2 shows an example of a two-level
process control stratification with two types of functional system, where the
continuous process controllers may be the same. When a particular section of
the process is assigned to the specific functional unit then, based on this fact,
there is a known number of process data and commands connected to this
functional unit.
Data and commands are determined according to the type of process used
and its performance. The usual distribution of process data and commands to
various functional system is outlined in Table 10.2. The definition of each
functional unit is also important, since the interrelations of the various process
units may require additional data transmission from other units. Dislocation of
these units requires a standardized bus structure in a distributed process control
system, or a standardized communication network, for the communication of
process data. Thus, a micro location of each unit has to be defined to prepare
for the required communication facilities.
Input and output process data volumes are defined for each functional
unit, usually as the decimal number representation of each signal or command.
Thus, a necessary consensus on data accuracy can be easily obtained with
process designers. Additional effort and mutual understanding among control
system designers and process designers must be made if a consensus on data
flow estimates is to be obtained. Some processes require a low rate of data
acquisition, and allow slow command executions. For example, one scanning
Table 10.2 Usual distribution of process data and commands to various

functional systems
Functional Status Count Analog On/off Set Count rate

systems pulse/rate command point /pulse output
Measurement
system x
Supervisory
system x x x
Chronological
event recorder x x
Remote
terminal unit x x x x x
Process
controller and x x
basic control
unit
Process
regulator x x x x
Process
governor x x x x x
Special
instrumentation x x x
system
Positioning
system x x x x
Protection
system x x x x
x: applies; -: does not apply
of process data takes 10 minutes or more, and 30 seconds or more for on/off
command execution for a natural-gas network. Data flow estimates are important
for functional units with a large number of data inputs and outputs and with
processes that generate and require control in small time intervals, such as the
control system of a DC-motor positioning unit in a hot rolling mill.
An example of estimates of data volumes and flows for a functional unit of
a crude-oil gathering station is given in Table 10.3, whereby two data flows
can be distinguished: the greater between a functional unit and controlled
process and the lesser between a functional unit and a superimposed control
unit.
Stratification of process control and basic system design 311
Table 10.3 Estimation of data volumes and flows of a crude-oil gathering station
Types of data Types of

data flow
Count Status Analog Set-point On/off
rate commands
Number of 180 200 50 10 10

inputs/outputs
Number of 6 4 5 2
decimal digits
per input
Total number 1080 200 200 50 20 Process
of decimal digits data flow
Estimated 100 10 100 30 30
scanning time, s
Estimated 40 80 8 6.5 3 Total
speed of process z 140 bit/s
data flow, bit/s
Number of 18 20 5 10 10
process inputs/
outputs
communicated
Number of 6 4 5 2 Communication
decimal digits data flow
per input/
output
Communication 2 3 2 2 2
overhead, digits
Total number of 144 80 30 70 40
decimal digits
Estimated 100 10 100 30 30
scanning time, s
Estimated speed 6.5 3.2 1.2 9 5.3 Total
of communication z 25 bit/s
data flow, bit/s
10.3 STRATIFICATION OF PROCESS CONTROL AND

BASIC SYSTEM DESIGN
A crucial point in control system design is the stratification of control which

consists of:
assignment of functional system to specific parts of the process;
assignment of priority levels to various functional systems;

definition of interrelations between given priority levels.
Thus, the stratification basically depends on the process involved in control,
which demands a multidisciplinary system approach to it. The stratification of
the process control for some production processes and plants is given in Table
lOA, along with the usual supposed number of control levels. The stratification
of process control also depends on:
the extent of control demanded from each control system;
the content of control applied in basic process units; the processes based
on special production machines and supplied with enough control facilities
do not demand a large overhead in the process control hardware and
software;
specific demands on regulation and safety measures issued for specific
process purposes;
traditional and proven solutions and techniques applied.
Major demands on control systems of some specific processes are, for example,
as follows:
1. Railway station control includes the number of basic control units equal
to the number of basic process units. Thus, the number of logic operations
and process control modules in the system are minimized; the connections
between modules are standardized; each module performs fail-safe opera-
tions due to possible critical situations in the railway traffic.
2. Thermoelectric power plant includes process controllers and regulators
assigned to each process functional group; control system and governing
system are interconnected for coordination of different thermoelectric plant
subsystems; separate systems for block supervision, measurement and
chronological event recording are also included in plant control.
3. Coal mine controllers are primarily installed for safety purposes; the
supervisory and control system is used for the coordination of safety
measures and control of coal production.
4. Crude-oil measurement station includes process controllers for the execu-
tion of production function, while the local control and supervisory centre
allows shutdown and start-up procedures in the plant.
5. Natural-gas field station includes process controllers and regulators to
execute production function and to send data to the remote terminal unit;
commands for the operation of the natural-gas station are obtained from
the superimposed control centre that coordinates overall natural-gas
production and distribution.
6. Refinery plant includes process controllers, process regulators, batch
controllers and various supervisory systems, while the central control system
determines and coordinates the overall process state.
Table 10.4 Stratification of process control for different processes
Type of industry or Process or plant Product or service Stratification control

service levels
Effective energy Thermoelectric Electric energy Process controllers

production and power plant heating steam Process regulators
distribution Control system
Governing system
Supervisory system
Hydroelectric Electric energy Process controllers
power plant Process regulators
Supervisory and
control system
Regional electric Electric power Transformer and tap
network distribution compensation
controllers
Regional distribution
centre
Coal production Coal mine Coal Mine controllers
Safety control
system
Supervisory and
production control
system
Production of crude Crude-oil Crude oil and Process controllers
oil and natural gas measurement natural gas Local control and
station supervisory centre
Natural-gas field Natural gas and Process controllers
station higher Remote terminal
hydrocarbons unit
Petrochemical Refinery Hydrocarbon Process controllers
industry derivates Process regulators
Central supervisory
and control system
Iron and steel Steel furnace Steel Process controllers
industry Process regulators
Central control
system
Rolling mill Steel products Process controllers
and regulators
Central control
system
Table lOA (continued)

service levels
Cold strip mill Steel products Process controllers

and regulators
Central control
system
Non-ferrous Copper reactor Raw copper Reactor controller
metallurgy furnace Process control and
supervision system
Copper electrolytic Anode copper Process controller
refinery Process supervision
system
Mineral processing Glass furnace Glass Process controller
industry (nonmetal) Batch controller
system
Salt production Salt Process controller
Process supervisory
system
Cement factory Cement Furnace controller
Process control
system
Supervisory system
Metal-working Machine tool Rotating metal part Drive controllers
industry and regulators
Computerized
numerical control
system
Steel furnace Steel casts Process controller
Furnace regulator
Electrical and Isolation furnace Isolating part Process controller
electronics industry
Electronic chip Electronic chip Process controller
production Process regulator
Product quality
controller
Production line for Small rotating Product quality
small rotating machines controller
machines
Chemical industry Production of Inorganic colours Process controller
colours Process batch
controller

service levels
Sulphuric acid Sulphuric acid Process controller

plant Process supervisory
and control system
Pharmaceutical Antibiotic plant Antibiotics Process controller
industry Reactor batch
controller
Process supervisory
and control system
Paper and pulp Paper plant Paper Process controller
industry Process regulator
Process control
system
Supervisory system
Textile and leather Textile machine Textiles Process controller
industry Process supervisory
system
Rubber industry Rubber plant Raw rubber Process controller
Process" regulator
Process batch
controller
Process supervisory
system
Food industry Meat canning Canned meat Process controller
factory Process regulator
Process supervisory
system
Final product
controlling system
Industrial production Alcohol refinery Refined alcohol Process controller
of drinks Process regulator
Supervisory system
Printing industry Newspaper printing Newspapers Process controller
facility Supervisory system
Railway traffic Railway station Railway Basic control unit
transportation Process controller
service Supervisory and
control system
Water and natural- Water pumping Drinking water Process controller
gas supply system station supply Remote terminal
unit with local
supervisory panel
start
freezing
process
recovery
sequence
Figure 10.3 Structure of command procedures for process states of a continuous

cold mill (Kessler, 1987).
10.3.1 BASIC SYSTEM DESIGN
The basic system design is based on a scheme of process control stratification

taking into account each particular process; see Figure 10.3 for a continuous
cold mill (Kessler, 1989). The solution to this task consists of six main topics.
1. Control system hierarchy and coordination tasks.
2. Descriptions of each particular process control level.
3. Descriptions of data processing at each particular process control level.
4. Connection to other computers in the system.
5. Possible hardware solutions at each particular control level.
6. Possible software solutions at each particular control level.
10.3.2 COMMUNICATION HARDWARE AND SOFTWARE
In addition to the basic system design, a study of communication hardware

and software has to be elaborated especially for processes executed over a
larger territory. The communication description includes four topics.
1. Main data processing traffic directions and pertinent data volumes.
2. Communication network(s) for the implementation of data communica-
tion.
Cost/benefit analysis 317
1200-2400BD
~-------'-------------
1200-2400BD
Figure 10.4 Basic design and communication facilities of a hypothetical process

control system: ®, private line to be built; ©, leased line from PIT
network; ®, existing private line; ®, leased line from TV network.
3. Potential communication deliverers or services including the possibilities

for the erection of private communication lines.
4. Alternative communications and fall-back.
A scheme of the hypothetical three-level process control system with the
appropriate communication facilities is illustrated in Figure 10.4.
10.4 COST/BENEFIT ANALYSIS
A cost/benefit analysis is based on a separate estimate of the total control

system costs and benefits, along with comparison, optimization and trade-off
coordination. An engineer's experience and process knowledge is most needed
at this stage of the process development and, where possible, unbiased estimates
and calculations should be supplied, various process and process control models
analysed, and the consequences of their implementation studied and calculated.
The two main financial aspects of computer control applications are invest-
ment costs and operating costs.
10.4.1 INVESTMENT COSTS
Investment costs include all the relevant control system costs shown in Table
10.5 for a four-level computer process control network consisting of:
Table 10.5 Estimated investment costs for a four-level computer process

control network
Equipment Costs Total
Hardware Software Power Project leading,

costs costs supply and implementation
building and installation
costs costs
% % % % %
Controllers 8 1 2 12
Remote terminal
units 9 6 2 4 21
Modems 0.5 0.5
District centre
computers 5 6 0.5 2.5 14
Dispatch centre
computers 2 4 0.5 3.5 10
Communication
system 31 6 5 42
Totals 55.5 17 10 17.5 100
double computer dispatching centre;

seven district computer centres;
100 remote terminal units at process sites;
200 process controllers.
Hardware costs
These include costs for every major hardware element of the whole control
system, except those for process instrumentation.
Software costs
These include all costs needed for the design, implementation and documen-
tation of software, excluding the costs of the on-site implementation, installation
and all costs that appear one year after the initial system installation.
Installation, connection and operation costs

Power supply costs and expenditure for the buildings needed for equipment
operation are the necessary investment costs to allow proper system installa-
tion, power supply connection and correct operation of the system.
Implementation costs
These include the effort and materials needed for the connection of each
computer unit to the system.
Benefits 319
Project leading
These costs include all the effort and materials needed for the planning, phasing
and supervision of all activities on the computer control system development
and implementation.
Installation costs
These include the cost of the effort and materials needed one year after the
initial system installation.
10.4.2 OPERATING COSTS
Annual operating costs

These costs need to be studied in detail to be properly identified. The main
components are the cost of depreciation, maintenance, leasing, and the annual
cost of the building.
Depreciation cost is usually the main cost component since the system
depreciation covers a period of 5 to 10 years.
Maintenance cost includes the cost of all necessary maintenance staff, tools,
equipment and material for system scheduled and on-call maintenance; for
larger space distributed systems this also includes the additional cost of staff
transport.
Operator's costs are added with regard to the involvement of extra operators;
there are always some operators recruited from a previous operator's staff; the
involvement of the computer control system can raise the responsibility of the
operators and thus influence their wages.
Leasing costs are involved usually for the leased communication lines and
sometimes for computer equipment.
Annual costs for building include all expenses for buildings where the control
and communication equipment and extra operators are situated (e.g. heating,
electric energy and insurance costs).
A breakdown of the annual costs for the project detailed in Table 10.5 is
given in Table 10.6. Annual operating costs are approximately equal to 30%
of the investment costs for the example given in Table 10.5.
10.5 BENEFITS
Benefits can usually be estimated by analysing two main categories: intangible

and tangible benefi~s.
Intangible benefits
Intangible benefits are expressed as (1) better working conditions and
organization, (2) greater insight into process condition, (3) knowledge of
Table 10.6 Annual operating costs
Total costs %
Amortization for minimum 5 year period 68

Maintenance costs, including transportation costs 17
Operators costs - 2/3 new operators 10
Leasing costs - communication lines 3.3
Annual costs for buildings 1.7
Total annual operating costs 100
planning, process expansion and (4) similar benefit elements that can sometimes
be estimated as a half of the overall tangible benefit gained by computer
control.
Tangible benefits
Tangible benefits are estimated in money terms by means of an approximate
calculation of the expected goal of the computer control system application.
The following four methods of benefit estimation can be used:
1. An analytical method that includes the knowledge of analytical expres-
sions of the process to be computerized and analytical formulation of
goals obtained by computerization; by means of such analytical expres-
sions an exact estimation of benefits can be obtained.
2. Experience of the process experts that can indicate some of the best aspects
of possible benefits obtained by the installation and proper use of the
control system in a given process.
3. Literature searches for descriptions of actual control systems that indicate
the minimum expected benefit for the process under consideration.
4. The opinion of independent consultants involved in a specific production
process and plant similar to the process under consideration.
There are many tangible benefits, which may be categorized as:
process production increase;
decrease of operating expenses;
decrease of investment expenses.
A particular benefit may influence more benefit categories, especially intangible
benefits (e.g. better control and work methods and good working conditions
based on efficient decisions at all process operating levels) can be expressed
in money terms after the installation of the control system and can assume a
significant part of the total benefits. Nevertheless, the point of benefit analysis
is to calculate tangible benefits only.
Benefits 321
Table 10.7 Survey of main tangible benefits and the reason for their appearance
Reason for benefit Production Decrease of Decrease of

increase operating costs investment costs
More precise Increase of the Decrease of Lower investment

knowledge of system load factor process energy in system extension
control system consumption
states
Faster knowledge Better dynamic Decrease of Lower investment
of control system control of system process energy in new system
states production consumption components
Knowledge of more Better control of Decrease of Longer life of
detailed process quality of product process energy existing system
variables (narrower product consumption components
tolerances)
Better knowledge- Increase of Decrease of Longer life of
of system maintenance level maintenance existing system
components states (lower MTTR and costs components
higher MTBF)
Better system data Decrease of Lower investment
disposability accounting costs in new system
and manual data components
acquisition costs
A survey of the main tangible benefits and the reason for their appearance
are given in Table 10.7. Since there are large variations in the level of applied
computerized control and in the type of process involved, the main benefits
given in Table 10.7 have to be specified in more detail for each process or
basic process unit and evaluated in money terms by any of the four methods
of benefit estimation.
An example of the quantification of benefits by better knowledge of system
components states is illustrated in Figure 10.5 for computer control of crude-
oil production. Table 10.8 gives the statistical data on the main causes of
failure for three methods of crude-oil production, as well as the mean time
between failures and the time estimated for repair of the failure. A supposed
percentage benefit in the production increase gained by better knowledge of
system component states is given with the following expression
APR 100
=- - [ PRa,b,e,
8760
L ARdAi + PReAR4/A4 + PR b L ARdAi
3
i=! i=5
7
(10.1)]
where A PR is the yearly expected benefit from the increase of crude-oil

production, PRa,b,e is the total yearly production of crude-oil by means of
eruption (PRa), gas lift (PRb) and submersible hydraulic pumps (PRe) for
different fields; ARb in the following expression
eruption
PRa
equipment paraffin cleaner

failure, and layer
compressor repairment,
failure, pipe corrosion
failure of failures
valves
failure on
submersible
hydraulic
pumps
Figure 10.5 The state of system components that determine quantification of

benefits in crude-oil extraction.
L\Ri = Ri before - Ri after> i = 1, 2, 3, ... , 7 (10.2)

is given as the differences in the total down-time before (Ri before) and after
(Ri after) the application of the computer control system; Ai is a mean time
between failure for different failure types.
The difference in a faster failure discovery is due to on-line data acquisi-
tion. The variations in the time used for part-issuing and repair are due to the
expected faster response to failure because of computer message switching.
According to equation (10.1), a total of 2.9% in the crude-oil production
increase can be estimated for the considered group of crude-oil fields.
Nevertheless, a pessimistic estimated figure of 1% in the production increase
is considered to be more realistic.
References 323
Table 10.8 Estimated values or upper and lower bounds of MTBF and MITR before
and after control system installation for crude-oil fields
Repair time (hours)

Failure MTBF (year) Failure Part Repairment
type discovery issuing
B A B A B A B A
Paraffin 12-24 1-24 0.5-6 50-100 40-100

cleaner
failure
Layer 8 8 4500 310
repairment
Pipe 3 3 12-24 1-24 0.5-6 50-100 40-100
corrosion
failure
Failure on 0.5 0.5 12-24 1-24 0.5-6 50-300 50-200
submersible
hydraulic
pump
Surface 30 30 24-48 16-40
equipment
failure
Compressor 9 9 12-24 1-24 0.5-6 72 64
failure
Valves in 3 3 12-24 48-72 24-48
a casing
failure
B = before computer application

