Book of Design Water System

Download as pdf or txt
Download as pdf or txt
You are on page 1of 233

WATER AND WASTEWATER SYSTEMS ANALYSIS

DEVELOPMENTSIN WATER SCIENCE, 34

OTHER TITLES IN THlS SERIES (Volumes 1-3 are out of print)


4 J.J. FRIED
GROUNDWATER POLLUTION
5 N. RAJARATNAM
TURBULENT JETS
6 D. STEPHENSON
PIPELINE DESIGN FOR WATER ENGINEERS
7 V. HALEK AND J. SVEC
GROUNDWATER HYDRAULICS
B J.BALEK
HYDROLOGY AND WATER RESOURCES IN TROPICAL AFRICA
9 T.A. M c M A H O N AND R.G. MElN
RESERVOIR CAPACITY AND YIELD
10 0. KOVACS
SEEPAGE HYDRAULICS
11 W.H. GRAF AND C.H. MORTIMER (EDITORS)
HYDRODYNAMICS OF LAKES: PROCEEDINGS OF A SYMPOSIUM
12-13 OCTOBER 1978, LAUSANNE, SWITZERLAND
12 W. BACK A N D D.A. STEPHENSON (EDITORS)
CONTEMPORARY HYDROGEOLOGY: THE GEORGE BURKE MAXEY MEMORIAL VOLUME
13 M.A. M A R l f l 0 A N D J . N . LUTHIN
SEEPAGE AND GROUNDWATER
14 D. STEPHENSON
STORMWATER HYDROLOGY AND DRAINAGE
15 D. STEPHENSON
PIPELINE DESIGN FOR WATER ENGINEERS
(completely revised edition of Vol. 6 in the series)
16 W. BACK A N D R. LkTOLLE (EDITORS]
SYMPOSIUM ON GEOCHEMISTRY OF GROUNDWATER
17 A.H. EL-SHAARAWI (EDITOR) IN COLLABORATION W I T H S.R. ESTERBY
TIME SERIES METHODS IN HYDROSCIENCES
18 J.BALEK
HYDROLOGY AND WATER RESOURCES IN TROPICAL REGIONS
19 D. STEPHENSON
PIPEFLOW ANALYSIS
20 I. ZAVOIANU
MORPHOMETRY OF DRAINAGE BASINS
21 M.M.A. SHAHIN
HYDROLOGY OF THE NILE BASIN
22 H.C. RlGGS
STREAMFLOW CHARACTERISTICS
23 M. NEGULESCU
MUNICIPAL WASTEWATER TREATMENT
24 L.G. EVERETT
GROUNDWATER MONITORING HANDBOOK FOR COAL AND OIL SHALE DEVELOPMENT
25 W. KINZELBACH
GROUNDWATER MODELLING: AN INTRODUCTION WITH SAMPLE PROGRAMS IN BASIC
26 D. STEPHENSON AND M.E. MEADOWS
KINEMATIC HYDROLOGY AND MODELLING
27 A.H. EL-SHAARAWI A N D R.E. KWIATKOWSKI IEDITORS)
STATISTICAL ASPECTS OF WATER QUALITY MONITORING - PROCEEDINGS OF THE
WORKSHOP HELD AT THE CANADIAN CENTRE FOR ISLAND WATERS, OCTOBER 1985
28 M.JERMAR
WATER RESOURCES AND WATER MANAGEMENT
29 G.W. ANNANDALE
RESERVOIR SEDIMENTATION
30 D.CLARKE
MICROCOMPUTER PROGRAMS FOR GROUNDWATER
31 R.H. FRENCH
HYDRAULIC PROCESSES ON ALLUVIAL FANS
32 L. VOTRUBA. Z.KOS. K. NACHAZEL. A. PATERA ANDV. ZEMAN
ANALYSIS OF WATER RESOURCE SYSTEMS
33 L. VOTRUBA AND V. BROZA
WATER MANAGEMENT IN RESERVOIRS
DAVD STEPHENSON
Water Systems Research Group, University of the Witwatersrand,
1 Jan Smuts Avenue, Johannesburg, South Africa

ELSEVl ER

Amsterdam - Oxford - New York - Tokyo 1988


ELSEVIER SCIENCE PUBLISHERS B.V.
Sara Burgerhartstraat 25
P.O. Box 21 1, 1000 AE Amsterdam, The Netherlands

Distributors for the United States and Canada:

ELSEVIER SCIENCE PUBLISHING COMPANY INC.