A = after computer application
The comparison of costs and benefits gives the main results of the analysis:
the feasibility of the planned computer installation;
the extent of the computer hardware and software to be installed and its
installation phasing in order to fulfil the expected benefits.
A partial computerization of a process that can exhibit higher benefits is
sometimes preferable over the total planned computerization. Yet, only a realistic
cost/benefit analysis is the basis for the implementation of any computer
control system.
REFERENCE
Kessler, R. (1989) Automatisierungstechnische Praxis, 31, 10, 461.
11
Computer control system design
11.1 INTRODUCTION
The design of computer/computer communications, remote terminal units,

process controllers, data transmission units and control dispatching centres is
a major design activity for all large and many medium-sized computer control
systems. Small and medium process control systems only partly demand this
type of work since communication design can be omitted because these systems
do not demand this facility. Nevertheless, some additional activities exist that
are common to all main designs of all large and many medium control systems.
They are as follows:
1. Elaboration of technical documentation of the control system and com-
munication installations for
obtaining building requirements for the system erection;
obtaining designers' requirements from water supply authorities,
electrodistribution authorities, communication and other official au-
thorities and services;
obtaining permissions for the system installation parts to cross over
roads, railroads, waterways and similar constructions.
2. Design project for different crossings of the system installation over roads,
railroads and waterways.
3. Design projects of special system objects and constructions.
4. Design projects of typical system objects (e.g. remote terminal housings).
5. Elaboration of technical conditions for system erection.
6. Elaboration of safety protection conditions and protection countermeas-
ures.
The basis for the design of technical documentation of the process or plant
control system is:
conceptual design of the process or plant;
specifications for enquiries for basic process parts (e.g. reservoirs, mani-
324
Communication design 325
folds, vessels, pumps, valves, transformers, motors, switch-gears, fittings,

isolators, different auxiliary tools);
chart of the supposed system installation site in different proportions be-
tween 1:5000 and 1:50 000;
time schedule and content of the appropriate technical documentation;
process control requirement specification.
The contents of general technical documentation usually required for the main
design of process control systems are given in Table 11.1. Documents are
needed for protection of work, health and safety personnel and the environment,
explosion proof and flame proof protection, fire protection and protection of
electrical and telecommunication cables, maintenance of the proper working
conditions of a system and all similar protection countenneasures.
The necessary general documents are usually obtained from state or local
government authorities, or inspectors' offices, and are normally prepared by
professional design organizations with the help of the user's designers. The
documents serve as the basis for the system creation and for a more elaborate
system design as described later in this chapter. The results of the main design
are then used for the completion of the whole control system, since the de-
sign data on equipment are used as the specifications of general technical
documentation.
11.2 COMMUNICATION DESIGN
Data and speech transmission is a typical demand on communication connec-

tion. An example of the connection of two data transmission stations to a
master station or to a dispatch centre is given in Figure 11.1 for a radio link.
A similar topology, and a much simpler situation concerning the equipment
needed, would exist for a wire connection on leased lines or private lines
where modems would be connected to the cable line equipment. Thus, to
allow proper communication of process data and commands from the data
transmission unit to the master station and back through a radio link, the
design of communication links has to be perfonned. The main activities on
the design of communication links are given in Table 11.2.
Each cable communication connection has to be defined in the system with
its microlocation on the map. The appropriate data transmission speed has to
be calculated according to requirements for the given data volume, scanning
and refreshment speed, and to the speed and amount of commands to be
issued. The data transmission speed has to be unified against the expected
quality of data transmission. The alternative connections for data transmissions
(e.g. public, leased or private lines) have to be taken into consideration and
applied when necessary alternative connection is expected in case of fail-
ures on regular data connections. A very valuable solution is to use speech
Table 11.1 General technical documentation
Process documents Process crossings Documents of typical Safety protection

documents process objects documents
Content of process Process design Process crossings design Process object documents Process design and
design documentation documents documentation object documents
Standardized process part Standardized process part Typical object documents Typical object documents
documents documents
Technical conditions for Technical conditions for Technical conditions for Technical conditions for
system erection system erection system erection system erection
Safety protection
documents:
sources of danger
danger estimation
safety design
countermeasures
Basic auxiliary Process isometry map Crossing isometric map Object isometry drawing Design task
documentation
Geographic map for Geographic map for Geographic map for Geographic map for
process approach process approach object approach system overview
Tracing schema(s) Tracing schema(s): Tracing schema(s): City planning conditions
with surrounding Crossing surrounding object surrounding needed
orientation longitudinal map map
List of contents
route profile Longitudinal route profile longitudinal profile
basic design data and Characteristic sections characteristic sections System technical
erection conditions needed needed description
Mechanical drawings
Process crossing schemas Design data on process Design data on objects List of applied technical
with other public utilities in crossing points such as names, numbers, norms and standards
revision numbers
Geological and
geomechanical reports on
soil type and carrying
capacity
Basic equipment Amount and estimate of Amount and estimate of Amount and estimate of List of pressurized tanks
documents equipment equipment equipment
Amount and estimate of Amount and estimate of Amount and estimate of List of equipment with
works as per unit of works as per unit of works as per unit of detailed specification of
work work work danger sources
Mounting and testing Mounting instructions for Mounting instructions for Mounting instructions for Authorization documents
instructions given technical given technical given technical
Report of work for local
conditions at system conditions at system conditions at system
authority
erection site erection site erection site
Nomination of workers
Testing instructions for Testing instructions for Testing instructions for
given system technical given system technical given system technical Document on correctness
conditions conditions conditions of work execution
Building work Building work Assembly log-book
instructions for given instructions for given
system technical system technical
conditions conditions
Technical conditions Typical drawings of all needed details (buildings, electrical, mechanical) Technical conditions for:
Instructions for erection and control of all phases and categories of work protection at work
Annual testing instructions protection of life and
health of people
environment protection
explosion proof
protection
protection of roads,
electrical and
electronic cables
328 Computer control system design
master
station
process
process
Figure 11.1 An example of a connection of two data transmission units to a
master station through a radio link: IT~Rxl, radio transmitter/receiver; [Qj, data
transmission unit; ~,multiplexing- emultiplexing equipment; Dc antenna
equipment; iI, j;, h, h, the frequencies of the radio link.
communication lines as alternative data transmission lines. The detection of

failures on wire connections has to be specified, together with the specification
of equipment for the testing and measurement on lines. The supervision of the
communication network and the preventive maintenance and equipment also
have to be described.
The quality of data transmission through the radio link depends on many
factors that have to be taken into account during communication link design.
The frequency plan has to be stated and the description of scanning must take
into account two contradictory facts:
1. Improved scanning facilities are possible with a higher number of radio
link frequencies.
2. A higher number of radio link frequencies enhances the possibility of
interference and crosstalk with other radio link systems.
A preliminary number and type of radio links has to be officially determined
by the appropriate state or government services specialized for issuance approval.
Communication design 329
Table 11.2 Main activities for data communication design
Wire connection Radio connection
Definition of network configuration and microlocations

Quality of data transmission Frequency plan, description of scanning
Calculation of connection
Antenna diagram
Alternative connections in the network, information workload calculations
Detection of failures on data communication links
Testing and measurements of data communication on real connections
Supervision of communication network and preventive maintenance
Specification of the equipment, Technical description of connections,
measurements and works specifications of equipment, devices,
location, infrastructure and power
supply
The description of interference measurements on preliminarily approved radio

links has to be specified by the time of the preliminary definition of the
frequency plan. The measurements of interferences have to support the calcu-
lation of radio link connections. The description of scanning of radio stations
can be defined and preliminary data on the speed of data transmission and
data acquisition can be given when stating the number of frequency pairs.
These data are compared to the expected data acquisition speed, as required,
by the data flow in the system, allowing proper corrections in the frequency
plan.
The quality of signal reception depends on the magnitude of the radio wave
at the receiver, on the shape of the received signal and on the signal-to-noise
ratio. Two main tasks of the radio link calculation are:
1. Calculation of the magnitude of the radio wave at the receiver and the
statement of conditions of its optimum reception.
2. Research of signal distortion in order to apply countermeasures for its
decrease.
Data transmission is suitable through telephone connections since they satisfy
the required bit rate that is usually higher than 300 Bd and transmission secu-
rity required to be between 10-5 and 10-7 • Therefore, the quality of radio link
connections for data transmission can be conveniently compared with the quality
of the standardized international telephone connection.
The radio link connection can be calculated, based on the power relation of
signal to noise, according to the following equation
P = Po - Ko + Go - F + Gp - Kp - C - Z - B - Ar + Arg + V + Wo (11.1)
where
Po is the transmitter power at the receiver output or the peak power at the
output for pulse modulated systems;
Ko is the cable attenuation of the transmitting antenna;
Go is the transmitter antenna amplification factor because of its direction
and compared to the isotopic radiator;
F is attenuation due to propagation;
G p is the amplification factor of the receiver antenna compared to the
undirected antenna;
Kp is the thermal noise of the receiver antenna;
C is the thermal noise constant and equal to 174 dB m;
Z is the noise number of the receiver;
B is the bandwidth influence of the modulating signal;
Ar is the noise gain factor of the receiver;
Arg is the signal gain factor and automatic gain control of the receiver;
V is the gain factor determined by the modulation system;
W 0 is the psophometric constant of the approximate value of 2.5 dB that
is not applicable to data transmission.
An example of the power balance of the radio link for data transmission is
shown in Figure 11.2, which shows an approximate signal-to-noise ratio of 87
dB for the fading of 20 dB.
The design of radio links varies for different transmission frequencies. The
most used frequency bands for data transmissions are (1) super high frequency
band (SHF) from 3 to 30 GHz, (2) ultra high frequency band (UHF) from 300
to 3000 MHz and (3) very high frequency band (VHF) from 30 to 300 MHz.
A similar procedure may be applied to these frequency bands for radio link
calculation. The basis for calculation is given in equation (11.1). The basic
unknown variable is F, the attenuation factor of the radio link or the magnitude
of the radio wave at the receiver location. The basic procedure for the calculation
of the radio link attenuation factor is given in Table 11.3. An example of the
radio link calculation is illustrated in Figure 11.3 and further detailed in Table
11.4.
11.3 DATA TRANSMISSION UNITS
The main design of data transmission or process control units consists of a

number of steps. The documents needed for data transmission units are applied
to a specific process control, as set out below.
1. General description of functions and tasks of data transmission units.
2. General description of hardware.
Data transmission units 331
80 dB m
60
Go
40
r~
p
20 0
F
0
Arg
-20
~~~r-----r----
Gp :
I
-40 L.-_~ I
f L.._ ~
I
I
I signal/noise signal/noise
-60 I I ratio without
,L.. __ JI ratio with
fading fading
r;~,---.1.-1----L1_
-80
-100 V I
-12 .J Ar ~----------------------
Wo
~
-160~ B
-180P
Figure 11.2 An example of signal-to-noise determination for a specific radio link.
3. Description of connection to process devices and final control elements.

4. Description of data processing functions.
S. Description of the communication with the superimposed computer level.
6. Description of the mechanical construction of the unit.
7. Description of the power supply, working condition and environmental
working conditions.
Table 11.3 Calculation of radio link damping
Steps Name of step Comment
The tracing of the radio According to disposable microlocations of

link profile transmitters and receivers - the determination
of the link with the free first Fresnel zone; the
effective earth radius is 4/3 of real radius
2 The calculation of the free According to the relation for the isotropic
space damping factor radiator
F = 20 log 4m!A
r - modified distance between antennas, m
A, - wavelength, m
3 The calculation of the The calculation of the damping is given with
additional damping due to the approximate effective height
barriers on the path
4 For the case of a complete The damping factor is calculated according to
shadow a passive repeater the relation
can be used with the gain
factor G F = FI + F2 - 2G
where FI and F2 are damping factors from the
transmitter to the passive repeater F1 and from
the passive repeater to receiver F2 respectively
5 The calculation of The calculation of damping for higher
additional damping due to frequencies
tropospheric scattering
The calculation of additional clear height of
the radiated field above the earth according to
the relation
h = h l/2 + dldz/a where
h = additional height;
hi = 548~dl . d 2/(d l + d 2). f, the value of the
first Fresnel zone; d I, d2 are distances in km
from path ends;
f = frequency of the transmitted wave;
a = 20 for rough terrain
a ::: 40 for very smooth terrain
I Fresnel zone free
/
~ ~ I-- t::;?
~ I--
~
v--- V /"
f-"
-- I
~
~ V V
~ ~ 1j'V
---------- -------------V\----
V- ---- ---- - ~ ---- 1"""-
V ~ V'v
----- t-'"
....-- ~
--- ---- f' ./\
~~ ~
-- --- / \1
II
~ \
1\ V
/
~/
I "
Ml 0= 17.223km M2
Figure 11.3 Profile of radio connection MI-M2 from Table 11.4.
Table 11.4 Short-form calculation of radio connection between locations MI and M2
Connection place and distance data
Length of connection 17.223 km

Ml M2
Height above sea level 993 m 1283 m
Height of antenna above sea level 1013 m 1303 m
Elevation angle -1 °1'21"
Radio link data
Frequency of radio link 450 MHz

Maximum power of transmitter lOW
Required power for phase modulation
input at signal to noise ratio 20 dB -110.0 dB m
Maximum signal-to-noise ratio 45 db
Type and gain of both radio link antennae yagi, 12.0 dB
Damping of antenna cable for both devices 6.0 dB
Quality of connection
Minimum required signal-to-noise ratio 4.00 dB

Time availability of connection 99.99%
Calculation of connection
Damping of the free space, I Fresnel zone

free (Figure 11.3) 110.2 dB
Damping of cables and connectors 7.0 dB
Damping of filters and tolerances 3.0 dB
Total link damping 120.2 dB
Gain of 1. antenna 12 dB
Gain of 2. antenna 12 dB
Reserve for slow fading for 80% of time 1.8 dB
Reserve for fast fading for 99.99% of time 10.8 dB
Receiving level for signal-to-noise ratio of
40.0 db at 99.99% of time -77.4 dB
Minimum required transmitter power where
signal to noise equals 40.00 dB for 99.9% of time 18.9 dB m
The power of transmitter of 10 W or 30 dB m will completely satisfy the required

specification of connection MI-M2.
Table 11.5 General description parameters of data transmission units (DTU)
Object, plant or part of plant

controlled
2 Type and number of process data Scanning Number of
received, scanning rate rate inputs/outputs
status
count rate
analog
on/off command
set-point
3 Type of communication with
superimposed computer system
4 Type of data polling on side of
superimposed system
5 Type of the work of the controlled
object before and after the
connection of DTU
6 Battery back-up (YIN)
7 Attachment to regulations and
standards
8 Reserve and test equipment,
type, function and number
8. Statement of warranty, descriptions of putting the unit into work, failures,

testing and maintenance.
9. Specification of
data volumes;
each input and output signal;
each process variable;
work and equipment needed;
distances between process connectors and data transmission units.
Table 11.5 gives the general description parameters of functions and tasks of
data transmission units.
A general description of hardware covers the following topics:
1. Description of modules, way of mounting and construction.
2. Description of standard modules and functional units.
3. Description of process connections.
4. Description of unit standard work and standard features.
Table 11.6 An example of input/output subsystems data
Subsystem Connection type Resolution Time Requirement of

and signal requirement process
connection
Analog input Analog input 0.05% ± 1/2 of Time of Resistance in

subsystem signal 4--20 rnA least significant analog-digital DTU part
from process bit, or 10 bit conversion of loop
transmitter; resolution ~ 200 f..I.s ~ 125 n
differential
signal
Analog output Analog output 8 bits Time of digital Resistance in
subsystem signal 4--20 rnA to analog regulator circuit
with power conversion ~ 125 n
supply from and filtering
DTU ~ 500 ms
Digital intput Closed and open Filter time Relay type of

subsystem contacts on constant connection to
process joints; between DTU with
optoisolated 3 and 20 ms maximum
1.5 kV current 0.05 A
and maximum
voltage 60 V
Digital output On/off signals Time delay of Relay type
subsystem for driving of command of output
relay-type maximum connection from
commands 100 ms; pulse DTU with
output signal maximum
300 ms - 3 s current 0.05 A
and maximum
voltage 60 V
Data transmission unit

5. Description of analog and digital input and output subsystems.
6. Description of parallel or serial communication unit and modem for the
connection with a superimposed control system.
7. Description of the power supply unit.
An example of input/output subsystem data is given in Table 11.6.
A description of data processing functions usually only covers data processing
of process data and data processing for data communication. Data processing
of process data is usually time initiated, or process initiated, and every type of
signal is processed by the individual user's program. Common user's functions
are given in Table 11.7.
Communication with the superimposed control level can be realized either
as a parallel communication through a common bus-to-bus converter, or as
a serial communication through a communication unit, modem or a cable or
Table 11.7 A description of common user's functions
Type of process data User's program functions Remark
Analog input data Periodic data acquisition With internal cycle < 1 min
Signal filtering Corresponding filtering
algorithm
Supervision of the signal Dead band of 0.5 + 1%
dead band
Reporting on demand To superimposed level
Status input data Change of state control With internal cycle < 30 s
Status signal processing After shorter time delay
Count rate data Change of state control For maximum input rates
Data processing Acceptance of all input signals
On/off commands Acceptance of command According to own algorithm or
from the superimposed level
Control of on/off output 0.2 to 10 seconds after
command acceptance
Control of on/off command Usually optional function
execution
Set-point value Acceptance of set-point According to own algorithm or
value from the superimposed control
level
Control of set-point value According to given address
Programming of time A time-programmed function
function normally in small steps
radio link connected to other data transmission units, or to a dispatching cen-

tre computer. The serial communication is usually a halfduplex connection
on the handshaking principle whereby the data transmission unit exchanges
synchronization characters for the constant supervision exchange of mutual
communication.
The superimposed control centre, or master station, usually sends requests
to data transmission units according to the given scheme of polling routine,
and the data transmission unit always responds to all questions and messages.
An example of messages from a master station and data transmission unit is
given in Table 11.8, whereby a complete set of different polling routines can
be assembled. These polling routines allow the acquisition of data in a given
cycle scanning time and the given response time to spontaneous process events.
Table 11.8 Messages from a master station and data transmission unit
Messages of master station Messages of data transmission unit
Status check instruction Statue O1essage

Spontaneous infonnation
Required type I priority sche01e COO1pleted cycle response
Required type 2 priority sche01e Executed response
Ti01e synchronizing O1essage Event recorder O1essage
trans01ission
direction
analog data 8 analog values DTU master
station
..
lon/off command
-
a set point
2 3
Figure 11.4 Data transmission by messages, the example of message block types:
m start code; IADRI, address word; 51617 ,5,6,7 - 8 bit data word.
The respond message from the data transmission unit contains the fixed address
of the data transmission unit that serves as the identification label.
The communication of on/off commands and set-point values is performed
by interrupting the scanning cycles. An example of different data exchanges
Data transmission between data transmission units and a master station has
to be protected against noise and crosstalk on communication lines and links.
These are some recommended data on security measures. For example, code
security with single horizontal and vertical parity, and with single parity and
BCH code, bit length checking by supervision of the bit phase position and
the pulse and pause length, and finally start code restriction of the receiver.
The speed of communication can be set to any of the needed standard

communication speeds (e.g. 300,600, 1200 or 2400 bit/s), respecting the time
requirements for process data acquisition, display data refreshment, and time
requirements for the execution of commands and set-point settings in the
system. The communication unit has to fulfil the CCITI V24 recommendations
concerning its input/output characteristics. The modem connected to the
communication unit has to fulfil the demanded standard data transmission
speed and mode of work (synchronous or asynchronous), protection means
against communication line breakdown, and the remote test of the communi-
cation state according to the CCITT V52 recommendation.
The description of the mechanical construction is based on the usually modular
assembly of process control systems. The modules are put into standard racks,
which are assembled into either stand-alone or wall-mounted cubicles. Access
to all modules should be possible from at least one side. The process cabling
has to be defined, with cabling inputs to the cubicle and mechanical protec-
tion means of cables given. The degree of protection of cubicles is usually IP
54 that ensures protection against water splashing from all directions, protection
of operators against contact with live parts, and protection of electrical circuits
against solid foreign particles.
Power supply and back-up facility is described in the voltage and power
requirement specifications, along with the way it connects to the data trans-
mission unit.
The data transmission unit is usually installed in a separate room with a
temperature ranging from O°C to 40°C and humidity margins (less than 95 %
of relative humidity). The permissible content of dust and chemicals, and
possible water splashing, have to be specified since they should not exceed
the given protection degree.
The warranty, commissioning, failures, testing and maintenance have to be
described and supported by a whole row of procedures fulfilled by the sup-
plier, such as:
separate testing of electronic components, modules and data transmission

units;
simulation of all process data inputs and outputs of the controlled process
according to the given factory test;
commissioning the system at the system purchaser's site with a detailed
description of tools and equipment and complete procedures needed;
check-up of the entire control system from the system deliverer during a
mutually agreed time period.
The definition of failures in data transmission units has to be given, as well

as the measures for their detection and repair. The testing equipment should
allow the identification of the following situations for each data transmission
unit.
Table 11.9 Expected result of data specification and system description
Description Expected result of data specification Countermeasure

and description
Description of Functions and tasks not completely Do not begin with design
functions and described until all present and future
tasks New functions added later functions and tasks are
known
Description of Description not based on any specific Specify hardware system
hardware hardware system
Description of Lack of precise catalogue data on Contact device and
process devices process devices and final controlling element producers and
and final elements obtain exact data
controlling
elements
Description of Vague definition of data processing Insist on unambiguous
data processing functions description of data
functions processing functions
Description of Standard data given only Insist on proof of each
mechanical given data
construction,
power supply
and working
conditions
Specification of Signal representatives, mean distances, Check each signal,
each input main process variables, standard work distance, process variable,
signal, distance, and equipment specified and specify actual work
process and equipment needed
variable, work
and equipment
needed
1. Content of each memory location.

2. Content of each register of the processor.
3. Two-way communication between data transmission unit and master station
by means of simulating both systems separately.
The repair of the unit has to be simple, as is, for example, the changing of the
faulty functional module.
The description of data transmission units is not as ideal as that stated
above, and Table 11.9 shows the expected results of an actual description
procedure and some obvious countermeasures. Table 11.9 also shows the re-
petitive and tedious efforts needed to obtain the required specifications.
Designing control and dispatching centres 341
11.4 DESIGNING CONTROL AND DISPATCHING CENTRES
The design study of control and dispatching centres usually contains the fol-
lowing subjects.
1. Hardware description.
2. Software description.
3. Description of function perfonnance.
4. Hardware and software specification.
Hardware available on computer market and additional data on process re-
quirements are used for the improvement of hardware description obtained in
the basic system design phase.
Software description has resulted from the functional description of the
system given in the feasibility study by its extension and specification and by
adding the parts connected with the software already available on the market.
A control and dispatching centre represents the highest level of the produc-
tion control system where all strategic decisions concerning process control
are executed. Strategic decisions are taken by the production management and
represent the official production policy. The connection between strategic de-
cisions and their implementation in everyday production practice is maintained
by process dispatchers through their control and dispatching console usually
supplied with the following:
1. Process video display monitor(s).
2. Printer(s) and tenninal(s).
3. Hard copy unit.
4. Message exchange facility.
5. Functional and alphanumeric keyboards.
6. Telephone connection facilities.
The state of the process is shown on the print-outs and video display moni-
tors. Data on the process state are gathered in the centre by the communica-
tion equipment or the common system bus from remote tenninal units, or
distributed data processing units and controllers. The refreshment of the pro-
cess database is perfonned in time intervals between 5 seconds and 1 minute
for analog and status data and between 1 minute and 10 minutes for counterstate
data. The acquired process data are presented on monitors in the fonn of
mimic diagrams or data sets. Important process states presented on monitors
can be documented by copies issued by a hard copy unit. Process data in the
database are analysed and composite alarm states issued as warnings, alanns
or critical alanns for each process event that requires such data processing.
The established alarm is displayed to the dispatcher on monitors or a print-out
of a line printer or printer. Historic data fields are automatically created for
certain important process data during the last 12, 24 or 48 hours, as required.
VI COLD
AIR
HOT
STEAM
(8 bar) It V2 FROM/TO
COWPER
OXYGEN
(12 bar)
TO BLAST HOT
FURNACE f - - - - - - - - . . : >....... AIR
Figure 11.5 Cold air, steam and oxygen control of a blast furnace - a block
scheme for panel presentation.
These historical data fields can be shown on monitors and logged at the op-
erator's demand. The dispatcher can issue set-point settings and on/off com-
mands to process final control elements on the basis of the given process
strategy, process data, historical logs and his/her own control and process
knowledge and experience. The issuance of set-point settings and on/off
commands is characterized by time and process variable values and both are
controlled by the dispatcher on process monitors. The automatic process soft-
ware mechanisms ensure the process integrity.
Design of control software is based on process algorithm definition (see
Figure 11.5 for a part of blast furnace control), detailed control scheme (see
Figure 11.6 for the same control part as given in Figure 11.5) and on imple-
mentation. An example of a PID control algorithm software module is given
in Figure 11.7.
The dispatching centre should also allow the following additional func-
tions:
1. Message switching between operators at various points in the system.
2. Centralized execution of larger engineering calculations.
3. Issuance of preliminary pro forma invoices on quantities of consumed or
sold product to customers.
4. Statistical elaboration of dispatchers' work.
5. Modifications and developments of software for remote terminal units,
district centres and control or dispatching centres.
OXYGEN
0,
close:
r---~---,
valve 2 ~---6---------""1
i
lb J
L _______ ( _
L _______ Y!l_____ J
ss
valve 2
__ -. alarm +
u
manual
j L~
.Q._--V
------PLC control part
Figure 11.6 Cold air, steam and oxygen control of a blast furnace - details of
functioning.
The features and functions of the centre are based on real-time operating
systems that usually provide multiprogramming, priority scheduling,
multitasking, disk-based operation, checkpointing, power failure restart and
contingency exits (Table 11.10). These facilities improve the performance of
the system, increasing its peak load capacity while maintaining its ability to
meet absolute real-time deadlines.
The servicing of input/ output data processes is a very difficult aspect of
application programming. The operating system can support the efficient solution
of input/output programming efforts and requirements by supplying a com-
prehensive input/output device independence. Device independence is sup-
ported by the system software as outlined in Table 11.11. The goal of device
independence is to provide a package of services that allows the construction
of a task that can substitute devices routinely without necessitating a single
change in the task's code.
CONNECTION PlAN
17EA YHO
19EA YHU
~
15EA YN
18EA YH
+
rrnfd r:
n
2EA TN/TA
+
XD e-- Norm. x
n
~
I ~
n
1 EA KP
3EA TV/TA Y 3M
~
4EA V
j-B-G]- -=--
j
L n
13EA XDIF f----
HI' - Norm. X
DY
28EB
ACHD
YNF
• t
16EA YR
Figure 11.7 Controller block; PID algorithm implemented on Teleperm M AS230

control system from Siemens: ACHD, automatic/manual mode; KP, proportional
coefficient; TA, scan time; TN, reset time; TOTZ, dead band of positioning
increment; TV, derivative action time; V, D part gain; XD, effective control
difference; XDIF, D part input; YA, value of manipulated variable; YH, value of
manipulated variable in manual mode; YHO, upper limit of manipulated variable
in manual mode; YHU, lower limit of manipulated variable in manual mode;
YN, corrective manipulated variable; YNF, correction of manipulated variable;
YR, actuator position; DY, positioning increment.
The following allows a further expansion of the possibilities of system

programming.
1. Sharing of common routine is used when a number of tasks include a
common routine by means of a mechanism that permits all tasks to share
a single physical copy of the routine code; such shareable libraries are
installed in an operating system at the system generation.
2. Support of user-prepared re-entrant routines is included in the system
when user-prepared library routines are re-entrant.
3. User control of external task scheduling is performed after task request,
running for a given delayed time after the annunciation, or after clock
unit synchronization, at a specific time of day, instant or immediately.
4. Operator control of the operating system gives the system operator full
and simple access to the system, provides for emergency on-line software
fault servicing, for initialization and control of the system software and
user tasks.
5. Comprehensive error recovery provides for error detection and recov-

ery services during the system operation.
6. Protection of program and data is made possible by the operating system
through a number of check-ups to ensure the integrity of each task.
7. System generation process. Tailoring the system to future requirements
is usually made by means of the system generation process where a
collection of system services is tailored to meet the local process physical
constraints and user's performance.
Table 11.10 Basic features and functions of dispatching centre operating system
Software facilities Feature System advantages Comparable advantages Comment
Multiprogramming Queue building for system Concurrent usage Lack of dead time intervals as The multiprogramming of
resource demands of system resources compared to single tasks is accomplished by
programming systems dividing available memory
into a number of named
fixed partitions
Priority scheduling Scheduling of task according Faster execution of Tasks can be scheduled by Interrupts are created by
to their priority by means of task on higher system clock intervals - such process events thus
interrupts priority level systems lack fast response and enabling faster system
possesses a lot of overlay response
Multitasking Multiprogramming of two or Parallelism of Better condition of system data Multiprocessor systems
more tasks that need to execution of processing need intertask parallelism
communicate among actions such as in order to exploit the
themselves and synchronize data refreshment availability of more than
their activities and searching one CPU
Disk-based operation The extension of the Enables common Building of overlay tasks, Disk-based configuration
executive main memory and file system, check- extension of main memory needs a development of
data interchange medium by pointing and rapid abilities while still maintaining special software measures
means of external disk-based initiation of tasks system response time for conditions of disk
operation requirements memory failures such as
file copying and
maintaining
Checkpoint Preemption and rolling out The loading of The system possesses much Optional per task-based
of lower priority tasks to processor with as higher workload still operation
disk and their later roll-in much work as it maintaining real-time
and restoration at the can possibly absorb commitments
previous preempted point
Power failure restart The ability of system to The reduction of Service disruptions due to The operating system has
smooth out intermittent shutdown periods power failure are often to support the register
short-term power due to automatic lengthy, reducing system restoration by notification
fluctuations by tripping power failure equipment effectiveness and of power failure to some
volatile registers contents restart demanding auxiliary power users' programs by means
and their restoring after supply of contingency exit
power restoration mechanism
Contingency exits Subroutines automatically Improvement of System lacking contingency Contingency exit
entered after occurrence of structural design of exists are poorly responding to conditions are synchronous
an anticipated condition or a program and I/O terminations and to illegal such as an attempt to
anticipated asynchronous response efficiency instructions execute illegal instruction
condition of a task or asynchronous such as
an I/O termination
Table 11.11 Device independence means usually supported by system software
Device Feature System benefit Comparable advantages Comment

independence
facility
I/O language I/O service request for Besides the device I/O language communicates
logical interface with independence it makes with the file system and
devices (like OPEN, CALL, higher language record I/O package
GET, PUT, CLOSE) implementations on
existing operating system
possible
Common file A collection of system File system manages file The substitution of similar
system services needed for I/O storage, i.e. finding, device-types can be regarded
transaction between opening, closing of files, as less flexible than common
programs and named holding file name directory, file system
protected collection of and redirecting a named
records file
Record I/O Managing of buffering, Using the I/O language the The entire process is
package blocking and device control user task communicates its transparent to the user level
functions I/O requirements to record task
I/O package
Record I/O builds the I/O
requests to a device driver
to which the data are to be
sent
Device drivers Device drivers accepts Drivers accept requests New devices can be built Error recovery is included
requests and performs the from the file system, record and included into operating when needed
physical functions implied I/O and user tasks. system
in the requests issued to it
Logical unit Logical unit numbers and The task connects the Higher level languages and Systems based on over 16
numbers physical unit directory are logical unit number to a comprehensive I/O services kwords usually include
Physical unit tables that enable the inter- device driver and a simplify user task both higher languages and
directory group change of devices during physical unit number programming I/O service capabilities
the runtime
12
Cost-effective system selection
12.1 INTRODUCTION
As outlined in Chapter 1 the main effort, aside from that of the system designer
(61 %), comes from the system user (39%).
Three factors contribute to reducing the time required to manufacture and
assemble digital control system.
1. Application of system analysis and synthesis.
2. Application of a continuous spectrum of digital and computer hardware
solutions ranging from relay logic, as the simplest solution, to distributed
parallel processor architecture (Brajak, 1990) as the most advanced.
3. Application of feasible software tools and procedures, including system
design and sophisticated system simulation using workstations and personal
computers.
The application of system analysis and synthesis is discussed in Chapters 10
and 11, and the application of feasible software tools and procedures on
p. 358. The procurement and testing of hardware is outlined on p. 352.
Overall, the tendencies in production and marketing of hardware and software
are the separation of investment cost from software costs after system com-
missioning, and from annual expenditure figures; an increase in the number
of original manufacturer equipment solutions on the market; and an effort to
standardize system components.
Investment costs for hardware and software are given in equation (4.1)
in Chapter 4, where the total sum of hardware and software expenditure or
efforts is constant. Efforts can also be described in terms of manpower or
engineer months, or similar units. The initial expenditure Tn for n identical
systems consists of hardware and software expenditures, that is
nH + S = Tn (12.1)
where H is investment costs of one system and S the total investment costs of
software. The optimum marginal costs of hardware ~H and software ~S
expressed as
350
Introduction 351
n~H = ~S (12.2)
are fulfilled in practice only for simple repeatable on/off process controllers,
process control systems and similar control systems of a rather low complexity
and high repeatability of functions. More interesting results are obtained when
discussing the optimum of marginal costs for unrepeatable control systems,
that is
~H = ~S (12.3)
Equation (12.3) explains the natural tendency of the marginal cost of un-
repeatable process control systems to be the least expensive when hardware
and software marginal costs tend to be approximately equal.
Since there are other expenses (e.g. the cost of software one year after
system operation, SI' and annual expenditure costs, AI), equation (4.1) in Chapter
4 can be extended to
H + S + SI + Al =T (12.4)
The type of expenditure can be minimized bearing in mind that usually
Al = aH + b (12.5)
where 0.05 < a < 0.3 is a servicing costs factor of hardware and b is a constant
number that for known systems is usually between 0.1 T and 0.3 T. The
software costs one year after system operation, S I' are
(12.6)
where 0.05 < c < 0.2. Thus the minimum marginal costs of equation (12.4)
can be set for
~H(1 + ea) = ~S(1 + c) (12.7)
This result shows that by using equations (12.5) and (12.6) the minimum cost
of one control system can be obtained when marginal costs for hardware and
software are equal and when a total variable part for hardware servicing costs
equals the variable part of the software cost one year after installation. Usu-
ally this one year for software is not enough for proper system functioning.
On the contrary, some software systems possess the ability of open-endedness.
These open-ended control systems (e.g. add-on functions, sensors and ac-
tuators) are different for costing since they require a complete separability of
physical and functional system realization. This separability can be obtained
by applying software system solutions such as flexible functional units,
multilevel system organization and system modification flexibility as referred
by the OSI concept. The minimum marginal cost appears when
~H(1/e + a' + b' + f) = ~S(1/e + b' + d') (12.8)

352 Cost-effective system selection
where e is the depreciation period, a' is the servicing cost factor of hardware,
b' is the cost factor of operators referred to hardware and software, f is the
cost factor of hardware extension per year and d' is the cost factor of software
extension per year.
12.2 BUYING AND TESTING HARDWARE
Digital control systems designed for general application are mostly manu-
factured as modular units consisting of modules, racks, panels, cubicles, or a
group of cubicles (Figure 12.1). The distribution of the equivalent number of
transistor functions and basic application design varieties are illustrated in
Figure 12.2. The degree of system integration D given is dependent on the
approximately estimated number of equivalent transistor functions T as the
greatest integer of the following expression
D = [log T - log Yo + 0.9] (12.9)
where Yo is the mean number of equivalent transistor functions of the applied
integrated circuits. Thus, for the system with 108 equivalent transistor func-
tions and the mean number of 104 equivalent transistor functions of the ap-
plied integrated circuits, an approximate fourth degree of system integration
can be expected. Thus, the matching of different process control functions to
the appropriate degree of system integration is primarily defined by the required
number of equivalent transistor functions. The complexity of integrated circuits
plays an important role, since the same number of equivalent transistor functions
can be obtained with a different degree of system integration using different
integrated circuits in the basic system modules, i.e. by the application of
small, medium, large and very large scale integrated circuits.
Nevertheless, all process control systems have to be provided with a power
supply unit, an input/output interface facility and a visual display presenting
the process state and controller state, units composed of elements with a different
degree of complexity of integrated circuits and other electronic, magnetic,
optoelectronic and optical components.
The systems are usually based on the general principle of internal hardware
compatibility which dictates that data exchange must be performed through
standardized bus terminations, lines and connections. The standards applied
for bus conceived systems are not strictly obeyed and accepted, yet the ma-
jority of systems recognize separated data, address and control wires or cables
for parallel conceived buses. Process control modules are uniformly designed
and usually purpose-made (e.g. status or analog input modules). There is also
a tendency to design universal modules with some microprocessor-controlled
functions. A universal module includes status and analog input and output
functions and a communication facility with a central data processing unit
through a system bus connection. The mechanical design of control systems
integrated circuit a basic unit with
a different number of
transistor equivalent functions
printed circuit:- up to three racks mechanically

a basic module circuits
rack up to 30 modules
panel:- a group up to three racks mechanically

of racks put together
cubicle including power supply,

cooling and up to
three groups of racks
a group up to five cubicles

of cubicles ,c----rL----r"'~-____1'--_i!:.....-__f'
Figure 12.1 The modular design of digital control systems.

--
number of transistor functions
10 2 10 4 10 6 10 8 10 10
application systems
10 12
~ ____ ~.,02
: : 10 6
integrated circuit
I :
~----- : I
basic module
t::::_::~_O': controller on a board
computer on a board
t:::::~_____-~.Ir10
controller
rack
computer
i
r----------- 10
5 9
controller with I/O interface
mounting panel
: ~ computer with I/O interface
r----------- controller with I/O interface
I 106 10 10 and peripheral units
cubicle 1 ____________ ~ computer with I/O interface
: ~ and peripheral units
r------------ 7 lOll and external memory
L_______________ }O~ I computer system with I/O
group of cubicles
: ~ interface peripheral units
1
1--- - -- -------- ~ and external memories
Figure 12.2 Estimated number of equivalent transistor functions for different

degrees of system integration:~, range of the number of equivalent transistor
functions.
is mostly based on standardized frames for modules, on racks for the plug-in
of the group of modules and on cubicles for the rack mounting. Process input/
output connections from/to modules are made on the front side of the mod-
ules, which are firmly tied to the basic frame in the rack. The connections to
the process are elastically designed to prevent bursting and wearing of wires
and cables under vibrations and shocks sometimes present at the process site.
The systems designed for the special conditions of electrical, mechanical and
chemical interference demand double and sometimes multiple shielding of
electronic components because of noise and crosstalk effects caused by elec-
tromagnetic incompatibility, or by aggregates of dust particles in the atmos-
phere. The module components are regularly tested for their correct working
functions before being mounted into modules or systems. Incorrect compo-
nents are correlated with the supplier's name and their serial production numbers
to avoid possible systematic errors in system functions. Specific control sys-
tems demand specific components with higher temperature and/or humidity
and shock/vibration margins to be built in and applied. There are still some
components of process control systems that have shown specific advantages
concerning their availability, power consumption features and resistance to
Buying and testing hardware 355
Table 12.1 Commonly used components in process control systems
Preference
Function Commonly used Preferable Avoidable
Electrical resistance layer resistor metal film resistor wire resistor

Electrical ceramic capacitor tantal capacitor electrolytic capacitor
capacitance
Visual indication light-emitting diode segment display electric bulb
(LED)
Internal information combination of bus standardized bus cable connections
exchange and cable connection
connections
Electromagnetic exciter winding power transistor or relay or contactor
power silicon-controlled
transformation rectifier (SCR)
Process pressure electromagnetic capacitor mechanical pressure
sensor pressure sensor pressure sensor sensor
Final position optical position magnetic position mechanical position
sensor sensor sensor sensor
Trimming of digital selection computer input trimmer
system function switches device potentiometers
Volume occupation ultrasound sensor photocell radiation sensor
sensor
environmental damage. Table 12.1 outlines some of the commonly used, pre-
ferable and avoidable components for mounting and use in process control
systems. A similar survey of some input/output devices for process control
systems is given in Table 12.2. Some mechanical input components are not
reliable enough for use in process control systems and, wherever possible, are
interchanged and substituted by more reliable electronic, optoelectronic, optical
and magnetic components.
The hardware part of a process control system is optimized in both its
reliability and price by decreasing the number of integrated circuits and other
electronic parts, enabling simple hardware production and maintenance. The
application of large and very large scale integrated circuits decreases the overall
component number, and changes the type of efforts needed for the control
system design from hardware to software. The development and implementa-
tion of the classical software function in hardware elements allows an easier
programming of process control functions and a faster program execution.
The basic building blocks of control systems are modules. The complexity
and lay-out of modules vary from manufacturer to manufacturer. Some attempts
Table I2.2 Input/output devices
Function Preference
Commonly used Preferable Avoidable
Rotating final regulated induction stepping motor DC motor with

control element motor brushes
Data read-in magnetic card floppy disk paper tape reader
Position encoder absolute position incremental position analog position

encoder encoder encoder
Data input device push buttons keyboard light pen
Display device alphanumeric display mimic diagram on 'blind' scheme

video display unit
Acoustical alarm alarm bell acoustic signal on loudspeaker
terminal
Print-out printer laser printer plotter
to standardize the dimensions and mechanical design of modules have been

made by introducing, for instance, NIM and Eurocard mechanical standards.
However, this field of technology is still being developed, and standardization
is premature. The best known technologies of the process control modules are
the double-layer printed circuit board and SMD (Surface Mounted Device)
technology, with dimensions that vary approximately between 100 mm and
300 mm in width and between 150 mm and 400 mm in length. The number
of integrated circuits and other electronic components also varies from module
to module. The simplest modules (e.g. input status module or output relay
modules) can easily be tested. Other, rather more complicated central processor
modules such as a specialized microprocessor-controlled module from the
axis control of a machine tool control system are more difficult to test. Com-
plicated processor modules can be completely tested only when put under
testing conditions with all the other modules.
The production of double-layer printed circuit modules is outlined in Table
12.3. The percentages of the total production time interval are given as estimated
values for corresponding production steps.
Summing up the particular activities on module production, it seems that
the total efforts of testing take approximately between 50 and 80%, and for
module production approximately between 20 and 45%.
The production of devices and systems from electronic modules demands
the mounting of modules into racks, frames, mounting boards and cubicles, as
well as performing the following tests.
Buying and testing hardware 357
Table 12.3 The production of double-layer (DL) and surface mounted devices (SMD)
printed circuit modules
Operation Description of operations Approximate percentage

of duration of
operations (%)
DL SMD
Electronic component input component control 10 10

test functional control
supplier control
Production of printed board cutting 12 10
circuit board process of illumination, etching
and neutralization
boring and finalizing of holes
final mechanical processing
Mounting of electronic cutting and bending of component 28 5
components terminations
placing and fixation of components
control and packing of nonsoldered
modules
Soldering warming up and soldering machine 5 5
preparation
placing of modules into soldering
machine
soldering
cooling and storing of modules
Correctness test visual test of correctness 24 35
testing under normal working
conditions
repair or destruction of modules
Functional testing functional test of modules under 21 35
critical conditions
repair or scrapping of modules
final test and storing
1. Functional test of a device or system that proves the basic functioning of

modules, devices and systems.
2. Extended testing of devices or systems that should prove the functioning
of modules, devices and systems in an extended range of system working
parameters (i.e. supply voltages, ambient temperature, humidity, vibration).
3. Factor test of devices or systems performed by the simulation of all process
data conditions, usually as the worst case design concerning the number
of simulated process events.
The technical advantages of bus-oriented and modular systems and devices
over purpose-made nonmodular and cabled systems and devices lie in the
change of hardware functions and in smaller repair time intervals. The dis-
advantages of bus-oriented and modular systems and devices are their price
margins, where purpose-made hardwired and cabled systems are more con-
veniently priced for smaller devices and systems containing less electronic
and mechanical parts and demanding less labour for their production.
Overall tendencies in the production of hardware for process control systems
and devices seem to be:
increasing complexity of system modules;
increasing use of microprocessors and microcomputers;
replacement of software functions by firmware or hardware built-in
functions;
increase in the number of specialized systems and devices and their available
functions;
increasing standardization of modules, devices and communications.
12.3 DESIGNING AND TESTING SOFTWARE
Designing and testing software for process control systems is based on the
following system specification documentation.
1. Functional specification, which should give a precise and concise de-
scription of the system's outer functions.
2. Software requirement specification, which should give a precise and
concise description of the requirements of the software function of the
system.
3. Data specification, which should cover the description of each signal and
command that will be presented at any time to the system.
4. Software/hardware interface specification, which should cover the
description of each hardware part such as register, input/output unit
or similar point where data are identified by software functions of the
system.
Data processing specification describes the status, transformation and margins
of each data processing function in the system. This specification is based on
the graphic representation of software objects and relations, such as that given
in Figure 12.3 for the HOOD method (Heitz, 1989).
The global content and interrelation of the different types of documentation
are given in Table 12.4, where it can be seen that there is a considerable
volume of documentation needed to satisfy the correct design and implemen-
tation of software. A special region of activity represents the development of
software tools and systems for the design and testing of software (Figure
12.4). The activities for the software design and testing in Figure 12.4 are
given in Table 12.5. The overlapping of activities (symbolically denoted in
Figure 12.4 by combinations of letters) are as follows:
HOOD Buhr
I~ entry
1\ N=e J ~
Name
/
active object task
Name Name
Op ... 1
Op ... 2
I
passive object package
~ 'include' relation 'include' relation
data data
~~~ EiJ~ EiJ
'use' relation with 'use' relation with
data flow data flow
except 1
I •
exception exception
L
generic unit generic unit
(not represented) alarm (not represented)
alarm
'\ Name
entry
interrupt interrupt
Figure 12.3 Comparison of graphic representations of HOOD method and Buhr
symbols for software tasks and modules.
Table 12.4 Designing and testing software for process control systems
Functional Software requirements Data specification Data processing Software/hardware