52, Vanderbilt Avenue
New York, NY 10017, U S A .

ISBN 0-444-42945-X (Vol. 34)


ISBN 0-444-41669-2 (Series)

0 Elsevier Science Publishers B.V., 1988

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system
or transmitted in any form or by any means, electronic, mechanical, photocopying, recording
or otherwise, without the prior written permission of the publisher, Elsevier Science Publish-
ers B.V./Physical Sciences & Engineering Division, P.O. Box 330, 1000 AH Amsterdam, The
Netherlands.

Special regulations for readers in the USA - This publication has been registered with the
Copyright Clearance Center Inc. (CCC), Salem, Massachusetts. Information can be obtained
from the CCC about conditions under which photocopies of parts of this publication may be
made in the USA. All other copyright questions, including photocopying outside of the USA,
should be referred to the publisher.

No responsibility is assumed by the publisher for any injury and/or damage to persons or
property as a matter of products liability, negligence or otherwise, or from any use or opera-
tion of any methods, products, instructions or ideas contained in the material herein.

Printed in The Netherlands


V

PREFACE

A systematic approach to decision m a k i n g i n water resources p l a n n i n g


i s presented w i t h p a r t i c u l a r reference t o wastewater re-use.
Various methods of system s i m u l a t i o n a n d o p t i m i z a t i o n a r e a p p l i e d in a
number of case studies. Methods of analysis and numerical methods
(Chapter 2, 4 ) a r e described as well a s the b a s i s of p o l l u t i o n a n d water
quality (Chapter 1 , 3). The economics of desalination a r e a l s o discussed
(Chapter 7 ) .
The a u t h o r has considerable experience in p l a n n i n g water purification
and recycling systems i n an a r i d area, Southern Africa. Water i s at a
premium f o r m i n i n g a n d i n d u s t r i a l development a n d considerable money is
spent on water treatment o r use of poor q u a l i t y water. Careful management
a n d d i s t r i b u t i o n of water resources can i n these circumstances save a lot
of money. The general theory of o p t i m i z a t i o n subject to q u a l i t y c o n s t r a i n t s
i s presented i n Chapter 6.
The examples studied range from regional supplies (Chapter 10) to
internal re-circulation (Chapter 8). Groundwater and artificial recharge
are considered (Chapter 9) and stormwater quality (Chapter 5) and
sewerage systems (Chapter 1 1 , 12) a r e also covered. Computer applications
e x i s t throughout and a number of s i m u l a t i o n a n d o p t i m i z a t i o n programs in
BASIC a r e presented.
Chapter 13 i s on an often ignored subject, the necessity f o r s c i e n t i f i c
sampling procedures in monitoring water quality. It was written by
Professor Tom Sanders of Colorado State U n i v e r s i t y .
The theory and case studies should p r o v e of v a l u e in many aspects of
planning use of water resources with quality constraints. Wastewater
re-use and conservation therefore a r e promoted b y the approach adopted.
vi

CONTENTS

CHAPTER 1 . WATER QUALITY IN INDUSTRIAL SYSTEMS

Geochemical source o f p o l l u t i o n 1
Effect o f e v a p o r a t i o n on c o n c e n t r a t i o n s 2
Effects o f poor q u a l i t y w a t e r 2
Scaling 3
P r e d i c t i o n of s c a l i n g a n d corrosion 3
P r e v e n t i o n of s c a l i n g 3
Calcium c a r b o n a t e s c a l i n g 5
Sulphate s c a l i n g 6
A d d i t i v e s f o r t h e p r e v e n t i o n of s c a l i n g 6
Fouling 8
Control o f f o u l i n g 9
O i l emulsion b r e a k d o w n 10
Corrosion 10
Types of corrosion 13
Corrosion p r e v e n t ion 14
P o t a b l e water s t a n d a r d s 15
Agriculture and irrigation 17

CHAPTER 2. MATHEMAT I CAL MODELLING O F WATER QUAL I TY

Mass Balances 20
M i x e d a n d p l u g f l o w systems 21
Systems a n a l y s i s 24
T e r m i n a l concentration in a w a t e r c i r c u i t 24
A p p l i c a t i o n to a m i n e w a t e r c i r c u i t 26
Computer s i m u l a t i o n model 31
Mathematical b a s i s of model 31

CHAPTER 3. NON CONSERVATIVE PARAMETERS

Introduction 35
B a s i c mass b a l a n c e e q u a t i o n 35
Oxygen b a l a n c e in r i v e r s 37
Coupled equations f o r DO a n d BOD 37
Analytical solution 39
C a l i b r a t i o n of a m o v i n g BOD model 40
Oxygen b a l a n c e 40
Fie1d measurements 45

CHAPTER 4. NUMERICAL METHODS

S i m u l a t i o n o f H y d r a u l i c Systems 51
Two-step method 52
Demonstration o f n u m e r i c a l i n a c c u r a c y 52
I m p l i c i t f i n i t e d i f f e r e n c e schemes 55
Comments on f i n i t e d i f f e r e n c e methods 56
Numerical methods f o r t h e s o l u t i o n o f s i n g l e
differential equations 57
The E u l e r method 57
The m o d i f i e d E u l e r method 59
Runge-Kutta methods 60
M u l t i s t e p methods 61
F i n i t e elements 62
Boundaries f o r n u m e r i c a l methods 62
vi i

CHAPTER 5. MASS BALANCE O F STORMWATER POLLUTANTS

Introduction 64
Catchmen t d e s c r i p t i o n 64
Q u a l i t y Observations 66
Fa1lout measurement 66
Relationship between t o t a l p o l l u t a n t load a n d
r u n o f f volume 67
Chemical constituents 67
Mass b a l a n c e f o r event of 18 January 1985
on H i l l b r o w catchment 72
Mass b a l a n c e f o r event of 7 M a r c h 1983
on Montgomery P a r k catchment 73
Conclusions 77

CHAPTER 6. OPTIMUM ALLOCATION O F WATER RESOURCES SUBJECT


TO QUAL I T Y CONSTRA INTS

Int roduc t ion 79


The system 80
Solution method 82
Discussion 85
L i n e a r Programming Solution 85
The I i n e a r programming technique w i t h separable
programming a p p l i e d 91
S e n s i t i v i t y study f o r v a r i o u s acceptable TDS values 95

CHAPTER 7. ECONOM I CS OF DESALINATION OF WASTEWATERS

I n t roduc t ion 99
A l t e r n a t i v e s f o r optimal reuse of waste water 99
Selection of optimum d e s a l i n a t i o n methods 101
Relevant d e s l i n a t i o n methods 103
I n d u s t r i a l wastewater treatment 1 04
Reverse osmosis 104
Membrane d e s c r i p t i o n 105
EIect r o d i a I y s i s 105
Ion exchange 105
Cost a n a l y s i s 107
C a p i t a l costs 107
I n d i r e c t c a p i t a l costs 108
Running costs 108
L a b o u r costs 108
Membrane replacement 108
Conc Ius ions 111

CHAPTER 8. COMPUTER ANALYSIS JUST I F IES DESAL I NAT ION

I n t roduct ion 115


A p p l i c a t i o n of o p t i m i z a t i o n of water s u p p l y 116
Systems Analysis 118
General o p t i m i z a t i o n problem 121
Program a p p l i c a t i o n 122
Optimization of mine water system 123
Result o f a n a l y s i s 123
Appendix 8.1 128
MlNSlM Program f o r s i m u l a t i n g flow a n d TDS 128
in closed systems. 128
Tape o r disc management 128
MlNSlM l i s t of symbols 129
viii

A p p e n d i x 8.2 136
MINOP p r o g r a m f o r o p t i m i z i n g d i s t r i b u t i o n 136
MINOP l i s t o f symbols 136

CHAPTER 9. INTEGER PROGRAMMING PLANNING OF TREATED WASTEWATER


CONVEYANCE FOR A R T I F I C I A L RECHARGE O F AN AQUIFER

Introduction 141
Cost a n a l y s i s 146
Mathematical formu tat ion 149
Results 151
Summary a n d conclusions 153

CHAPTER 10. OPTIMAL PLANNING OF REGIONAL WASTEWATER TREATMENT

Introduction 155
T h e m a t h e m a t i c a l model 158
Optimization method 162

CHAPTER 1 1 . SIMULATION OF SEWER FLOW

Int r o d u ct i o n 166
Hydraulic analysis 167
F low measurements 167
H i g h e r income r e s i d e n t i a l 169
L o w income r e s i d e n t i a I 170
Apartment b u i l d i n g s 171
Commercial a r e a s 171
Industrial 172
Conclusions 172
Appendix 174
P r o g r a m SEWSIM 174
Effect of local p e a k s 1 74
Routing effect 175
Non-Circul a r Conduits 175
I nf low components 1 76
Data 177
Program output 177
Sample d a t a f i l e 186