specification specification interface specification
Definitions Process definitions Software definitions Data definitions used Data processing Interface definitions
used used definitions used used
System verbal Introductory Global software Data processing Global data Global interface
description functional description specification processing description
specification specification
Sensors Function, number and Limits, ranges, Connection tables; Variable status; Registers; bit and byte
types alarms, print-out, addresses; names variable locations, tables of
specification, timing transformation address equivalences
Final control Function, number and Limits, ranges Connection tables; Variable status; Registers; bit and byte
elements types indications, timing addresses; names variable locations, tables of
transformation address equivalances
Computer peripherals Function, number and I/O actions, timings, Connection tables; Data status and Registers; bit and byte
types indicators, print-outs files formats; data transformation locations; tables of
sources and address equivalences
destinations
Computer input/output Function, number and I/O actions, timings, Connection tables; Data status and Registers; bit and byte
units types constraints data files, formats, transformation locations; tables of
data sources and address equivalences
destinations
Control activities
control state Overall definition of Precise definition of Definition of status, Control data I/O address for control
system control states system control analog and counter margins, inputs and outputs
condition states inputs data for control evaluations,
states comparisons and
tables
control response Response scheme Types of control Definition of algorithms Control response I/O address for system
response for control response data processing interface
time constraints Time margins of Time delays and Timer and delay data Timer data Timer interface
control response functions specification processing addresses
operator control Operator control Operator-system Operator data Operator data I/O addresses for
modifications specification control relations processing operator interfaces
Data acquisition Function, number and Content of Location of data, data Processing I/O modules addresses
status types signalization; alarm parameter specification algorithm
margins
analog Function, number and Content of Location of data Processing I/O modules addresses
types signalization algorithm
counter Function, number and Production content Location of data, data Processing I/O modules addresses
types parameter specification algorithms
System reports and
visualization
reports Report functions Time requirements, Report data Report processing
content, scope _ specification specification
VDU content VDU functions Time requirements, VDU data VDU data 'processing
content, scopes specification specification
Mimic contents Mimic functions Time requirements, Mimic data Mimic data processing -
contents, scope specification specification
Functional Software requirements Data specification Data processing Software/hardware

specification specification interface
specification
System construction
reliability Overall system usage Global reliability Detailed reliability
specification estimate Security interface
security Overall Global Security data Security data specification
operator-system usage requirements specification processing
specification
human factors Overall specification of Verbal description of Operator's data and Operator-system Operator-system
operator-system operator activities operator-system communication data interface
interface communication processing specification
System resource Computer allocations Main memory - Main hardware Alternative

and function requirement, computer design hardware/software
types system solutions
System development
constraints
time System development Approximate System development
planning development time timing
resources Overall financial System price Split-up of system
resources for system costs
installation
methodology Overall description of Definition of Detailed data on system System test data
system erection acceptance test testing processing spec.
System maintenance System maintenance Maintenance timing Maintenance crew and Maintenance software
support and requirements tools specification and procedures spec.
System evolution Expected system Anticipated system Hardware and software
evolution changes modules needed for
system expansion
system
documentation
elaboration
according to
Table 12.4
of system C
definition software
of
application
software
software technical
specification prepared
main design of
software
EF
software F
production
FG
production of
documentation
and system
delivery G
Figure 12.4 Division of activities on process control software design and testing.
Table 12.5 Software design and testing
The name of the The description of Margins of effort (%)

activity the activity (estimation )*
Micro- Mini-
computer computer
A Design of Specification of data and data 10 6

functional processing
specification of The tasks of the system
software The tasks of application
programmes
B Definition of Analysis of existing software 5 5
system software Analysis of new needed modules
Definition of new modules
C Definition of Analysis of existing modules 10 10
application software Definition of new modules
D Development of Analysis of existing development 17 7
software for design software
and testing of Definition of necessary
software improvements and of the
execution of new modules
Execution of improvements or
new modules
E Main design of Block diagram of system on 25 40
software modular basis
Detailed specification of modules
with data inputs and outputs
Description of input and output
data and process data-base
Description of operator-system
communication
F Software production Design of programs and routines 25 22
on software elementary level
Connection of software elements
into modules
Connection of modules into system
G Design of Design of program documentation 8 10
documentation and Design of system manuals
system delivery Factory testing of system work
Putting the system into work
Finishing of program
documentation
Schooling of user's personnel
* For complete software production cycle

Designing and testing software 365
AB includes the fitting of existing application programs into the system

functional specification. It is expected at this level of activity that some of
the user's demands can be modified or even neglected if mutually agreed
upon between system designer and system user;
ABC includes the coordination of system tasks, application program
definition and user's system specification;
AC includes the fitting of the existing system software to the user's demands
and the opposite (e.g. the rearrangement of system hardware demands, or
the modification of some of the user's equipment demands);
BC includes the division of system and application software areas;
BD includes the fitting of development needs of the application program
with the existing software development tools;
BCD includes the coordination of software development needs with the
existing system for software development and creation of the necessary
needs for a new software development support;
CD includes the fitting of the development needs of the system software
with the existing software and hardware development tools and equip-
ment;
DE includes the coordination of efforts on the main design of software by
an already-designed development system. The end of this phase results in
software technical specifications: this phase covers the crucial transition
from functional specification of the software to the software structure given
in tasks and modules, i.e. the transition from 'what?' to 'how?'. The ac-
tivities in this phase demand creativity and experience as well as much
effort and generate many errors;
EF includes efforts on the coordination of works on the production and
testing of software with the main design of software;
FG includes the efforts on the coordination of works of final software
production and system documentation; very few documentation amend-
ments can be expected after the team leaves the job.
The basis of software production is oriented towards the system functional
specification and system software technical specification. Many procedures,
methods and techniques can be applied to obtain a full agreement between
designed software and its functional specification. These are structural pro-
gramming, top-down software design, bottom-up software design, outside-in
or inside-out software approach, object-oriented programming, HOOD method,
etc.
The design of software for process control systems can be made by three
types of systems:
1. Microcomputer development system.
2. Minicomputer in-house development system.
3. Time-sharing system.
Table 12.6 The advantages and disadvantages of software design systems
Software design Short-form Relative Relative

system description advantages disadvantages
Microcomputer Microcomputer with Software design by Small system

development system disk or floppy disk, system emulation
external memory Hardware
and full support of development
system software for PC-supported
program linking,
debugging and
testing
Minicomputer Minicomputer Many system users System emulation
in-house with disk external Powerful system not possible
development memory, full support functions and
machine of system software, software tools
program library, UNIX-supported
higher languages,
program linking,
debugging and
testing
Time-sharing Computer that Structural System simulation
system supports many programming not possible
terminals and Many system users
possesses the Powerful system
microcomputer functions and
simulation tools, software tools
higher languages,
program linking,
debugging and
testing
The short-form description of each of these development systems and their

comparative advantages and disadvantages are given in Table 12.6. The
connection between these three systems, and the connection of the microcom-
puter development system with the microcomputer boards and modules (Jovic,
1978), allows a multilevel software design by simultaneous hardware
development, system emulation and multiuser system simulation. The de-
scriptions already given in Figure 12.4 and explained in Table 12.5 are based
on the existence of an operating system, a microcomputer software development
system connected to the minicomputer house development machine and to the
microcomputer modules, and on the modular concept of software production.
The estimated values of efforts usually given in engineer month units are
given in percentages of the total programming effort. The number of assembler
language instructions varies approximately between 1000 and 1 million for
microcomputer and minicomputer systems. The complete cycle of the soft-
process
events
emulation
data and time
prom
programming
iI~~:' ~~~;:
t
Figure 12.5 The total software production cycle.
ware design is only applied to new software modules or module parts. The
total software production cycle is given in Figure 12.5.
The essence of process control systems is their interfacing towards pro-
cesses, which implies that a simpler programming can be obtained by the
orientation of system software and operating system towards I/O process data
processing. In such cases the concept of the automatic generation of the process
database can shorten the time for software development when the specifica-
tion of data is known and the decrease in software costs is possible. However,
user's real-time programs, extended real-time programs and parts of operator-
system communication programs can still present the largest part of the modified
or newly produced programs for each new application. Approximately one-
third to one hour of an engineer's time is still needed for each instruction of
Table 12.7 Trends in microcomputer software
Trends in software design The cause and reason of appearance
More frequent application of operating The existence of bigger microprocessor

systems main memory
More frequent application of The existence of feasible crossassemblers
crossassemblers on microcomputers and mainframes
More frequent application of The decrease of microcomputer and
mUltiprocessor systems microprocessor costs
The increase of software solutions by The increase of development of standard
built-in programmable firmware circuits software routines and programs
The dominant application of higher The development of interpreters and
languages for microprocessors compilers for higher languages and their
integration into operating systems
The increase of external memory The increase of reliability and robustness
applications of floppy disks and minidisks
The increase of number of autonomous The increase of microprocessor power
systems with •upwards' communication from 4 and 8 bits to 16 and 32 bits with
bigger instruction set
a newly developed, tested program to be included in the system work and

completely documented (Wolverton, 1974; Infotech, 1975). There is still a
tendency to devote more time to basic efforts for the technical clearing of
software functions than for coding and testing. As a result approximately 80%
of the time is spent on the first activity, leaving only 20% for the second.
Microcomputer software production, indicated in Table 12.7, influences the
decrease in a part of the newly developed software in the total software de-
sign of process control systems. It can be roughly estimated that the part of
newly developed software will not exceed between 10 and 50% of the total
number of source instructions, taking into account that the system software
and a major part of the operator-system communication are already developed.
Approximately 12% of the total software effort appears necessary for software
completion by the user, and for software maintenance up to one year after the
installation and operation of a system which is usually part of the user's
workload.
The testing of software can be performed by many tools and techniques,
the most common of which are given in Table 12.8. The main categories of
program to be tested are at program unit level, routine level, program and
system level, and acceptance level.
12.3.1 TESTING AT PROGRAM UNIT LEVEL

This must assure the programmers that all individual program modules meet
their technical specification. The test plan contains specifications for the routines
Table 12.8 Tools and techniques for testing software
Testing tool or Description Advantages and Remarks

technique disadvantages
Breakpoint Breakpoint returns control The ability to Breakpoint must

of program execution to the follow the states be set at the
breakpoint routine which of CPU registers first byte of an
can 'freeze' the state of Breakpoints can instruction
program disabling interrupts be set only in
and saving the contents of read-write
CPU registers memories
Hardware register The device that displays Fast and accurate The storing of
and circuit state and compares the state of insight into system behaviour
controller different registers and CPU circuit states data and the
circuit parts with the Requires the manipulation data
desired states knowledge of can be very time
digital operations consuming
in the CPU
On-line variable On-line revision of variable Control of On-line
revision and that has been searched by variable state by indications of
variable search its value its value variable states
by value Removal and can be obtained
insertion of but not the
variable reasons for this
Execution The initiation of program The ability of The feature is
initiation and execution can be changed testing only parts important when
transfer of as well as the control of of program larger parts of
control to any program execution program must be
point Execution can be done tested for error
from any program point
Insertion and Program register state The insight into Only small
removal of dumping at chosen points CPU register portion of
snapshot dumps that can be inserted and states at different program can be
removed program steps tested in this way
Tracing of Each call of subroutine and Efficient tracing This tool is
subroutine call each linkage done by of program efficiently applied
and of operating operating system traced by execution when faults in
system linkage this testing tool The reasons and operating system
ways of have to be tested
execution not
properly known
Input/output and The simulation of input! Testing of This tool can
interrupt output data transfers and program only simulate
simulation simulation of program segments to real-time I/O
interrupts simulated process and interrupt
events and situations
other program
interrupts
Testing tool or Description Advantages and Remarks

technique disadvantages
Incremental The execution of program The follow-up The expected

addition of new is done by step addition of program values of CPU
instructions new instructions and their execution in register should be
execution step-by-step known
manner
Not applicable
for higher
program
segments testing
to be used to ensure the completeness of the individual module. The ability

to define test routines clearly is an excellent test of the completeness of the
system technical specification.
12.3.2 TESTING AT ROUTINE LEVEL
This requires the specification of test procedures for the minimum number of
modules or units that form a routine. The integration of modules or units is
tested by simulating other software components already tested during the
program unit testing.
12.3.3 TESTING AT PROGRAM AND SYSTEM LEVEL
The test of the program must include all details of the test plan concerning the
order in which new modules are to be interfaced and included in the proven
system components. Requirements for testing the software subsystem as it
grows with the addition of new modules must also include detailed test stimuli,
along with the desired results of these stimuli on the part of the program
being tested. The test of the software system includes all the necessary data
on program stimuli and responses with off-line and on-line simulating process
data and man-machine communications data.
12.3.4 TESTING AT ACCEPTANCE LEVEL
This must include all the necessary data on software system behaviour at the
time of system acceptance test when installed.
The order of software design and testing is given in Figure 12.6, in which
the work is as follows:
1
all technical specification
of software
development
test ,b~
/ development
test b~
program b3
level 1
integration
.. te-st-an-d--
pre operation
b3
software b 1
system level
!!.l!~~~!.i9!.l
test and
/ preoperation
b4
6
a12
system operation
and maintenance
Figure 12.6 The order of software design and testing (Boehm, 1976).
all - the change of technical specification due to changes in the code at

the program unit level;
ab - the change of function and test of program unit level due to
change of technical specification;
afl - the change of function of program unit level due to change of
code of program unit level;
ar2 - the change of code and test of program unit level due to change
of function of program unit level;
a~l - the change of program unit level code due to changes in routine
level code;
a~2 - the change in program unit level function and test due to changes
in routine level code etc.;
b} - the work on program unit level coding and debugging;
b~ - the work on program unit level test development;
so that the total time needed for software design and test efforts on program
unit level T SEU is equal to
(12.10)
where
Ii is the inverse of the number of engineers or programmers engaged
in the program unit level coding and debugging and
l~ is the inverse of the number of engineers or programmers engaged
in the program unit development test.
The time needed for total software efforts T SE is equal to (Figure 12.6)
4
TSE = Ai . Lo + L (Aj+l + Bj ) L j - 1 + A(; Ls (12.11)
j=l
Where Lo equals the inverse number of engineers or prammers engaged in

technical specification, and Ls equals the inverse number of engineers or
programmers engaged in the program documentation and program docu-
mentation changes.
12.4 HUMAN FACTORS IN SYSTEM ASSEMBLY
Basic activities in the assembly of industrial control systems have already

been described (pp. 352-372) Fulfilment of given delivery terms, keeping to
the financial margins and the control system quality within specified and known
limits, are essential to the professional quality of the produced systems.
The design production and procurement of process control systems are
basically individual work and group work in a working team. A working
team usually consists of people with various skills, knowledge, education and
social background (Berne, 1975) and who perform two different types of work:
external and internal group work.
Human factors in system assembly 373
Table 12.9 Group work classes and processes
Class Nature of State of Group process

group work environment and goal
External work Productive Nonthreatening Activity

Procedural and
creative work
on material
Nonproductive Threatening Defence from
external work external process in
process order to preserve the
group
Internal work Nonproductive Either Defence from
internal work internal process in
process order to preserve the
group
12.4.1 EXTERNAL GROUP WORK
This consists of two main activities:

1. Productive work on group tasks such as software specification and test-
ing.
2. Unproductive work, termed 'external group process', such as the prepar-
ing of lists of group representations at the meeting of a common group
representative.
12.4.2 INTERNAL GROUP WORK
This consists of nonproductive work in a group, termed 'internal group process'.

The basic relations between group work classes and processes are given in
Table 12.9.
The productive work process comprises procedures and activities such as
system specification, testing of hardware modules and discussions on working
plans.
The external work process exists when a group is threatened by external
disruptive forces that tend to change the group state, even destroying the
group. The internal work process in either threatening or nonthreatening cir-
cumstances consists of specific interactions between group members. These
interactions (named 'transactions') tend to preserve the group state. They are
caused by some group member(s) changing the actual group state. Four types
of group can be distinguished as described in Figure 12.7, namely simple,
compound, complex and complicated groups.
simple group
leadership region ~
internal boundary group
membership region space
external boundary
~ external environment
compound group
minor boundaries
complex group
minor boundaries
complicated
group common group boundary
common group region
Figure 12.7 Four types of group.
Simple group
A simple group consists of group space, including a leadership and membership
region, an internal and external boundary, and an external environment.
Compound group
A compound group consists of the same basic group elements as for the
simple group above, and an additional internal and external boundary, which
means more priority sub-levels.
Complex group
A complex group consists of the same basic group elements as for the simple
group above, and additional minor boundaries splitting the leadership and
membership region into smaller subspaces, which means more subgroups.
Complicated group
A complicated group consists of the same group elements as the complex and
compound groups, in addition to at least one group region common to some
other group.
12.4.3 GROUP STATE
A group state is defined as follows.