CHAPTER 12. SEWERAGE SYSTEMS MANAGEMENT

L e a r n i n g Simulation Program 190


Optimization 192
Optimal Control a s a L i n e a r Programming Problem 193
Sewer M a i n t e n a n c e D a t a P r o c e s s i n g in J o h a n n e s b u r g 195
A p p l i c a t i o n t o J o h a n n e s b u r g ' s System 197
P r o c e s s i n g of Sewer M a i n t e n a n c e D a t a 198

CHAPTER 13. WATER QUAL I TY MON I TORlNG NETWORKS

Necessity f o r Networks 204


M o n i t o r i n g System F r a m e w o r k 205
F a c t o r s in N e t w o r k D e s i g n 205
S e l e c t i o n of Water Q u a l i t y V a r i a b l e s t o M e a s u r e 206
Sampling Station Location 207
Sampling Frequency 21 1
Discussion 215
ix

AUTHOR INDEX 21 7

SUBJECT INDEX 21 9
This Page Intentionally Left Blank
1

CHAPTER 1

WATER QUALITY IN INDUSTRIAL SYSTEMS

GEOCHEMICAL SOURCE OF POLLUTANTS

Many of the chemicals ?n solution in water originate from the


surroundings. M i n e r a l s which form rocks may be dissolved b y water in a
s u i t a b l e environment. Acidic waters i n particular a r e known to dissolve
certain chemicals in the rock, Exposure to a i r , which contains oxygen,
assists the reaction, Iron sulphide i s one such chemical which can be
o x i d i z e d to sulphate. B a c t e r i a a r e also thought to p l a y a n important p a r t
i n the leaching of sulphides.
The s o l u b i l i t y of chemicals i s also dependent on temperature and the
t o t a l dissolved s o l i d s i n the water amongst o t h e r factors. I n many cases
the r a t e of dissolution i s slow. I t may take years to dissolve all the
s u l p h i d e from a rock sample. On the other h a n d c h l o r i d e s dissolve very
r a p i dIy .
When a chemical compound dissolves in water the ions appear as
positively charged metal or cations and negatively charged anions.
Solubility depends on other charges present and is expressed as a
s o l u b i l i t y K. In p a r t i c u l a r water i s ionized a s follows: (Brownlow, 1979).
H 20 = H+ + OH-
The log of the hydrogen H+ ion concentration i s termed the pH:
pH = -log(H+)
Water w i t h a pH below 7 i s acidic. I t may be rendered so b y many
factors. For example absorption of carbon d i o x i d e C 0 2 from the a i r forms
carbonic a c i d which c o u l d reduce the pH as l o w as 3.0. O n the other hand
water d r a i n i n g from limestone o r s i l i c a t e m i n e r a l s may h a v e a pH g r e a t e r
than 7 ( P e l l e t i e r , 1964).
The process of l e a c h i n g s u l p h i d e from minerals i s self stimulating as
s u l p h u r i c a c i d i s formed i n the process. On exposure of s u l p h i d e b e a r i n g
horizons ferrous s a l t s o x i d i z e to the f e r r i c s t a t e a n d s u l p h i d e i s o x i d i z e d
to sulphate:
4FeS2 + 1502 + 10H20 + 4FeO(OH) +' 8H2S04
I n m i n i n g environments, bacteria thrive in the acid mine water and
further promote the oxidation of Fe and S. The bacteria thiobacillus
ferro-oxidans oxidize both Fe and S whereas thiobacillus thio-oxidans
oxidizes only S (Mrost a n d L l o y d , 198Oj.
EFFECT OF EVAPORATION ON CONCENTRAT IONS

The r a t e of concentration of t o t a l d i s s o l v e d s a l t s b y e v a p o r a t i o n may be


p r e d i c t e d f o r any ambient c o n d i t i o n s u s i n g psychrometric relationships. In
a d d i t i o n to e v a p o r a t i o n in cooling towers, e v a p o r a t i o n of water takes p l a c e
in industrial systems and particularly ventilated systems. Thus if the
dry-bulb temperature of air i s 31OC a n d the air i s conveyed in at 38%
relative humidity, it will increase in water content by 5 g water
v a p o u r / k g of a i r . For a n a i r d e n s i t y of 1.2 kg/m3 and a ventilation rate
of 700 m’/s, the amount of water absorbed b y the a i r w i l l be 5 l i t r e s p e r
second ( B a r e n b r u g , 1965).
The loss of water b y e v a p o r a t i o n leaved b e h i n d s a l t s which may have
entered the system in dilute solution initially. There is therefore a
c o n c e n t r a t i n g effect, and the rate of increase in concentration in time
depends on the volume of storage in the system. Thus if there is a 12
h o u r retention a n d the flow r a t e of service water i s 100 l i t r e s p e r second,
the volume i n the system w i l l be 100 x 3600 x 24/1000 = 8640m’. If the
e v a p o r a t i o n loss i s 10 l i t r e s a second which i s 864 m 3 / d a y the i n i t i a l rate
of concentration w i l l be 10 percent of the initial concentration per day.
The s a l t concentration would increase indefinitely unless the water was
replaced. The r a t e of concentration i s u s u a l l y offset b y the f a c t t h a t there
i s a source of p u r e r water used f o r make-up b u t even this adds to the
t o t a l load unless there i s a blowdown (Porges, 1971).
The concentration of total dissolved solids at equilibrium will be a
function of the e v a p o r a t i o n r a t e as well as the p u m p i n g d i s c h a r g e rate
a n d r a t e a t which s a l t s a r e introduced as a r e s u l t of make-up water a n d
l e a c h i n g of chemicals from the environment (Van Staden, 1970).

EFFECTS OF POOR QUAL I TY WATER

H i g h t o t a l dissolved s a l t s concentration in water g i v e s r i s e to a number


of problems. The nature of the problems varies, but in all cases the
economic consequences o f poor quality water are severe. High chloride
concentration in mine water is suspected to be one of the causes of
corrosion of pipework and equipment. Sulphates and carbonates in the
waters give rise to scaling and blockages. Scaling is common in heat
exchange equipment. I n many systems there may be s c a l i n g in some a r e a s
a n d corrosion in others. Plant has frequently to be r e p l a c e d a f t e r o n l y a
few years i n service i n many systems because of these effects. The recent
development of mechanized equipment has g i v e n rise to new f e a r s of the
3

consequences of poor q u a l i t y water. Many of these machines a r e designed


to operate hydraulically using oil-in-water emulsions. Emulsion stability
and the hydraulic circuits could be already affected by poor quality
water.

SCAL I NG

S c a l i n g i s the phenomenon of chemical deposition on submerged surfaces.


The deposits are due to crystallization or precipitation. Scaling takes
place because of the dissolved salt concentration exceeds its saturation
l i m i t and i s usually a r e s u l t of an excess of chemicals in solution which
c o u l d be caused b y e v a p o r a t i o n loss of water, l e a c h i n g of chemicals from
surroundings, o r a change i n temperature.
The s c a l i n g i s also a f u n c t i o n of other parameters such as pH, total
dissolved solids concentration, alkalinity, time a n d flow v e l o c i t y .
The chemicals most frequently causing scale are calcium carbonate
(CaCO ) and calcium s u l p h a t e (CaS04). Calcium carbonate is particularly
3
insoluble whilst solubility of both salts is highly dependent on
temperature. Figure 1 . 1 illustrates the effects of temperature on the
solubility of these salts. The solubility is influenced by chlorides and
other ions i n solution. Other chemicals, particularly oxides of magnesium
(e.g. Mg(OH)2), iron, aluminium and silica a r e also sometimes found in
scales (Betz, 1980).

P r e d i c t i o n of S c a l i n g and Corrosion

T h e factors affecting the e q u i l i b r i u m of calcium carbonate i n solution


have a complex interdependence. L a n g e l i e r (1954) developed an e q u a t i o n to
predict the tendency of calcium carbonate to form a scale. Ryzner (1944)
p r e f e r r e d to express the equation in terms of pH. However there a n y many
i n f l u e n c i n g effects a n d such formulae can o n l y o f f e r a g u i d e to the l i k e l y
b e h a v i o u r of the water, p a r t i c u l a r l y i n respect to corrosion.

Prevention of S c a l i n g

One way of p r e v e n t i n g s c a l i n g o r corrosion would be to d e s a l i n a t e the


water. Possible methods of desalinating mine water are ion exchange,
reverse osmosis, e l e c t r o d i a l y s i s a n d thermal procedures. Although these a r e
expensive t h e i r possibi I i ties a r e b e i n g re-assessed.
4

Scale prevention is currently normally undertaken by pH adjustment


and, where necessary, a controlled bleed (waste) from the system to
p r e v e n t excessive concentration of the d i s s o l v e d s a l t s . T h i s treatment may
be supplemented with the use of scale and corrosion preventative
additives.