1. Leadership that possesses the unilateral right to take the initiative, to
impose sanctions and power to enforce them; leadership and group can-
ons control the group behaviour. Basically, three types of leadership exist:
responsible leader, such as the software project leader;
effective leader, such as the most experienced and skilful system
designer;
psychological leader, such as the oldest member in the software team.
Leadership can be classified as delegative, effective, executive, personal,
primal, psychological, responsible, or subordinated.
2. Membership that possesses the state of accepting initiative, sanctions and
power of the leadership; group members demand certain qualities of the
leadership; membership can be categorized in relation to the volition of
the candidates as accidental, obligatory, optional and voluntary.
3. Major internal boundary is the constitutional, psychological and spatial
distinction between different classes or individuals within the member-
ship and leadership.
4. Common group boundary is the constitutional, psychological and spatial
distinction between different groups concerning a common group region.
An individual can belong to many social structures, where the distinction
between various social structures can be quantified by the probability of
prediction of the class of people in one's neighbourhood at any given moment.
The quantification of a social structure is the internal distribution of a member
in a structure. The usage of these two parameters for open congeries and
enclaves is shown in Figure 12.8. The distribution probability offers the
possibility to compare groups of different degrees of organization. A group
with the 'internal' distribution probability 1 may be called a completely or-
ganized group, since there are as many roles in the organizational structure as
there are slots in the manning table (Figure 12.8). This means that each member
neighbour distribution
,
predictivity predictivity
•
,
mass
...
•
~
open
groups
t i
t
crowd
,
•
<1
ttttttttt •
parties
1 ~ enclaves
(~i~!~ it 1 ";;;1
•
~-,_/
t (: i) t
Y
1 1
't~~
,
I
I
I
completely organized
groups
Figure 12.8 The quantification factors for open groups and enclaves.
has a special position different from those held by other members, and knows
precisely the responsibilities and privileges of all the team members.
The distribution predictivity may also be expressed in percentages. In a
software team with a project manager and three equally responsible and
privileged programmers, the distribution predictivity Dp is
Dp = Ns/NM = 2/4 = 0.5 or 50% (12.12)
where Ns is the number of slots (project manager, programmers) and NM the
number of team members (four team members). The higher the distribution
predictivity, the better the group efficiency. Group efficiency is the potentiality
of working economically. Comparing the number of slots with the number of
members, a given group can be termed as 'overmanned', 'fully manned' and
'undermanned' .
12.4.4 GROUP DYNAMICS
Any working team is always under certain group dynamics, in which there are
three main forces:
1. Group cohesion as an organizing force, including internal organizing
proclivities.
2. External pressure and agitation as disorganizing forces.
3. Internal disorganizing proclivities.
The strength of these forces is difficult to measure and evaluate. The amount
of external pressure and agitation can be taken as a constant in an organized
society and may influence the group in a way that causes absences of group
members in a group work. The percentage of such absences measured over
long time intervals seems to be about 3% (Berne, 1975). Internal disorganizing
proclivities mostly manifest themselves as absences of group members in a
group work. The amount of absences due to internal, disorganizing proclivi-
ties measured over a long period of time seems to be about 9% (Berne, 1975).
Thus, the regular group with a proper group cohesion seems to possess the
ability to attract its members for 88% of the productive group work, as measured
by the time parameter.
A working team is organized for particular working activities and its mem-
bers pass through a series of processes from the start of the group's existence,
the most important of which is understanding the particular group states and
elements. The internal process in a group results from the tension between
two sets of mental images:
1. The image of a group member about what the group represents (e.g. a
working team for the production of software for the microcomputer on-
line positioning system).
2. The image of a group as one of the members sees it (e.g. working alone
on the positioning problem with very loose connections with other team
members).
Group organizers must try to achieve the reality of a group existence as close
as possible to the image of what it should be. Because of differences in the
images of group organizers and a group member, tension may exist. The basic
function of group organizers is to regulate group work by issuing the working
canons that can be the same for several working teams. The basic characteristic
of teams working on the design of computer control systems is a rather short
period of existence compared to other types of working teams. Thus, a process
control working team will not produce heroes, as other social groups tend to
do, since the group does not last long enough for this to develop. A process
control working team possesses a number of rules and processes that last a
much shorter time than with other working groups but the following are
particularly important:
1. The role of a programmer's personality is much more emphasized.
2. Listening to the other team members is more emphasized.
3. A programmer's training and schooling are much more important and last
longer (Weinberg, 1971).
4. Negligence is the programmer's worst enemy (Gramatke, 1989).
12.4.5 GROUP AUTHORITY AND INDIVIDUAL BEHAVIOUR
The existence of a group rests on group authority and its culture. Group au-
thority is defined by a leadership and group canon. There are three types of
leaders corresponding to the three aspects of the group structure.
1. Responsible leader who fills the role in the organizational structure and
officially corresponds to parental elements of group member psychology.
2. Effective leader who makes decisions and corresponds to adult elements
of group member psychology.
3. Psychological leader who fits the group's image of a leader and actually
corresponds to emotional elements of group member psychology.
Programming teams seem to have a very flexible leadership since in specialized
programming knowledge there is always a possibility of anyone taking over
the effective leadership.
Group culture is part of the group canon and an important factor in group
cohesion. Group culture can be divided into the following:
1. Technical culture used principally in productive work and including all
sorts of useful artifacts, techniques and intellectual operations.
2. Group etiquette that deals with standards of behaviour and ways of pre-
senting an acceptable persona and of reinforcing and guiding other team
members.
Figure 12.9 The symbolic scheme of common regions of emotional, rational and
traditional parts of group culture: P, parental, traditional, etiquette part of
individual and group behaviour; A, adult, technical culture part of individual and
group behaviour; C, child, emotional, group character part of group behaviour.
3. Group character through which the individual particularities of each team

member can be presented and tolerated.
Group etiquette and character and respect are important for members. While
group etiquette covers the traditional behaviour pattern, group character covers
the emotional aspects of group member behaviour (Figure 12.9).
A software group working on on-line computer applications can be classi-
fied according to parameters given in Table 12.10.
The role of an individual team member is always complex and deserves
special attention. Individual behaviour in a group is based on the ego state
control of an individual and can be classified as extero-psychical (parental or
estimating), neopsychical (rational or cold) and archeopsychical (childish or
emotional).
The structural analysis of each person is necessary for precise and clear
relationships in a team. The main motives for work are also important and,
Table 12.10 Characteristics of a software working group
Aspect Group elements Characteristics Type of group
Structural external boundary permeability open

membership obligatory
internal structure complexity complex
manning completeness fully manned
undermanned
organizational completeness completely organized
structure
Dynamics cohesion state active
strength strong
agitation and potentiality relaxed
pressure
activity intense
Work nature of forces direction productive activity
Authority canon democracy democratic
leadership fluidity fluid
constitution rigidity flexible
rigid
culture permissiveness permissive
firmness firm
Psychological situation necessity obligatory
attitude ego state rational
demeanour seriousness serious
Sociological activity descriptive technical
according to Weinberg (1971) and Couger and Zawacki (1979), they are listed
with almost equal importance as material compensation, work challenge,
working and social conditions, and competence of control and execution.
Teams members are in everyday, mutual interactions (named transactions)
that can be usefully and adequately classified according to complementarity,
direction, pureness and intensity. Complementary transactions can last for
an unlimited time; crossed transactions can stop this process and force the
formation of transactions at other psychological levels.
The structuring of a team member's time includes the following stages
(Berne, 1967).
1. Withdrawal, including fantasies and imaginations.
2. Rituals, ceremonials and ceremonies.
3. Activity, work procedures.
4. Pastimes, group jokes and talks.
References 381
5. Games.
6. Intimacy.
7. Script.
The corresponding classification of behaviour, and three possible ego-controlled
states, limit the number of possible individual behaviours. An individual may
seek social contacts and time-structuring in a team and will primarily struc-
ture the time to obtain the maximum pleasure from work in the group. After
participating in the group, an individual will correct his/her work in the group
according to natural flexibility and adaptability. The share of each team member
is thus programmed by mental group image, social habits, idiosyncratic mani-
pulating patterns, specific long-term goals or, more simply, by the group culture
and individual character structure. An individual team member usually takes
the initiative in group work when recognition is given to the image of lead-
ership. This leads to a series of efforts to match the group image and individual
script where each phase of adaptation is well defined.
12.5 PROGRAMMING TEAMS
There are some generally accepted rules and specific situations concerning
programming teams.
1. The essence of what makes a good program is the same as that which
makes good friendship, understanding and support.
2. When writing programs the programmers are actually tested for their
understanding of the programming language.
3. Good programmers are trained, not born.
4. Teaching of programming should be based on teaching understanding.
5. Computer languages tend to force programmers to think and behave
differently when executing a specific task.
An actual example of software team structuring is presented in Figure 12.10,
and staffing of the project is given in Figure 12.11. The most crucial moments
in project staffing are time instances t3 and t6, since then there are urgent
changes in team structure.
REFERENCES
Berne, E. (1967) Games People Play, Grove Press, New York.

Berne, E. (1975) The Organisation, Structure and Dynamics of Groups, Grove Press,
New York.
Boehm, B. (1976) IEEE Trans. Comput., C-25, 12, 1226-41.
Brajak P. (1990) Personal communication.
bidding and
system reporting to contract
enquiry management agreement
job definition coordination of work

+ ~ system concept definition
+ -
functional specification
operating system bid enquiry data I main software
definition + primal system documentation + design
+
system leadership team
designer definition definition
definition {software and
hardware designers}
hardware
system integration documentation production
,
system programming system integration putting into work system
~l ~ and testing work
database programming minor corrections
~
system
technician
definition
Figure 12.10 An actual example of software team structuring.

time intervals
t2 - t\ t3 -t2 t4 -t3 ts - t4 t6 -ts t7 -t6
team leader II ~
II I
system
designer
I I
I
I
i
software
designers II
II :-------t
hardware design
and engineering
staff
system
technician L
I
II
Figure 12.11 The staffing of the project illustrated in Figure 12.10.
Couger, J. and Zawacki, R. (1979) Datamation, March, 149.

Gramatke, H.P. (1989) Automatisierungstechnische Praxis, 31,11,538.
Heitz, M. (1987,1989) HOOD: Hierarchical Object Oriented Design for Development
of Large Technical & Realtime Software, Manual 2.2, CSI Ingenierie, Toulouse and
Darmstadt.
Infotech (1975) Infotech State of the Art Report: Data Base Systems, Infotech Infor-
mation.
Jovic, F. (1978) ISEMEC 78 Symp., Ljubljana, C2/l.
Weinberg, G. (1971) The Psychology of Computer Programming, Grove Press, New
York.
Wolverton, R. (1974) IEEE Trans. Comput., C-23, 6, 615-36.
13
The integrated approach
13.1 INTRODUCTION
A successful computer control application depends on meeting the very varied

needs of many groups. The initial aims, the design, the hardware, the software,
the communications links, the commissioning and the satisfaction of the process
operators have all to be met. Failure in anyone area can mar the whole
project. This final chapter covers the commissioning and discusses all those
practical problems with which commissioning engineers are so much more
familiar than design engineers. Process control systems are chosen, bought
and applied by specific global decision criteria. Figure 13.1 shows the relative
weighting factors. The most preferable criteria are (1) all from one deliverer,
(2) price, (3) maintenance and (4) total deliverer experience from same or
similar installation (Koch and Hoffman, 1978).
The next crucial point in system functioning is an inevitable connection of
the process control system with process devices (i.e. transmitters, transducers
and final control devices). Many imperfections of process control behaviour
result in the calling out of the process control service staff, although the
imperfections lie in the process and process control interface that are not a
part of the process control system. An example of such a case is given in
Figure 13.2, where the faults occurred on numerically controlled machine
tools. The service staff were called in for 61 % of the system faults, although
if the source of error is known only 37.8% are usually called out. The criteria
for the choice of process control systems and problems of connection to the
process and process control system seem to conceal a much more subtle process
that derives from the hardware and software design phases and that manifests
itself in the phases of mounting, testing, installation, system operation,
commissioning, training of personnel and maintenance. This process results in
the existence of faults, imperfections and misunderstandings of the process,
process control equipment and between the manufacturer and the user. Thus,
a gap exists between the planned and actual efforts and the realizations of
the process control systems. The uncertainty of planned and realized efforts
expressed, for example, in man months and in months both for software and
385
all from One deliverer
tot.1 deliverer
standard
application
programs
Figure 13.1 Relative weighting factors for decision criteria of process control
system deliverer.
mechanica l, h ydraulics
38%
Figure 13.2 The distribution of causes of failure on numerically controlled

machine tools (Baisch and Hellwig, 1979): D ,faults responsible for
NC service calls.
Mounting and installation procedure 387
costs (man-months)
software 175%
./
(
hardware
,--''------..
115%
100%
planned
~ ~ ~
17%
-----
53%
5 25
actual
Figure 13.3 The usual uncertainty of planned and realized efforts and time for the
design of process control systems (Koch and Hoffman, 1978).
hardware, is shown in Figure 13.3. The phases of mounting, testing, installation,

system operation, commissioning, training of personnel and maintenance have
to be carefully planned and related to all the main tasks and efforts. Errors
remaining in the process control system will cost three times more to correct
after the system is complete, than during its production (Figure 13.4). However,
they are responsible for only 1.78% of total process down-time (p. 441) in
power plants (Eitz and Heining, 1989).
13.2 MOUNTING AND INSTALLATION PROCEDURE
The mounting and installation procedure usually copes with two main problems:
1. Provision of the adequate installation team, installation equipment and
tools to secure adequate installation work and operations.
2. Design of adequate installation rules and additional equipment and
accessories to secure adequate behaviour of the installed equipment in
dangerous, polluted, noisy and electromagnetically incompatible process
control environments.
To solve these two problems, the following general procedures may be adopted.
388 The integrated approach
maintenance
r equirements __ - -;. .,.,

'!ll - - - ".,/ '" ,
specific:~~n
'd~ "./
" "
""
, I
oe'it'>b ,," ¢O /
./ ....'/)0 I
".... ~'" I
~~J$' /
.~~ / ~
'" / 0
,
I
Figure 13.4 The relative costs of software errors at different stages of system
design and functioning (IEEE, 1981).
1. Organization of the working team according to actual work needed.

2. Elaboration of technical documentation for the provision of mounting
works containing:
control system structural and functional schemes;
principal schemes of the system power supply, automatic regulation,
signalization and control;
description and technical drawings of all control system parts, such
as cubicles, control boards, central and peripheral computer equip-
ment;
schemes and descriptions of all outer electrical connections and con-
duits between system parts;
detailed lay-out of all system parts and electrical connections and
conduits;
specification of all system parts and accessories, such as sensors,
final controlling elements and automatization accessories, electrical
apparatus, cubicles, panels and boards, conduits and connection arma-
tures, cables for the power supply and signalization, basic mounting
materials, nonstandard equipment;
calculation of mounting accessories, tools and work;
additional explanatory documentation on work, materials, tools and

schemes applied.
3. Provision of adequate technical equipment, instruments and work for
production of mounting part.
4. Mounting of cubicles, panels and all other system parts.
5. Mounting of tubes and shelves.
6. Mounting of electrical cables.
7. Mounting of sensors.
8. Mounting of special process instrumentation.
9. Mounting of final controlling devices.
lO. Control and correction of mounting work including tests for the explosion-
proof, flame-proof and aggressive atmosphere and other special demands.
An electrical scheme of a module is given in Figure 13.5.
The user's personnel involved in the process control installation should
divide their responsibilities between the modernization of existing equipment
and installation of new equipment.
The primary work responsibilities are:
survey of the present field instrumentation, specification of new instru-
mentation, replacement of additional instrumentation, ordering as required;
survey of communication cable routes and ordering of cables (areal and/
or underground);
installation of new instrumentation hardware;
installation of electrical switches and conduits;
interconnection of switches, conduit wires and cables in junction boxes;
preparation of the remote telemetry station and control computer sites
(building, electrical outlets, cables);
installation of all telemetry instrumentation, communication facilities, remote
terminal units and central computer, and check-up of their physical
connections, power supplies and mechanical requirements; the equipment
should be installed so as to allow a continuous unattended operation.
Installation testing and commissioning efforts amount to approximately 35%
of the total system efforts for control equipment in power plants (Eitz and
Heining, 1989). The quality of mounting works is decisive for the operation
of the entire installed system during its whole life cycle. Special care must be
given to the installation of process sensors and communication hardware. The
installation of process sensors requires
definition of instrument and sensor sizes;
definition of process parameters and possible requirement for explosion-
proof instruments;
ability of sensors and instruments to be easily removed and repaired;
Figure 13.5 An example of an electrical scheme of a module.
type of hardware
switches I----~~~~mnnnn'll"m~~~
meters
instrument
cables
R TU and cen tre

sites
telemetry 1--_.",
cables
RTU and centre
installation t----------------------,==:mmrmmrr==
1 5 1 11
time (months)
Figure 13.6 An example of an installation run by a supervision control system:

RTU, remote terminal unit; D, order and delivery time; ~,foreman; ~,
electrician; IIIIJ, others.
ability of sensors to operate efficiently under very different process variable

or process environment conditions;
ease of adjustment; .
ability of field personnel to repair and maintain system components with
the minimum of training and without special tools;
minimum operational problems in connection with process and environ-
ment media, such as viscosity, paraffination, freezing water, insects;
attempt at standardization, if possible.
The sequence of installation and appropriate work is:
1. To determine the delivery date of required hardware.
2. To order equipment according to the expected delivery date, and allow
for the time lag.
3. To establish the sequence of installation and priorities so as to gain ex-
perience in larger process areas (Figure 13.6).
4. To assign responsibilities to the area foreman, area electrician, electronic
technician and electronic engineer.
5. To review frequently the state of the installation work and determine
major problems and their solutions.
signal Zm2
transmitter _
:
'....1 - - - - transmission line - - - - - i• • 1_
,
signal receiver
Figure 13.7 The model oj a transmission line with concentrated parameters in a

noisy environment (VDI, VDEII976): i$, signal or noise equivalent voltage
generator; ~, impedance; Um' U ml , U m2 , measurement signal, V; Zml' Zm20
impedance oj measurement signal transmitter, D; Zl/' Z12' Zw Z220 longitudinal
impedances oj the transmission line, D; Ztl, Zt2, transverse impedances oj the
transmission line, Q; Zcl' Zc2' connection impedance oj the in-phase disturbances,
D; Zc3' connection impedance oj the out-oj-phase disturbance, Q; UN1 , UN20 in-
phase and out-oj-phase noise voltage, V; Zrl' Zr20 receiver impedance, Q; Uro
receiver voltage, V; UDM , the part oj measurement signal at the receiver, V; UDN •
the part oj noise voltage at the receiver, V.
Special care must be given to the electromagnetic incompatibility of the process

control system and its environment. A model of a transmission line with
concentrated parameters in a noisy environment is given in Figure 13.7.
Interference voltage UNl acts at both lines synchronously (e.g. through capacitive
connection) and changes the voltage at both ends of the receiver. Interference
voltage U N2 acts only at one line (e.g. through the galvanic connection) and
changes the voltage in part of the transmission line to approximately
(13.1)
Voltage U N2 is serially connected with the action of voltage Um and changes
the voltage of one line of the transmission line compared to another. The
potential sources of interference voltages are galvanic, inductive and capacitive
connection and electromagnetic radio frequency interferences.
Galvanic inteiference connection For example, when many transmission lines

are connected through one common line, the voltage drop on the common line
varies according to current loops through all other transmission lines and thus
disturbs data transmission.
Inductive interference connection For example, cables of a transmission line

when the voltage is inducted from the variable magnetic field from another
transmission line.
Capacitive interference connection This is where voltage changes on the

transmission line are induced through capacitance from voltage changes on
another transmission line.
Electromagnetic radio frequency inteiferences from 1 MHz to 100 MHz These

are very critical for the work of sequential circuits of the computer and computer
peripherals. They are emitted and received by the specific parts of the hardware
by fast changes of currents and voltages in the frequency range, approximately
between I MHz and 100 MHz; the influence of such an interference can be
measured for each specific installation or modelled by the disturbed parameter
model of the circuits, their connections and mechanical lay-out; the highest
amplitudes of radio interferences are usually measured within the range from
1 MHz to 4 MHz and from 10 MHz to 30 MHz.
There are some common means applied for the protection of process control
installations.
1. Electrostatic discharge protection in the form of surface conductors or
semiconductor materials placed above the computer and peripherals' logic
circuits.
2. Exclusion shielding of computer and peripheral logic circuits from electro-
magnetic radio frequency interferences in the form of a metal cover with
ribbon cable connectors for the signal and power supply connection.
3. Cable shielding method for the separation of galvanic, inductive and
capacitive interferences by means of metal shields around cables. An
example of the relative improvement by cable shielding is given in Figure
13.8.
4. Twisted cable method for the decreasing of inductive interferences. The
number of wires is between two and approximately 30.
5. Grounding of equipment can be of three main types:
grounding by a protective conductor made to prevent unallowable touch
voltages;
neutral conductor that is a star connection of different neutral conductors
of various devices and units;
grounding of the shielding and electrostatic discharge protection that
is a star connection of all shieldings. An example of grounding of a
computer process control system is given in Figure 13.9.
6. Separation of potentials carried out by relays, transformers, optoelectronic
elements and/or fibre optics put between separate parts of the process
control system, usually at its process input and output connections. The
following goals are attained by the separation of potentials:
prevention of galvanic connection and interference;
prevention of the influence of inductive interferences;
installation of symmetric transmission lines;
F (f}Jdb
unshielded cable
o /"II.
-40 ./"" V- /~
/'
/
shielded cable
-80 I
/' measurement
./ sensitivity limit
-120
."
"
-160
;;
""
10
Figure 13.8 An example of relative improvement of interference influence by cable

shielding: length of transmission line 100 m; distance from interference source
10 cm; internal resistance of interference transmission path 10 kQ; internal
impedance of the interference source 100 D; radius of the lead 1 mm; radius of
the shield 4 mm.
- decreasing of electromagnetic radio frequency interferences and stray

capacitances.
7. Decreasing of power supply interferences is performed by the separation
of signal and power supply cables and by placing them rectangularly at
crossing points.
Recommendations for decreasing interferences that have already burst out in
the computer hardware include:
build up of symmetric signal transmission paths in order to press back
synchronous interference;
filtration of input and output signals, as given in Figure 13.10;
software tools for interference decrease such as (Dugi, 1990, personal
communication):
initialization of interrupts and I/O ports at each operating system cycle;
minimum use of interrupts;
check-up of stack printer;
predict even 'impossible' values of variables and pointers;
control the calculation results by an independent program whenever
feasible;
plausibility check-up of all input variables;
refreshment of watchdog in the main program routine;
waiting until hardware watchdog reset occurs after error detection.
- ----------
Ir AC network AC co;;~- -: - - ------ - ----
I connection
I ' space ectlOn space protected
. from---- - -~-- -- - -~-
high fr;q~-ency - --- "1
sIgnal
220V/llOV AC' ~ ". • mt"fmoce, icooo«UoJ
I ~req~i;~cy ~ A ~p~cP I
I ~fiIter
I '
I
I
i.
1 ..... ------1-----
I
• I accumulator
eqUIpotential I
ground
lead :
I cable
zero lead
protective lead
connection at side
of leading body
functional grounding of process
ground devices
Figure 13.9 An example of grounding and shielding of a computer control system.