3000

2800

2600

2400

2200

zoo0

1800

1600

-E
-I
1400

1200

1000

800

600

400

200

100

50

0
0 20 40 60 80 100 120 140 160
TEMPERATURE O C

Fig. 1.1 Effect of temperature on s o l u b i l i t y of s c a l i n g s a l t s


5

Calcium carbonate s c a l i n g

The f a c t o r s a f f e c t i n g the e q u i l i b r i u m of calcium carbonate in s o l u t i o n


have a complex interdependence. In a d d i t i o n to temperature, the presence

Procedure :
glven temp. OC
TDS mgll
Ca mgll
Alkallnlty
proceed
1-2-3-4-5

PHS
Ryzner Stablllty Index RSI=lpH,-pH
Calclum Carbonate S a l l n g Ilkely If LSI>O
and R S l c 6
Colrorlon Ilkely If RSI>O

Fig. 1.2 L a n g e l i e r S a t u r a t i o n Index Chart f o r Carbonate S c a l i n g


6

of other dissolved solids, especially total alkalinity a n d pH, affect the


tendency to form scale. A sudden reduction in pressure such as at a
nozzle can induce s c a l i n g , and suspended m a t t e r i n the water may serve
as n u c l e i f o r scale formation. L a n g e l i e r developed an equation to predict
the tendency of calcium c a r b o n a t e to from scale. The L a n g e l i e r S a t u r a t i o n
Index i s LSI = pH - pHs, where pHs i s the s a t u r a t i o n pH. I f the LSI is
positive, there is a tendency to scale and if it is negative, calcium
carbonate tends to dissolve. The pHs i s c a l c u l a t e d from the e q u a t i o n :
pHs = pK(C,T) + pCa + pAPk
where K(C,T) i s a f u n c t i o n of the temperature and total dissolved solids,
and represents the second dissociation constant and solubility constant
which can be computed from thermo-dynamic considerations. pCa is the
negative logarithm of the calcium content, and pAPk is the negative
l o g a r i t h m of the e q u i v a l e n t concentration of the a l k a l i n i t y . The LSI can be
computed r e a d i l y from F i g u r e 1.2.
Ryzner proposed a d i f f e r e n t arrangement of the terms in the Langelier
equation. The Ryzner S t a b i l i t y Index ( R S I ) i s :
R S I = ZpHs - pH
If the RSI is less t h a n 6, scaling tendency increases, and if it is
g r e a t e r t h a n 8, corrosion i s in f a c t more l i k e l y .
There are many other effects influencing scaling and corrosion,
however, and such formulae can only provide preliminary guides
particularly i n r e l a t i o n to corrosion.

Sulphate s c a l i n g

The s o l u b i l i t y of v a r i o u s forms of calcium sulphate i s higher than that


of calcium carbonate but it is also highly dependent on temperature.
Calcium s u l p h a t e occurs in three different crystal line forms: dihydrate,
CaS04.2H20 (gypsum), herni-hydrite CaS04.)H20 (plaster of paris) and
a n h y d r i te, CaS04.
The solubility of the hernihydrite and anhydrite decreases with
temperature (Figure 1.1 1. The solubility increases with chloride
concentration and i s affected b y t o t a l dissolved solids.

A d d i t i v e s for the p r e v e n t i o n of scale

I n most systems, especially once-through systems, dernineral i z a t i o n or


softening the water w i t h r e s i n o r Zeolite i s not economically j u s t i f i a b l e . In
some cases chemical i n h i b i t o r s a r e used to p r e v e n t the formation of scale.
7

These agents control deposits by preventing crystal growth, even in a


supersaturated solution. The basic mechanisms of scaling and deposit
control a r e :

(1 1 Control of i n t e r p a r t i c l e a t t r a c t i v e forces e.g. dispersants.


(ii) Control of particle-to surface forces, e.g. surfactants or
wetting agents. These involve electrostatic forces. They act
non-stoichiometrically and hence low concentrations are
possible. They are used more for preventing fouling than
scal i n g .
(iii) Control of p r e c i p i t a t i o n rate, e.g. flocculants. These are high
molecu l a r weight pol ymers.
(iv) Retardation of c r y s t a l growth, e.g. polyphosphates.

Some of t h e reagents used a r e l i s t e d below:

Polyphosphates: Applied in rates from 0,5 to 5 mg/O. Absorbed onto the


surfaces of growing c r y s t a l s and i n i n c i p i e n t c r y s t a l n u c l e i . They increase
the apparent solubility of scale forming salts. These are successful for
carbonates and h y d r o x i d e s b u t not f o r sulphates.