computer input devices and circuits

A/D converter instantaneous
analog filter digital filter with integrator AID converter
TITI = 10- 1 T/T / T/.1 T = 50 "T/.1t = 10 4
1'\,
10°
3 ~ ,\,
F (w) 10- 1 r'\ 1\ \r\ \ lA,
Y ~ ~
\"'r" \ \f\h,
,,
I'
,,
,
, ,
10- 1
computer output devices and circuits
-I"-
analog filter delay
T/2rrT I T/T
3 ~
F (w) 10-1
'\ \ )",
l
3 ~ /'r'I '
\ I
10 2
w
wT
Figure 13.10 The filtration effects of input and output signals: F( m), filter
amplitude characteristics; m, circular frequency; T, sampling time; T 1 , time
constant of analog and digital filters; .1T, AID integrator time constant; T, pulse
width of the instantaneous AID converters.
13.3 TESTING AND RELIABILITY
Testing of process control systems installed at the site usually begins with a
step-by-step procedure of commissioning specific parts of the system, starting
from the basic process unit level. The need for testing lies in the system's
faults and behaviour not consistent with the system requirement specification.
The main causes of system faults, and remedial procedures, are given in Table
13.1. The testing procedure and commissioning of the whole control system
require about 11 % of the total efforts in the system realization (Koch and
Hoffman, 1978). For this reason, and in view of further problems in system
maintenance, the study of system faults and the reliability of the system is
considered here in more detail.
Testing and reliability 397
Reliability of and consistency with the system requirement specification

seem to be the most sought-after aspect of a control system (Gailbraith, 1977).
Many efforts have been made to produce as good a product as possible.
Failure is the inability expressed as the probability of a functional unit
(component, circuit, instrument, equipment, system, plant) to perform its outer
function under stated conditions for a certain period of time. Failures are
categorized as major failures demanding the change of system working
conditions and minor failures (all other failures). Failures are caused by a
defect-imperfection of the functional unit (IEC Publication 271C, 1985). The
probability density function of failures f(t) is equal to
f(t) = r(t)/N (13.2)

where r(t) is the total number of failures in the stated time interval t, t + L\t,
and N is the total number of equal observed units. The function of the
cumulative failure distribution F(t) is
F(t) = J~ f(t) dt (13.3)
Reliability r(t) is the ability of a functional unit to perform its outer function
under stated conditions for a certain period of time. The reliability is expressed
as a probability, that is
R(t) = 1 - F(t) = J:0 f(t) dt (13.4)
The failure rate z(t) is the ratio of total number of failures in a given time
interval to the number of correct functioning units, that is
z(t) = [F(t + L\t) - F(t)]/R(t) L\t (13.5)
The instant failure rate z(t)~t--*l = A(t) is
A(t) = f(t)/R(t) (13.6)
The mean life time mL(O, 00) equals the mean time between failures for
irrepairable units, that is
mL(O, 00) = J; R(t) dt (13.7)
To calculate real data collected on tested reliability, the following relations

should be applied:
• down-time, tD
K
tD = L L\t k (13.8)
k=!
Table 13.1 The main causes of system faults
Remedies of faults
A B
Physical faults A Component failures Test of modules

and components
B Intermittent Test of system Test of modules and
malfunctions components in system in nonstandard
nonstandard working working conditions
conditions
C External Measurement Measurement of
interference and calculation external interference;
of external shielding, grounding
interference; and screening of
shielding, components and
grounding and modules
screening of
components
Man-made D Design faults in Design calculations Design calculations

faults hardware control; comparison control; comparison of
of actual and actual and calculated
collected parameters of
component components and
parameters functions
E Design faults in
software
F Interaction faults
Remedies of faults
c D E F
Measurement and
calculation of
external interference
to components,
modules system and
communications;
shielding, screening
and grounding of
I/O circuits,
sensors, control
procesing system
and peripherals
Change of shielding Component and
and grounding; modules redesign;
better choice of lay-out redesign
system lay-out
Control the changes Test the system

in hardware for all possible
redesign and exploitation
include them in circumstances,
software change the faulty
behaviour by
change in software
Control the changes Test the system for Better schooling of
in hardware all possible operator operating personnel;
redesign and interactions; change better know ledge of
include them in the man-system system operation and
man-system communication functioning
descriptions descriptions and
include protective
algorithms
where Lltk is the particular unit down-time from the beginning of the output
to the time when the unit was returned to service, K is the total number of
outages for N units that are repairable, thus K ~ N;
• repair time, tR
K
tR = L LltRk (13.9)
k=l
where LltRk is the time spent on the actual maintenance of the kth unit; excluded
are waiting time for spare parts and for getting the maintenance personnel as
well as for system recovery time;
• per cent availability, Ap
Ap = l00(t T - to)/tT (13.10)
where tT is total time the N units were on test, that is
tT = N· tt (13.11)
where tt is the time of the system test;
• mean time between failures, MTBF
MTBF = (t T - to)/K (13.12)
and
• mean time to repair, MTTR
MTTR = to/K (13.13)
The scheme of time domains represented for system reliability under test
conditions is given in Figure 13.11.
System failures can be and usually are random events. There is inevitably
the tendency to estimate, calculate and predict their rate in order to organize
the production process and personnel needed for its correct working. For a
great number of systems and system functional units with a rather small rate
of mutually independent failures, a very good approximation for the unit failure
rate can be given by a nonstationary Poisson process where the nonstationary
failure rate A.'(t) is given as (Jovic, 1972)
A.'(t) = An + Bne - ent (13.14)
Reliability R(t) can thus be represented using equations (13.4) and (13.6) as
A.'(t) = R'(t)/R(t) (13.15)
or as
R(t) = exp [- f~ A.'(x) dx] (13.16)

Testing and reliability 401
total time
~---------~--------~
(' "
operable
down time time
repair awaiting
time repair operating idle time and
time time system switched!
off time '
waltmg
time for
~. ':i0:~on
waiting,
time for
spare getting:
parts maintenance
staff ;
system maintenance I system available
Figure 13.11 The time domain model of system test activities.
The mean time between failures (MTBF) is then given by
MTBF = J: R(t) dt = Jo~ exp [ - J~ A'(x) dXJ dt (13.17)
Practical data on the reliability of process control systems and their parts
are given in Table 13,2 for MTBF and the mean time to failures, MTTF, for
fast repairable systems (Moore et al. 1978; Data General Corporation, 1976;
Musa, 1980),
For
A'(t) = const = An (13.18)
MTBF is
MTBF = l/An (13.19)
The reliability of function of serially connected units Rs is equal to
Rs = R\ . R 2 · R 3 · .. •· Rn (13.20)
The reliability of function R p of at least one unit out of n parallel connected
units is
(13.21)
Table 13.2 Reliability of process control systems
MTBF .Ii MTBF MTTR .Ii MTTR

hour hour/year hour hour/year
Minicomputer Nova, h 4900

Data General Corp. (with front
console)
Central process control system h 300 -300 10 -2
in power generation with disks
s 200 -200 0
CNC machine tool control h 800 10
systems
s 200 5
Microprocessor remote h 8000 20 (total)
terminal units
s
Typical operational software s 50 -50
under maintenance (extended
real-time software)
h = hardware
s = software
The failure rate calculation of a hardware module of the process control system
is given in Table 13.3.
The reliability of software should be expected to increase its MTBF with
time so eventually no single failure should be expected after some time. A
practical case shows that:
sometimes some program or data sensitive faults cannot be found by any
practical process of testing (SouCek, 1989);
some software faults, although well known, are not removed from the
installation and process operators seem to get used to them;
some new software faults are created by the process of system changes
due to the unfulfilment of system requirements or the implementations of
changes in system outer specifications.
Software failures in a real-time process control system application are
automatically detected and the system restarted in a much shorter time than
after hardware faults. This is shown as the shorter mean time to repair, MTTR
(Table 13.2). Software testing procedures must follow the usual software fault
sources (Goodenough and McGowan, 1980):
construction errors, i.e. faults of software components as implemented, to
satisfy expected specifications;
Table 13.3 Hardware module failure rate
(a) (b) (c) (d) (e) (f)

Component type Nominal failure Temperature Electrical load Corrected failure Quantity of Total failure rate
rate X JO-7/h correction factor correction factor rate components in x JO-7/h
the module
Metal film resistor 0.5 1.0 0.4 0.2 60 12

Wire wound resistor 5.0 1.0 0.6 3.0 10 30
Polyester capacitor 0.8 0.5 0.5 0.2 40 8
Ceramic capacitor 10.0 0.5 0.6 3.0 18 54
Electrolytic capacitor 10.0 0.6 0.5 3.0 3 9
Transistor 5.0 1.0 0.6 3.0 11 33
Diode 1.0 1.0 0.6 0.6 12 7.2
Contact 2.0 1.0 1.0 2.0 150 300
Relay 10.0 1.0 0.8 8.0 10 80
Solder joint 0.5 1.0 1.0 0.5 1200 600
Integrated circuit 1.5 1.0 1.0 1.5 30 45
Total module failure rate = 1178.2 x 1O-7/h

1
MTBF = 1/1178.2 x 10-7 = 8.488 X 10 3 h
specification errors, i.e. faults to specify accurately the intended behaviour

of a software module;
functional design errors, i.e. faults to establish an overall design that can
meet identified requirements;
requirement errors, i.e. faults to identify user needs accurately, including
faults to communicate these needs to software designers.
A common objective in process control software testing is to determine whether
a program is correct, i.e. whether the program produces specified outputs at a
specified time when presented with permitted inputs in given operating system
conditions. The tests used to prove the program correctness are based on the
following common principles.
1. Exhaustive test that uses all or, if not possible, a large number of program
input state combinations to test the program functioning.
2. Selected program flow test that uses the main or all program paths and
branches by passing through them to test their functioning.
3. Software fault test that hypothesizes certain software faults and selects
test cases that will fail if the faults are presented.
The reduction of construction errors consists in reducing the number of data
in the algorithm design, data lay-out and access considerations.
The specification of algorithms and data structures is helped by object-
oriented and structured programming and by a software configuration man-
agement model - mostly under the UNIX operating system (Cmkovic, 1989).
The reduction of design and specification errors can be achieved by software
design methods such as control flow, data flow, input and output structures,
internal database structure, stimulus/response threads through the system,
hierarchical reduction of system functions, system state analysis and state
transition analysis, dependencies between program functions and major system
features that should be easily changeable.
The testing of hardware functions can show some design errors and can
also avoid their repetition. They are:
component malfunctions when improperly designed after redesign;
module malfunctions when improperly used in system after redesign;
system malfunctions when improperly designed after redesign. Redesign
cannot influence such system design errors that include ageing effects and
damage of technical components due to exposure to severe and nonspecified
environmental conditions.
Proper system testing based on realistic reliability calculation can avoid large
expense in further stages of system functioning.
13.4 SYSTEM COMMISSIONING
There are many ways of commissioning a process control system. For example,
using a step-by-step procedure, or making the entire system run immediately
System commissioning 405
at lower or simulated production. This task is one of the last major and time-
consuming operations in control system design, resulting in the user obtaining
an insight into the functioning of the whole system. There are many rational
and irrational expectations and worries concerning control system functioning
and features (Koch and Hoffman, 1978). They are set out below.
1. System safety and reliability.

2. Fast reaction time.
3. Larger working memory.
4. System with video display units.
5. Data security.
6. Hardware compatibility.
7. Better documentation.
8. Testing possibility in on-line work.
9. Higher and faster software availability.
10. More comfortable bit handling.
In addition, the users who make their own application programs expect to
have easy learning and fast programming facilities, easy change and add-on
programs, and better transparency, portability and issuing of documentation.
The commissioning procedure may demand fulfilment of some of the user's
specified desires that may not yet have been included in the system design.
The advantage of the software-based control system lies in the possibility of
fulfilling these desires, even if the system has already been designed and
installed. The contractual side of these last modifications is a matter of mutual
consent between the system designer and the system user, taking into
consideration the facts previously put into control and system functional
specifications.
The commissioning and the final testing procedure are usually performed
together. Some parts of the process control system can be commissioned after
the final test, while others are still in the course of testing procedures.
Commissioning usually requires the operation of .the whole production process
with a decreased rated production and a cumulative increase of production in
agreement with the process operator and the control system supervisor. Some
additional measuring and data logging faCilities can be used for the purpose
of process safety and control. These facilities can be pen recorders and simple
visual displays, or can be made by a separate computer measuring and recording
system (see Figure 13.12). In such a case, the separation of analog and status
signals from process devices must be provided.
The preparatory work for system commissioning should be based on
drawings showing the manner in which the equipment is to be affixed, grounded,
accessed and connected to the electrical power supply, air-conditioned and
supplied with all other technical means. These drawings and required works
are done separately from system installation drawings and work descriptions,
and should be described and prepared by the control system contractor. The
t independent process measurement and
~--,---,...--~
recording system
production
process
process control system
Figure 13.12 Supervision of a process control system by an independent functional

system.
system user is usually obliged to prepare all other officially required permissions
for system work and system connection to any public utility network or system.
The expenses incurred during system commissioning usually fall on the
customer.
13.5 COMMISSIONING AND SAFETY
The commissioning of a control system is usually based on taking-over tests

carried out at the end of the procedure. The system contractor or designer
notifies the user in writing when work on taking-over tests is ready, and
should be given in sufficient time to enable the user to make any necessary
arrangements in process or staff appointments. The test should take place in
the presence of both parties. Technical requirements are to be specified in the
contract, or as specified in contract amendments as mutually agreed upon. If
technical requirements are specified differently, then the general practice for
computer process control in the country where the system is installed should
govern. Taking-over tests usually include the following items.
1. Hardware test assesses the functioning of the units connected into the
system. System parts connected on one chassis are controlled by the mutual
exchange of test data. The parts more sensitive to errors in data transmission
(i.e. external memory parts) are checked by special test protocols or
procedures that check the worst case of system behaviour, including the
change of the power supply between guaranteed margins and simulating
the worst case of allowed mechanical and temperature strains.
Commissioning and safety 407
2. Firmware test assesses the read-only memory unit content by using a

parity check or check sum means, and reports on the possible errors found
during the check. The content of some types of read-only memory changes
with time, particularly during the first period after installation.
3. Software test assesses the programs and data stored in read-write memory
and disk and floppy disk memories, usually by applying check sum and
functionality checking, emulating process input data and operator input
data to the programs and activating programs one by one with these
emulating data.
4. Total functional test includes the testing of inner and some outer system
functions, usually in a stepwise procedure starting with the functionality
of data acquisition, reduction, monitoring, print-out and visualization and
ending with the test of functionality of data communication and command
execution. The connection of almost all process sensors and final control
elements allows the completeness of all test procedures; some process
input data can be omitted if not available and simulated from panels and
keyboards to the process database; additional software test features have
therefore to be added to the system.
5. Test of software facilities (i.e. debugger, compiler, interpreter, editor, etc.)
can be performed by requiring their execution and by carrying out small
characteristic programs that test their functioning.
6. Test of system outer functions assesses the system functioning as given
in system requirement specifications. It demands the simulation of all
process data inputs and operator interventions, as they are given in system
requirement specifications; a special documentation mutually agreed upon
by the system user and the system designer has to be previously issued
where these inputs and the operator's data are connected with outer system
functions.
7. Test of process models and modelling features includes the testing of
operator-model communication, presentation of model functioning, test
of model functioning, test and presentation of change of model parameters,
and test and presentation of change of model configuration.
8. Test of the system behaviour takes place when the simulation of process
variables in protective conditions is given. In addition, the test of system
safety functioning is performed when, for example, a voting system (i.e.
two out of three or system sensitivity to data plausibility) is checked.
If as a result of such tests the control system functioning and feature are
found to be defective, or not in accordance with the contract or contract
amendment or any written material issued and mutually agreed upon, the
system contractor should with all speed and at his/her own expense make
good the defects or ensure that the works comply with the contract. If required
by the user, the necessary tests should be repeated thereafter at the contractor's
expense. The installation provision of any instruments, apparatus, raw material,
water, fuel, lubricants of all kinds reasonably required for the final adjustment
and taking-over tests should be provided by the user.
As soon as work on the control system is completed in accordance with the
contract, and has passed all the taking-over tests made on completion of the
system erection, the user takes over the system and the guarantee period
starts to run. The supplier issues a certificate, named a taking-over certificate,
and certifies the data on which the work has been completed and the system
has passed the test.
13.6 TRAINING PERSONNEL
Adequate training of the user's personnel is essential for the successful

management and exploitation of a computer control system. The training can
take place at the system designer's plant, at the user's site or at both sites.
The best insight into system functioning and usage can be obtained from the
designer's engineers at their plant, where courses for training user's personnel
usually take place. The final expectation of the user is concentrated on a
better system maintenance, modification and function extension. The training
can be organized for system operators, dispatchers, system engineers and
maintenance staff. The main topics at the seminars for the user's different
staff are shown in Table 13.4, and the prerequisites, goal and content of a
seminar are given in Table l3.5. Courses should be well supported by equipment
for exercising and system documentation to enable discussions and a fast
involvement of the user's staff into practical problems of system implementation,
operation and maintenance.
13.7 MAINTENANCE
Maintenance of computer control systems seems to be primarily the respon-

sibility of the system user. The only exception can be the computer, its per-
ipherals and leased communication circuits. The company personnel are
intensively utilized on specifying, modifying, purchasing and installing process
equipment, sensors and final control elements, field communication circuits,
company-owned communication facilities and electrical interface units. Trouble-
shooting of the total system to isolate the problem area is also performed by
company personnel. In all these areas there may be a need for a specialist for
the coordination analysis of a particular problem. In particularly isolated areas,
it may be more practical to use an outside firm to service some components
of the system or to help during overhauls.
The user personnel are able to service field equipment more efficiently
through:
equipment standardization;
knowledge of system and process operations;
familiarity with the maintenance area covered.
Table 13.4 Seminar for process control system users
Seminar topic System operator System Maintenance System dispatcher Seminar duration
engineer and engineer (days)
programmer
Minicomputer topics Process control system + + + + 3-5