Organic Phosphates: Simi l a r to polyphosphates b u t they a r e more s t a b l e i n


cooling tower systems. Phosphonic acids have proved particularly
successfu I .
Phosphate: React w i t h calcium to form i n s o l u b l e calcium phosphate which
p r e c i p i t a t e s out. For this reason i t s use has largely been replaced by
d ispersan ts.

Polymers (especially polyacrylates): Absorbed onto surfaces of crystal


growths. Effectively dispersants as they maintain small particles of
crystals i n suspension. Low molecular weight polymers h a v e r e c e n t l y been
developed f o r t h i s purpose.

The reagents may be used on combination, o r even together w i t h a c i d to


reduce pH. Carbon d i o x i d e can be added to closed systems to reduce pH.
F e r r i c c h l o r i d e i s also used.
Dispersants or sequestrants are sometimes used to prevent scale
formation of i r o n h y d r o x i d e o r o x i d e in p a r t i c u l a r . Chelants o r complexing
agents a r e used to i s o l a t e and i n h i b i t scale formers. It should be noted
8

that crystal inhibitors a r e not effective in e l i m i n a t i n g foulants entering


the systems as suspended sol i d s . Instead, agglomeration of these sol i d s
must be prevented by dispersants such as phosphonates and
I igno-su I phates.
The deposit of phosphates i n closed systems due to a d d i t i v e s can be a
problem. The d u r a t i o n of effectiveness of additives i s also unknown. High
velocities and turbulence can affect low-concentration dispersants in
particular.
The effects of chemical additives on the rest of the system will also
r e q u i r e consideration. Deposits may block pipes o r machines. Suspensions
may erode h i g h - v e l o c i t y jets. Reactions w i t h o t h e r chemicals may a g g r a v a t e
total dissolved s o l i d s problems. There may also be an effect on settlers
and demineralization plants.

FOUL I NG

Besides chemical precipitates there are many substances in suspension


which can settle out o r block p i p e w o r k a n d m a c h i n e r y . The deposits may
materialize in the form of films bridging openings or building up in
c a v i t i e s where water v e l o c i t i e s a r e slow. The m a t e r i a l deposited may be:

Sediment from ore o r the atmosphere t r a n s p o r t e d in suspension


Floc created b y chemical treatment
Iron oxide ( r u s t )
Chemicals used for scale or corrosion inhibition which subsequently
cause deposits
Oils
Foam from chemical r e a c t i o n s o r a e r a t i o n
B a c t e r i o l o g i c a l slime collected o r accumulated i n the system

The tendency to s e t t l e i s a function of particle size, shape, density,


water v e l o c i t y a n d a p p e r t u r e bore. Turbulence due to flowing water will
m a i n t a i n some p a r t i c l e s i n continuous suspension a l t h o u g h the concentration
w i l l b e highest n e a r the bed i n the case of p a r t i c l e s denser t h a n water.

Once p a r t i c l e s s e t t l e out they may stick to the surface, or to other


particles. Alternatively they may migrate along the bed. Under some
conditions the bed m a t e r i a l may move a s dunes w i t h p a r t i c l e s b e i n g p i c k e d
up b y the flow upstream of the dune c r e s t a n d deposited downstream. The
resulting rippled surface can aggravate friction loss in conduits. In
9

addition to the reduction in cross sectional-area, the capacity of the


conduit i s reduced due t o the h i g h e r d r a g on the perimeter.
Deposits may block f i n e pores o r o r i f i c e s completely. The gaps between
filter media particles rapidly block thus requiring backwashing. In
machines w i t h f i n e j e t s o r screens s i m i l a r blockages a r e possible.
Deposits may remain i n flocculated blanket form, or consolidate with
time and i n c r e a s i n g deposits p r e s s i n g down from above. They may s t i c k to
the surface due to chemical bonding.
B i o l o g i c a l matter such as b a c t e r i a l slime o r f u n g i can build up w i t h i n
a water system provided nutrients such as nitrogen, phosphorous and
sometimes carbon and silica, are present. They may be anaerobic (not
r e q u i r i n g oxygen f o r g r o w t h ) o r aerobic. Some b a c t e r i a t h r i v e on iron o r
s u l p h a t e and cause d e t e r i o r a t i o n .