Al
Macroassembler + + 3-5
A2
Operating system + 3-7
A3
Database systems + 3-7
A4
Programming system
and operator-computer + + 7-10
communication
A5
Microcomputer topics Distributed micro- + + + + 3-5
computer systems
Bl
Basic topics in + + 3-5
programming
B2
Specific systems: remote
terminal units and + + 3-7
process controllers
B3
System hardware and + 5-10
maintenance
B4
Table 13.5 Seminar breakdown
Contents Prerequisites Goal
Al Supervisory and control Basic knowledge Acquaintance with real-time

systems; real-time operating of minicomputer process control systems and
systems; application programs; systems their application
operator-system
communication
A2 Minicomputer instruction Basic knowledge Acquisition of basic skill of
repertoire; micro-assembler; of minicomputer program writing and testing
basic programming and systems tools
testing tools
A3 Elements of operating system, AI' A 2, Knowledge of real-time
basic characteristics of operating system
operating system and system
software, process database
A4 The organization of real-time AI' A 2, A3 Knowledge of real-time
database; data structures; logic database
and physical data organization
As Input/output data AI' A 2, A 3, A4 Knowledge of real-time
communications; program programming system
modules; process database;
operator-system communication
BI Microcomputer systems; Not required Acquisition of basic
system hardware, software knowledge on microcomputer
and configurations. process control hardware,
Programming languages. software and applications
Programming. Fault
location and correction.
Documentation. Applications
B2 Programming fundamentals. BI Basic knowledge of micro-
Real-time programming computer system programming
language and specificities. software and programming
Operating system. tools. Program testing and
Documentation programming correcting
list analysis. Programming
tools
B3 Basic configurations. Data BI Knowledge of specific
structures and communication. systems, data exchange and
Practical experiences. System testing up to unit hardware
features expansions module
B4 Hardware. Software. Fault BI Base knowledge on methods
detection and correction. and means of system testing
System programming. and repair
Documentation
Maintenance 411
I
total/ /
I
1/ I
/
/
/ 1
I I
( 1
operable : down:
r::=L, 1 1
operating idle switched scheduled inoperable external loss time
off :maintenance
I
I
I
1
1
i.hsupplementary
routme
maintenance maintenance
1 ,
\
, ",
r-....L.-----.---.--.," "'" f---,..----,I
"\,
system system rerun miscellaneous awaiting repair system
productive test ' , repair recovery
1
1
1
I
1
1
1
I
I
I
I
production program program

h
fault awaiting!
!
I
• I
production test : recognition maIntenance
personnel:
I
I
~
maintenance
..1
1
.'.
I
available
1 unavailable I
• •
1
Figure 13.13 Time domain model of system availability.
Standardization allows personnel to become more familiar with the types of

equipment, to specialize in the servicing of this equipment, and to calibrate
and maintain adequate spare parts without maintaining excessive stock.
Maintenance is a procedure that, besides personnel and equipment, requires
certain time intervals for its execution. The division of time intervals of the
system operation and down-time is given in Figure 13.13 and the maintenance
duties and respective time intervals are indicated in Table 13.6.
The available time from the user's point of view is a time interval during
which a functional unit can be used, as specified, with its outer functions.
Table 13.6 Maintenance duties and times
Maintenance routine Definition Formal description and Remarks

or interval expected margins
Scheduled maintenance Maintenance carried out in accordance with TMS = t) + t2

an established time schedule
Routine maintenance A part of scheduled maintenance carried out 0 < t) < 10 h t) - routine maintenance interval,
in a routine way h/y
Supplementary A part of scheduled maintenance carried out 0 < t2 < 10 h t2 - supplementary maintenance
maintenance regularly in addition to a routine maintenance interval, h/y
Fault recognition Time interval needed to recognize functional TFR = MTFR The time interval depends on
unit fault - mean time to fault recognition - 0.1 h < MTFR < 24 h system surveillance
MTFR
Awaiting maintenance Time interval needed for maintenance TAP = MITA The time interval depends usually
personnel personnel to arrive at the fault site - mean 0.1 < MITA < 24 h on time of day in a week at
time to arrival - MTTA system fault recognition
Repair Routine and time interval needed for fault TFR = MTTR Mean time to repair depends on
repair I h < MITR < 10 h personnel skill, experience and
schooling
System recovery Routine and time interval needed for system TSR = t7 The time interval for system
recovery after fault repair om h < t7 < 50 h recovery depends on the type of
system under repair
h/y =hour/year
Maintenance 413
probability (%)
100
90
80
70
60
50
40
30
20
10
minutes
10 1 3 10 2 3 103
Figure 13.14 The comparison of probability density functions of short-term

disturbances in an electrical network (Puendler, 1977): (a) data for 1976; (b) data
for 1972; Fd(t) according to equation (13.22).
The external loss time is the down-time caused by faults outside the
functional unit.
The rerun time is the part of the operating time that is used for reruns
necessitated by faults or mistakes in operating.
A mistake is a human action that produces an unintended result, which
occurs by the operator's intervention in an inappropriate way, or the operator's
nonintervention at an expected time. Statistical data for electric power
distribution (Datzkevic and Meltzer, 1978) show the distribution function Fit)
of time intervals of dispatcher interventions for frequency reglilation on busy
days as
1- e-o·76tl.88 for 0 < t :s;; 1 h
Fd(t) ={ (13.22)
1- e-o·76t °.82 for 1 h < t < 00
The probability density function of electrical network disturbances as a function
of time of an actual 110 kV network is given in Figure 13.14, compared to
Table 13.7 Maintenance procedures
Maintenance procedure Responsibilities and procedure Scheduling

description
Instrumentation and process device User's electronic and process On call

maintenance technician
Testing and proving of device
functioning
Replacement of device in case
of failure
Preventive maintenance and User's electronic and process Regularly
calibration of instruments and technician
process devices Testing, calibration and proving
of device functioning
Replacement in case of failure
Instrument and process device User's electronic technican and Regularly
repair designer's specialist
Repair of device and its testing
for correctness
Computer and computer peripherals, Designer's maintenance specialist On call and

maintenance; remote terminals, Testing, proving and functional regularly
maintenance replacement of defective part
System failure location User's electronic technician or On call
system engineer
Testing of system functioning and
localization of failure
Computer and peripherals, repair Designer's engineers and As required
technicians
data given in equation (13.22). The proper behaviour of the process operator,
or the fully automated process control system, can produce the calculated data
indicated in Table 13.7. In other words, the responsibility of the system
functioning, and thus of the system maintenance, lies with system operator for
most computer control systems. Therefore, the training and education level of
system operators and maintenance staff is a most important factor in the system
operation.
The training of the maintenance personnel at the user's site is justified
when the process control equipment is widely used. A better and more reliable
service will normally result by employing user's personnel for this purpose.
New skills and upgrading of skills are required in the field portion of the
system. Maintenance, calibration and servicing of basic process control units,
communication units and process instrumentation is the most required action.
Involvement in trouble-shooting and in preventive maintenance techniques
requires a more efficient use of time and work scheduling and an efficient
References 415
::Ie signer's data ___ hardware

equipmen t------..
software
t
MTBF data
MTTR data
required availability
--
data on rough calculation of the the working
system number and composition strategy of
f-+
,
failure of maintenance the production
critical state and repair crews process
mutual relations
concerning
responsibility
,
and competency
maintenance protocols
and maintenance books
,
for evidence of system
state and maintenance
-- --
designer
dat a designer's data
on spare parts working tools maintenance
and maintenance i--.. and materials instruction
equipment
,
the choice and
schooling of maintenance
staff and their working
tasks and responsibilities
Figure 13.15 The engineering of maintenance procedures and crew planning.
internal communication system. The user's electronic technician or specialist

should be familiar with maintenance, repair and servicing of the computer,
computer peripheral devices, and various input/output devices so as to locate
and determine the source of any problems within the control system. The
maintenance procedures are given in Table 13.7. The engineering of maintenance
procedures and maintenance crew planning is shown in Figure 13.15.
In order to specify maintenance tasks better a practical proposal for
differentiating between electrical, control, mechanical and process parts is
given in Figures 13.16 and 13.17.
REFERENCES
Baisch, R. and Hellwig, F. (1979) Ind. Prod. Eng., 3, 95.
Cmkovic, I. (1989) Automatika, 30, 3-4, 67.
(a)
:C
Type of Mounting scheme
equipment and border Comment
I
Protected t}m - measured
~
temperature m temperature
sensor
P ,
M :
,
Unprotected ~' t}m - measured
temperature m I temperature
sensor I
r:
P I
I
,
________~M~=~I~~C~___________
Pressure M ' C Pm - measured
transmitter pressure
P
Pm I IV instrument
,
valve
,
Flow Qp - measured
transmitter
Q~ flow
---~------
M~ C
r'
(b)
Type of Mounting scheme
equipment and border Comment
Pm pressure
Pm : measurement
IV" t}rn - temperature
Analytic ~m 'S"'~T~ measurement
instrument' ~ IV - instrument
P : valve
M ' c SO - sampling
device
cv - control valve
PM M---: C ___ M and actuator
Pneumatic control
valve ~
.
' ~
)-
I I
PM - pneumatic motor
EP - electro-pneumatic
_ _ _ _ _ _ _ _ _ _ _ _ .-!-:_E_P_ _CV
_ _--,--:-__c-:-o_n_t_r-:-o_l-,d_e;-vl_·c_e_
Electrically
C :x
EM - electric drive
A control, protection
and signalling
operated control CV -
valve
- M--rf
~ J. devices
E~;-- CV - control valve and
,E actuator
E--·;":·_-C ET - electric terminal
Connection part
~
terminals ET CT I CT - control terminal
part
Figure 13.16 Maintenance, the division between control and process equipment.
For the maintenance responsibility part in (a) and (b): P, process; M, mechanics;
C, control; E, electric
(a)
Type of equipment Mounting scheme
and connection and border Comment
Functionally defined E CB CB - control
border on board
control (OD~
!®®i
C ECS - electric
panel I® ®!
control and
ECS Ll:;:-=:0::;_.::..)_ _ _ _ _----1 signalling
.. part
,---------- - - - - - - -
J:J
-------------,---
Total maintenance
CB - control board
responsibility
1
JCT CT - connection
terminals for
- CB
control board
E C
Electric DC e - fuse
power supply E -I- C
R, 0, T - DC power
e I supply terminals
R~'-
T
o---t---~-----
(b)
Type of equipment Mounting scheme
and connection and border Comment
Electrical AC
power supply E .j, C
(monophase)
-:<2> ~:t:: I
j
e - fuse
Electrical e - fuse
three-phase EN - belongs to
power supply control
maintenance
for smaller
installations
Instrumentation M "I' C FC - fluid container

air/gas _ ~D IV I PD - preparation
preparauon 1
FC I Po device
j c Po, ~o, c - nominal values
j of auxiliary
1 ~o medium
1
Figure 13.17 Maintenance, the division between control and

electric and auxiliary equipment. For the maintenance responsibility part in (a) E,
electric, C, control; and in (b) E, electric, C, control, M, mechanical.
Data General Corporation (1976) Reliability Report on the Data General Corporation
Nova Line of Digital Computers.
Datzkevic, Yu. G. and Meltzer, M.M. (1978) Reports to the Academy of Science of
Uzbekh SSR, 15.
Eitz, A.W. and Heining, U. (1989) Automatisierungstechnische Praxis, 31, 416.
Gai1braith, J.K. (1977) The Age of Uncertainty Houghton Mifflin, Boston.
Goodenough, J.B. and McGowan, C.L. (1980) Proc. IEEE, 68, No.9, 1093.
IEEE (1981) Trans. Software Eng., SE-7.
Jovic, F. (1972) Registration of Nonstationary Poisson Process Data, PhD thesis,
Faculty of E1ectrotechnics, University of Zagreb.
Koch, G.R. and Hoffman, R.H. (1978) Ang. Inf, 6, 248.
Moore, C. et al. (1978) IEEE Trans. Power Appar. Syst. PAS·97, 4, 1115.
Musa, J.D. (1980) Proc. IEEE, 68, No.9, 1131.
Puend1er, K. (1977) ElektrizitiitswirtschaJt, 76, Part 17,579.
Soucek, B. (1989), Personal communication.
VDI/VDE 3551 (1976) Recommendation for Security to Disturbances of Signal
Transmission by the Installation of Process Computers.
Author index
ABB 254,295 Gorez, R. 58

Ahmed, 59 Gramatke, H.P. 378
American National Standards
Committee 146 Heining, U. 389
Arafeh, S. 82, 85 Heninger, K. 189
ASEA-ATOM 254 Hoffman, R.H. 10, 385, 396, 405
Hornbach, B. 254
Bailey, S.l. 151, 215 Huelsman, A. 154
Baisch, R. 386
Berne, E. 372, 377, 380 IEC Publications 397
Blickley, G.L. 125 IEEE 388
Bodlovic, P. 151 Infotech 368
Boehm, B. 371
Bogner, R. 62 lovic, F. 351,400
Borevkovic, B. 173
Brajak, P. 153, 350 Katz, M. 125
Brillouin, 291 Katz, P. 58
Kehler, Z. 41
Conover, D.W. 234 Kessler, R. 316
Couger, 1. 378 Koch, G.R. 10, 385, 387, 396, 405
Crnkovit, I 404 Koncar, R. 174, 184, 185
Krueger, H. 251
Dallimonti, R. 234 Kwakernaak, H. 55,56
DATA General Corporation 401
Datzkevic Yu.G. 413 Laduzinsky, A. 183
Debeljuh, V. 181, 183, 186, 187, 211 Leinbaugb, D. 199
DIN Coding of Information 268 Licbtberger, Z. 218
Dugi, 394 Livingston, W.L. 300
Lundberg, M. 183
Eitz, A.W. 389
Evans, C. 143 McGowan, 402
Martinovic, A. 183, 215
Gailbraith, 1.K. 397 Meltzer, M.M. 413
Goodenough,l.B. 402 Mikovic, M. 135, 169, 173, 174, 183, 185
419
420 Author index
Moore, C. 401 Skeen, D. 175, 215

Moore, G.E. 97 Smrkic, Z. 251
Muller-Limmroth, W. 251 Soucek, B. 402
Musa, J.D. 401 Stonebraker, M. 175, 215
Naylor, C. 294
Tanenbaum, A.S. 180
Novakovic, B. 54
Tunon, J. 185
Noyce, R. 97, 98
Peter, K. 125 Vranic, P. 181, 183, 186, 187,211

Peterson, L. 235
Peterson, M. 235
Warren, C. 98
Plecko, V. 121
Weinberg, G. 378, 379
Puendler, K. 413
Wilhelm, R. 268
Wilkie, 351
Reding, 84
Wolverton, R.W. 10, 183, 368
Rus, E. 173
Woodson, W.E. 234
Sajfar, J. 186
Sajfar-Tasic, J. 158 Zelenko, B. 225
Subject index
Acceptance routines 159 quantized 17, 18

Access paths 210 scanning 51
Acknowledgement information 73 time-continuous 18
Acoustical signalization 268, 276-7 time-discreet 18, 21, 22
Actuators 71-2 Analog-to-digital conversion 25-7, 51,
Ada language 183 52,59
Addressing Application programs 183-202, 268
external addressing systems 204, activation 189, 192
205 constraints 188, 189
goal address 213 direct communication 194, 199
hash addressing 212-14 multiprogramming constraints
internal addressing system 204 199-202
modes 143 requirements 188, 189
Alarm acceptance signal 45 structure 189-202
Algorithms tasks 183-9
control 79 timing requirements 189-202
data processing 17, 51, 59, 136 Arithmetic operations 151, 225
process control 54, 257 Artificial inferencing tools 289-97
serial or parallel 152-4 Bayes' statistics 291-2, 297
simulation 225 hypotheses 292, 296
software 188, 189 masoot wanner 292, 293, 294
Alphanumeric display units 241, 245, probabilities 291-2, 296
249, 250 search sequence 295
Amplitude modulation 117 ASEA Information, Sweden 183, 184
Analog devices 65 Automata 28, 37
Analog input circuits 94 Automation 257
Analog input processing 138 Available data processing time 59, 60,
Analog interface units 73 411
Analog output circuits 94 Available time 411
Analog output processing 138
Analog signals 16, 19, 59, 120 Background program activities 166
chahcter set 25 Baseband signalling 117, 119
continuous 17, 18 Basic interval 170
data processing function 260 BASIC language 179, 180
filtering 62 Basic process unit 79, 81, 248
421
422 Subject index
data processing ability 41, 49-62 hardware 111-21,316-17

description 41-9 interprocessor 78
hardware 41-2, 59, 61 modes 112-13
process paths 76 point-to-point 78
signal paths 76 software 316-17
software 41-2, 202-14, 203 speeds 339
Basic system design 316 see also data communication
Bayes' statistics 291-2, 297 Computer control systems 5-8, 9, 10,
BCH code 338 98-101
Bell 200 frequency shifting technique auxiliary field 103
120 block model software 5
Benefit analysis 319-23 costs 8, 11
Binary devices 65 design 324-49
Binary digital signals 18, 19 development cycle 5, 6
Binary expert systems 292, 293 functional systems 5-6
Bit (binary digit) 26, 147 grounding 393, 395
Block model software 5 hardware 5, 8, 104
Breakdown reports 69, 77 input/output structure 103
Buhr symbols 359 interrupt unit 102
Bump-free transfer 73 main memory 100
Bus-oriented systems 357-8 maintenance 408, 411-17
Buses 98, 101, 125, 341, 352 medium to large computers 108,
Bytes 147 110
processor 101
C language 179, 183 software 5, 11
Calculating steps 225 tasks 264, 266
Carrier Sense Multiple Access with see also process control systems
Collision Detection technique 125 Computer-controlled continuous process
CCnT language 254, 289 58
recommendations 339 Computerization 305, 323
Central processor unit 135, 148, 166 Computers
Chronological event recorders 244, as process control parts 28
249 as process controllers 37, 78,
Closed-loop dynamics 56 98-101
Coal mine controllers 313 task stratification 79-86
Coaxial cable links 121, 123 Computing power 143, 147
Code distance 113 Concentration points 218
Collision situation 213-14 Conditional program branch 152
Colour video displays 251, 252, 253, Connections
254 connecting request list 158
Commands, process output connecting tables 205, 208
devices 73, 74-5 input devices 67
Communication networks 113, 114 lateral controller connections 215,
loop 113, 114 218
multipoint 114 loop connections 28
point-to-point 113, 114 multiple communication connection
Communications 215
design 325, 328-30, 331 Contention resolution 213-14
Subject index 423
Control centres 337, 338, 341-5 Data input process hardware 62-70
Control economics 180 see also input devices
Control equipment 11 Data logging 241, 405
Control operations 264, 265 protocols 241, 244
Control system designers 309, 383 Data maintenance 210
Control tasks 78 Data management functions 159, 163
Control terminals 113 Data output process hardware 70-3,
Controlled systems 11 74
Controllers 28, 55, 57 actuators 71-2
Conversion routines 176, 178-9 commands 73, 74-5
Cost/benefit analysis 305, 317-19 electronic control devices 72
benefits 319-23 final control devices 70-3
investment costs 317-19 Data presentation 238, 240, 242-3,
operating costs 319 244
Count pulses 61, 171 Data processing 136-47, 157
Crude-oil gathering station 310, 311 algorithms 17, 51, 59, 136
Crude-oil measurement station 79-81 basic process unit 49-62
Crude-oillnatural-gas separator 42, 44 computing power 143, 147
CSMNCD technique 125 flow chart 192, 193, 194, 195
Cycle stealing 168 functions 138-41, 335
Cycle time 143, 146 hardware 137
safety of equipment 70
schedule 198
Data, definitions 6 software 147, 149-51
Data acquisition 53, 170, 202, 218, speed 53-4
309 variables 59
speed 329, 339 Data reduction 233, 235-6
Data communication 101-21 Data retrieval 210
asynchronous 112 Data transmission 325, 328-30
channel capacity 101, 105, 110-11 code security 338
circuits 10 1 units 330-40
codes 101 Data volumes 308-11
duplex model 113 Data-oriented work 5
facilities 121 Databases 162, 202-14
half-duplex mode 113, 337 access to 209
hardware 111-21 basic process software, and 202-14
modes 112-13 data plausibility 214
open-ended 124-5, 126-8 distributed, see distributed databases
security 113 facilities 209-11, 211
simplex mode 113 failure types 215, 217
smart process devices 125, 130 knowledge database of binary expert
synchronous 111 systems 292, 293
see also communication networks: organization 211-13
communications Datawords 111
Data concentration points 218 Debugging routines 177, 178
Data descriptors 212 Demultiplexer frequency 143
Data flows 308-11 Derived data 211
Data independence 210 Device independence 343, 348-9
424 Subject index
Device usage 199 see also artificial inferencing tools

Digital devices 65 Extended real-time functions 273, 278
Digital input circuits ~5 Extended real-time processing 268
Digital input processing 139, 142 Extended real-time programs 278
Digital output circuits 96 Extended real-time software 220-6,
Digital output processing 139 268
Digital signals 16, 17, 19, 59, 120 External addressing systems 204, 205
acquisition time 53 External loss time 413
binary 18, 19 External memory 204
character set 25
count 18, 19 Failure causes 386
scanning 51 Failure discovery 322
Dispatching centres 341-5, 346-7 Failure rates 397
Distributed concentrators 218 Failure transition 215
Distributed databases 214-20 Feasibility studies 305-23
advantages of 215 areas of investigation 306
concentration points 218 basic design system 316
control system 215, 216 communication hardware and
data plausibility 214 software 316-17
failure types 215, 217 contents of report 306
lateral controller connections 215, costlbenefi t analysis 305, 317-19
218 data volumes and flows 308-11
multiple communication connection design task 305
215 stages 306, 307
Distribution networks 81-2 stratification of process control
Distribution predictivity 375, 377 311-17
Distribution subnetworks 81-2, 83, Feedback systems 37, 39
88-89 Feedforward systems 37, 39
DIT 199 File handlers 178
Double-layer printed circuit modules File systems 154
356 Filtering
Down-time 322, 397, 400, 413 filtering time 50
Drivers (software modules) 149, 150 impulse response of filter 62
DS-8 system 218, 219, 220, 221 Final control devices 70-3
Dumping 178 Firmware programs 137, 407
Duplex mode of data communication First-in-first-out principle 189, 192
113 First-in-first-out queues 221
Fischer Controls pressure gauge
Editing facilities 176 transmitter 16
Electronic control devices 72-3 Flow calculation 68
Entry points 210 Flow charts 192, 193, 194, 195, 306,
Equivalent transistor functions 352 308
Ethernet 154 Flow pulse transmitter 42
Execution permission 149 Following up operations 264, 265
Executive devices 71 Formatting routines 178-9
Expert systems 153 FORTRAN language 179-80, 183,
binary 292, 293 185
Subject index 425
Fourier transform 117 control systems, and 6, 8, 9

Frequency division multiplexers 232-41, 252-6, 264-78
115-16, 118 external group work 372-3
Frequency modulation 117, 119-20 group authority 378-81
Frequency shift keying (FSK) 120 group culture 378-9
Full order observer 54-5, 56 group dynamics 377-8
Functional systems 5-6, 87, 92, 95-7, group state 375-7
310 internal group work 373-5
level 1 87, 92 leadership 375, 378
level 2 87, 92 operator-system dialog 265, 267,
level 3 87, 92, 95 271-2, 274-5, 276-7
level 4 92, 96 physical condition 236-7, 238
level 5 92, 97 programming teams 378, 381
system assembly and 372-81
Goal address 213 training 408-10, 414
Global memory 171, 172, 173 use of control system 232, 233
Graphic display units 241, 246, 249 visual perception 233, 237
Group culture 378-9 see also process operators
Guarantee period 408 Hysteresis 68
Guaranteed response time 200, 201,
202 Idle characters 111
IF-THEN-ELSE logic judgements 289
Half-duplex mode of data Incremental regulations 273
communication 113, 337 Independent data recovery 175
Hamming distance 113 Information, definition 6
Hardware Information carriers 6
buying 352-8 Information channel 10 1, 105, 110-11
control centres 341 Information parameters 65
costs 318, 350-1 Information processing 8
data transmission units 335-6 Input devices 62-70
design 385, 387 connection 67
dispatching centres 341-2 construction data 70
functions 135-6, 404 detecting device 63
module failure rate 403 electric system data 69-70
pointers 210 functions 65, 68-9
priority levels 166 measuring transducer 62-3
software, and 151-6 mounting data 70
testing 352-8, 403, 406 sensor 63
see also process hardware signal converter 63-4
Hash addressing 212, 214 signals 66
HDLe protocol 121, 124 standard process 64-70
Historic data fields 341-2 Input rate 137
HOOD method 359 Input time 137
Human operator 6, 8, 9, Input variables 221
audio perception 237 Input/output devices 162, 165, 356
communication with control system Input/output hardware 144-5
233--41 control levels 86-7, 92-7
426 Subject index
Input/output modules, intelligent 153 Logic indentifiers 210

Input/output subsystems 336 Logic operations 151, 282
Input/output units 158 Loop connections 28
Instruction repertory 143
Instructions 152 Macroinstructions 183
Integrated circuits 97, 352, 355-6 Maintenance 411-17
Integrity control 211 MAP (Manufacturing Automation
Intelligent input/output modules 153 Protocol) 124-5, 126-8
Interactive variables 221 Master stations/terminals 113, 115,
Interference 337, 338
interference connections 392-3 Mathematical modelling 39-40
power supply interferences 394 Mean time between failures 397, 402
voltages 392, 394 Mean time to repair 400, 402
Internal addressing system 204 Measuring transducer 62-3
Internal variables 194, 199 Memory
Interrupt processing 166 external 204
Interrupt response 163 global memory 171, 172, 173
Interrupted programs 158-9 microcomputer-based systems 151
Interrupts 166 random access memory (RAM) 137,
Inverted indexes 210 147, 173, 174
Investment costs 318, 350 read-write 151
ROM, see read-only memory (ROM)
Job management functions 159, 161 Memory cycle 143
Memory usage 199
key tree 212 Microcomputer-based systems 98, 106,
keyboard control system 186 287
computing power 147, 148
Labour force 13 memory 151
see also human operator: process operating system 166, 168, 173
operators: teamwork overhead characteristics 173
Laplace transforms 17, 23 software 368
Last-in-first-out principle 192 system parameters 173-5
Lateral controller connections 215, Microprocessors
218 cycle time 143, 146
Leadership 375, 378 speed of components 146
Level transmitters 42 Mimic diagrams 188, 250, 251, 341
Limit values 68 Mimic elements 268, 273, 278, 280
Line printers 241 Minicomputer systems 109
Linear 28, 32-3 computing power 147, 148
Linear systems 28, 34 operating systems 168, 178
causal 28 Mistakes 413
classification 33-5 MML (man-machine language) 254
dynamic 28 Model-following method 58
structuring 35-7 Modems 117, 121, 124, 328, 336, 339
time response 32-3 Modular systems 356, 357
time-instant 28, 29-32 Multiple communication
Linking facilities 179 connection 215
Load centre voltage 83 Multiplexer circuits 50, 115-16
Subject index 427
frequency division multiplexers Parameter tables 204

115-16, 118 Pascal language 179, 183, 254
time-division multiplexers 116, 118 Permission for program execution 149
Multiplexer frequency 143 Phase modulation 118, 119
Multiplexer scanning time 50-1 PL/l language 179
Multiprogramming 199-202 Planar transistors 97
Planning operations 264, 265
Natural-gas field station 313 Plausibility 214
Network file system 154 Positioning systems 184, 185
Noise 27, 105, 113, 392 Power supply interferences 394
Nonresource segments 200 Power transistor 72
Primary bucket 214
open-ended control systems 351 Printers 241, 247, 341
open-ended data communication Privacy control 211
124-5, 126-8 Probability density functions 413
MAP 126-8 Problem-oriented work 5
OSI model 125, 126-8 Process commands 268-9, 270, 279,
Operating systems 157-82 309
data 173 Process complexity 14
functions 162-73 Process control
hierarchical 175, 176 actions 258, 259
interrupt response 163 algorithms 257
mean response time 164 auxiliary data 73, 76-7
overhead 202 computer hardware 98-101
software 175-80 computer systems 97-8, 99
subsystems 157 data processing 147, 149-51
symbols 173-5 functions 268
time critical 161 strategic decisions 341
time uncritical 161 Process control systems
UNIX 152, 153, 404 automation 257
see also real-time operating systems basic concept 37-40
Optical signalization 268, 271-2, benefits 319-23
274-5 categories 15-16
OSI (Open System Inter-connection) closed-loop 37
model 125, 126-8 commissioning 404-8
Output control levels 86, 87, 92, 95-7
digital output circuits 96 costs 8, 15, 308, 317-19, 387
digital output processing 139 database information 208
rate 137 decision criteria 385, 387
time 137 design 3, 305, 316, 324-81, 387
see also data output process development 8, 10, 11, 13
hardware functions 4, 8, 265, 268
Overflow bucket 214 hardware 352-8
Overhead characteristics 173 maintenance 408, 411-17
operating system overhead 202 microcomputer-based 106
models 222, 223
Packet switching network 121, 123 modules 352, 354-8, 390
Parallel processor architecture 153 open-loop 37
428 Subject index
protection of 393-4 visual perception 233, 237

reliability 396-404, 402 see also human operator
software 135-6, 150, 175-80, Process parameters 223
358-68 Process planning 268
stratification 79-86, 308-9, 311-17 Process protecti.n systems 278-86,
supervision 406 393-4
systems approach 4-16 active 280, 282
testing 396-404 design criteria 282
see also basic process unit: computer dynamic 279, 280, 281
control systems passive 259, 280, 282
Process controllers 42, 43, 45, 49 static 279, 280, 281, 285
control algorithms 90 Process sensors 63, 389, 391
microcomputer-based 106, 109 Process signalizations 280
microprocess-based 109 Process signals 16-26, 65
Process data 204-8, 211, 252 distribution 68-9, 69
distribution 210, 309 Process simulation 226, 224
visualization 249, 250, 268, 395 Process software 149-51
Process designers 309 Process states 171, 236, 249, 257-64
Process devices 385 allowable 259
response times 51, 53, 54 available 270
Process events 61, 149, 268, 298 change of 259
Process hardware dangerous 259, 261
for data input 62-70 recognition of 254
for data output 70-3, 74 safe 259, 261
Process information 73, 76-7, 211 stable 279
Process modelling 4, 39, 220-6, supervision 265
262-3 total number 248, 264
Process operators transitions 270, 272
artificial inferencing tools 289-97 unallowable 259
audio perception 237 unstable 279
communication with control system unavailable 270, 273
233-41 useful information on 244, 248
control of work 298-9 working state 257, 259
formalization of role 289, 290 Process structure 220
function 231 Process transmitters 50
hardware functions 234, 235 Process varilbles 4-5, 11, 39, 257,
interaction with control system 280-2
252-6, 264-78, 287-30 calculation of 68
limits of functions of 299-300, 301 control of 11
long term memory 235, 237 detection 259
operator-system dialog 265, 267 model 222, 223
physical condition 236-7, 238 ranges of conditions 284
reaction times 234, 236, 298 Process working mode 220
reaction to process control systems Processes 313-5
299,300 definition 8
retention of knowledge 254, 255 description 8, II
role 287-301 energetic aspects 11, 12, 13
short-term memory 234-5, 237 environmental aspects 12, 13
skill 287 features 13
Subject index 429
human aspects 11 , 12, 13 Read-only memory (ROM) 137, 147,

informational aspects 12, 13 151, 407
materials/raw materials aspect 11, erasable programmable (EPROM)
12, 13 137
socioeconomic aspects 12 programmable (PROM) 137, 174
time-space aspects 12, 13 Read-write memory 151
types of 8, 11-16, 313-15 Ready state 171
PROCONTROL P12 coaxial cable link Real-time operating systems 149,
121, 123 158-73
Production plant, definition 4, 5 basic concepts 158-62
Production process functions 162-73
definition 4 hierarchical 175, 176
environmental aspects 15 structure 162-73
Production system see also operating systems
design study 5 Reconstructed state 54
development 5 Recursive filtering 62
feasibility study 4, 5 Refinery plant 313
preliminary study 4 Regulations 273
Productivity 305 Regulators 58
PROFIBUS DIN Vl9245 130 Remote file system 154
Program branching 152 Reporting 268
Program chains 194, 196-7 Resource blockage 200-1
Program communication links 175, Response time, guaranteed 200, 201,
177 202
Program counter 152 Rosemount Model 3051 smart
Program efficiency 147 transmitter 120
Program exchange 147
Program request routine 158 Safety 259
Program starting routine 158 commissioning and 406-8
Programmed tasks 158 data processing equipment 70
Programming languages 179, 183, 185 Sampled devices 65
see also under names of languages Sampling frequency 143
Programming systems 162 Scanning of signals 21
Programming teams 378, 381 SDLC protocol 121, 124
Protocols 241, 244, 406 Search sequence 295
PROZA 11 DIR software system 180, Semaphores 194, 199
181, 182, 185, 186, 211, 212 Semiconductor-controlled rectifiers 72
Semigraphic display units 241, 246,
Queue-oriented systems 218, 220 249
Queues 158, 166, 192, 218 Seminars 409-10
first-in-first-out 221 Sensors 63
Servomotor actuators 72
Radio link 321-5, 325, 328-30 Set-point regulations 273
Railway station control 315 Set/reset sequential functions 42, 45,
Random access memory (RAM) 137, 48
147, 173, 174 Shannon 26
Reaction times Signal converters 63-4, 65
operator 234, 236, 270, 298 Signal-to-noise ratio 50, 329, 330,
systems 269 331
430 Subject index
Signals 16-27 priority levels 166, 167

analog see analog signals process 149-51, 175-80
characters 25 programs 137
communication of 27 requirement specification 358
definition 6 testing 367, 368-72, 402, 407
description 21-3 Software team structuring 381, 382
destination 16, 17 Solenoid actuators 72
digital, see digital signals Speech transmission 325, 328
direction 16 Speed variators 72
distortion 329 Square pulse signals 18, 19
information rate and content 23, Squelch circuits 121
24-7 Standardization of equipment 408, 411
mathematical descriptions 23 State variables 39, 221-2
sampled 21 Stationary sinusoidal signals 19
scanning of 21 Status data 171
singular 18, 20 Status input processing 156
source 16, 17 Status signals 260, 264
square pulse 18, 19 Stiffness 40
stationary sinusoidal 19 Stop combination 111-12
time-discrete 21, 22 Street traffic control systems 45, 47,
types 17-21 48-9
unit pulse and unit step 19 Structured programming 404
Simplex mode of data communication Subsystems 28
113 Suspended state 171
Single supply operation 73 Synergism, hardware and
Singular signals 18, 20 software 154
Slave terminals 113, 115 Synoptics blocks system 45, 46
Smart transmitters 27, 120, 125, 130 System assembly 372-81
PROFIBUS DIN V19245 130 external group work 372-3
SMD (Surface Mounted Device) 356 group authority 378-81
Software group dynamics 377-8
algorithms 188, 189 group state 375-7
basic process 202-14, 203 internal group work 373-5
block model software 5 System commissioning 404-8
control centres 341 guarantee period 408
costs 318, 350-1 preparatory work 405-6
designing 358-68 safety and 406-8
dispatching centres 341 system design integrated approach
engineering calculations, for 151 385-418
extended real-time 151, 220-6 commissioning and safety 406-8
faults 402, 404 maintenance 408, 411-17
functions 135-6, 358 mounting and installation procedure
hardware, and 151-6, 358 387-96
intercomputer communication 149 system commissioning 404-8
knowledge engineering and expert testing and reliability 396-404
systems 153 training of personnel 408, 409, 410
message-switching 149, 151 System elements 8
microcomputers 368 System failures 397
operator-system communication 299 cumulative failure distribution 397
Subject index 431
mean time between failures 397, Time-function diagrams 21, 23

402 Time-programmed functions 149
mean time to repair 400, 402 Timeout transition 215, 217
System faults 396, 398-9 TLM (Siemens) 184
System integration 352, 354 Trace functions 178
System management functions 159, Traffic control systems 45, 47, 48-9
162 Training of personnel 408, 409, 410,
System parameters 173-5 414
System refreshing intervals 268 Transducers 65, 66, 78
System report programs 178 measuring transducer 62-3
System requirement specifications 183 Transformer tap movement 84, 86
System tables 175, 204 TRAP instruction 166, 178
System time response 199 TRAP processing 166, 168
Systems 28-37 Triacs 72
automata, and 37 Turnkey controls 180
causal 33
definition 28 Unit functions 32, 33
dynamic 34 Unit pulse signals 19
instant response 33 Unit step excitation 33
linear, see linear systems Unit step signals 19
nonlinear 28 Unit tag 159
structuring of 35-7 Universal logic systems 151
time-continuous 28, 32-3, 34-5 Universal modules 352
time-discrete 28, 29-32 UNIX operating system 153, 154,
404
Taking-over certificate 408 User's data 211
Taking-over tests 406-8 User's programs 162, 165, 175, 337
Tasks Utility routines 178-9
application program tasks 183-9
stratification 79-86 Valve positioners 72
task chains 200-2 Valves 72
task priority 172 Video displays, colour 251, 252,
task states 171, 173 253, 254
Teamwork 8, 377-81, 387, 388 Visualization 241, 244-52
Technical documentation 325, 326-7 display units 245, 246, 251
Technology actions 259, 261 hardware equipment 241, 244
Telephone network 124 protocols 241, 246
Testing Voltage ratio transformer 84, 85
acceptance level 370-2
hardware 352-8
program and system level 370 Waiting state 171
program unit level 368 Word length 143
routine level 370 Working body system 278
software 367, 368-72
taking-over tests 406-8 X.25 154
Thermal power unit 84, 86, 90, 91 X-Look 154
Thermoelectric power plant 313 X-Windows 154
Thyristors 72
Time-division multiplexers 116, 118 Z-transforms 17, 23, 62

Jovic 1992 Process Control Systems Principles of Design Operation and Interfacing

Cargado por

Información del documento

Derechos de autor

Formatos disponibles

Compartir este documento

Compartir o incrustar documentos

Opciones para compartir

¿Le pareció útil este documento?

¿Este contenido es inapropiado?

Copyright:

Formatos disponibles

Jovic 1992 Process Control Systems Principles of Design Operation and Interfacing

Cargado por

Copyright:

Formatos disponibles

Process Control

To the memory of the distinguished men and teachers

English Language Edition Consultant:

Ignl SPRINGER-SCIENCE+BUSINESS MEDIA, BV

ISBN 978-94-010-5380-8 ISBN 978-94-011-3116-2 (eBook)

Part One Systems, Processes and the Role of Process Control

1 Signals, systems and process control

2 The basic process unit 41

3 Stratification of control tasks and data communication 78

3.9 Practical suggestions and recommendations 121

Part Two The Role of Software in Process Control Systems 133

4 The relative roles of software and hardware 135

5 System software 157

6 Application programs and databases 183

Part Three The Man-Machine Interface 229

7 Reduction and visualization of data and procedures 231

8 Process management and control 257

9 The role of the operator in process control systems 287

Part Four System Design 303

10 The feasibility study 305

11 Computer control system design 324

12 Cost-effective system selection 350

12.1 Introduction 350

13 The integrated approach 385

Author index 419

Subject index 421

free-standing or networked throughout the plant. Additionally,

The following tables and figures have been reproduced here

A process control system is made up of a process involved in a controlled

1.2 A SYSTEM APPROACH TO PROCESS CONTROL SYSTEMS

production units controlled in an appropriate way constitutes a production

1.2.2 THE COMPUTER CONTROL SYSTEM

A typical computer control system for a power plant is illustrated in Figure

working change in production

Figure 1.1 Development cycle of a control system: - - - -, control system

means of the collection, storage, processing, dissemination, interpretation and

analog digital drive controls ':;:'

Figure 1.2 A typical computer control system of a power plant.

Due to specific limitations on speed and operable information capacity when

1.2.3 TYPES OF PROCESS

A process is defined as a set of operations that perform physical or chemical

process and interface

I 59% 41% I total efl

Figure 1.5 Production environment aspects.

Human Variable response Handtool Increasing

investment production material

The influence of a process element, a system part or a system upon other

00- signal source

1.3.2 TYPES OF SIGNAL

B . time discrete analog

C • binary digital signal

r' D • count digital signal

L-______~'____________~______~______~ ______~L__ _ _ _~t

count signal instant:

Figure 1.9 Singular signals.

L3 --- ---------------------------~=== ~ output

L2 -- ------- ----------------- - - -----

LI--- ----------------------------/--===~ forced output

Figure 1.10 Signals in a process liquid tank_

1.3.3 TIME-DISCRETE SIGNALS

Any analog or digital signal can be observed using a scanning procedure,

u(z) =L u(n)z-O (1.4)

1.3.4 A DESCRIPTION OF SIGNALS

Signals usually represent complex interactions between process parts or complex

L-~'__~~~ ~L _ _ _~t