Control o f f o u l i n g

Deposits i n machinery and p i p e systems can be prevented o r reduced b y


controlling particle attraction forces, preventing settling by turbulence,
i n s t a l l i n g s e t t l i n g b a s i n s o r keeping the p a r t i c l e s out of the system (e.g.
closed c i r c u it s ) .
Dispersants are used to control particle-to-particle and
p a r t i cI e- to-surface forces. They n e u t r a I i ze e l e c t r o s t a t i c a t t r a c t ion charges
o r create r e p e l l i n g charges. One problem w i t h these i s that if there are
sedimentation b a s i n s i n the system they may h i n d e r s e t t l i n g there.
High concentrations of dispersants may in f a c t be used f o r desludging
systems. Surface w e t t i n g agents a r e sometimes used to p r e v e n t deposition of
o i l and grease.
Biological fouling may be controlled by disinfection. Shock dosing
treatment appears more e f f e c t i v e a n d economic t h a n continuous dosing.
Chlorine i s widely used as a biocide to combat bio-matter. It is an
o x i d i z i n g agent a n d reduces the pH when dissolved i n water by forming
hypochlorous a c i d a n d h y d r o c h l o r i c acid. A free r e s i d u a l c h l o r i n e content
of less than 1 mg/t i s usually sufficient i f contact p e r i o d i s a n h o u r o r
more. Hypochlorite i s also used occasionaly.
Non-oxidizing biocides act b y surface a c t i v a t i n g o r b y c a u s i n g s u r f a c e
lesions i n the metabolism. Into this category fall quarternary ammonia
compounds ( q u a t s ) .
10

0 I L EMULS ION BREAKDOWN

Emulsions of oil in water are used for driving prototype machinery


amongst other things. The emulsions consist o f oil dispersed in water in
the form of minute droplets. The emulsion is stabilized by electrical
charges on e m u l s i f y i n g agents.
Chemicals (such as polymers of opposite charge polarity) break
emulsions by neutralizing repulsive charges between particles
(coagulation), precipitating or Crystallizing out emulsifying agents or
altering the e m u l s i f y i n g film so t h a t i t can readily be broken. Cations
and cationic polymers are particularly effective in separating dilute
o i I-in-water emulsions. Once charges h a v e been n e u t r a l i z e d , oi I droplets
a n d suspended s o l i d s w i l l be absorbed on the surface of floc o r w i l l break
out a n d f l o a t on top thereby d e s t r o y i n g the emulsion p r o p e r t i e s .
Although it is desirable to maintain the emulsion oil-in-water
suspension w h i l s t in service, after the emulsion i s d i s c h a r g e d to waste it
may be d e s i r a b l e to separate the o i l and the water. This should take
place a t controlled locations to prevent subsequent slime and caking in
machinery f u r t h e r in the cycle. Acid a n d a l u m i n i u m s u l p h a t e ( a l u m ) h a v e
been used to break oil-in-water emulsions. The acid lowers the pH to
about 3 a n d alum coagulates the o i l b y n e u t r a l i z i n g the charges. Lime i s
added to raise the pH again and the aluminium is precipitated as
aluminium hydroxide. Cationic polymers are preferred and often used in
double a i r f l o t a t i o n (DAF) u n i t s which c o l l e c t s the o i l on the surface.

CORROS ION

Corrosion is the attack and degradation of metal by chemical or


electrochemical action. Pipework and machinery are subject to corrosion
due p r i m a r i l y to highly saline or acidic water. The d e s t r u c t i o n may be
general o r i n i s o l a t e d p o i n t s . I t may reduce the l i f e o f p i p e and steelwork
b y many years.
I r o n corrodes in water as follows: It replaces the hydrogen ion in
water since i t i s less noble i.e. i t i s less c a t h o d i c :
Fe + 2H20 = Fe(OHl2 + H2
I n the presence of oxygen, which i s usually i n solution in water, the
ferrous oxide i s oxidized further to ferric hydroxide, Fe(OHl3. This is
insoluble, b u t i s u l t i m a t e l y changed to f e r r i c oxide, Fe203. The reaction
manifests a s p i t s i n the i r o n surface, a form o f oxygen corrosion. (Figure
1.3).
11

TABLE 1.1 Nernst s c a l e of s t a n d a r d e q u i l i b r i u m p o t e n t i a l s r e l a t e d


to the s t a n d a r d h y d r o g e n electrode a t 25OC
(Metal immersed i n a normal s o l u t i o n of one of i t s s a l t s )

Metal Electrode r e a c t i o n s E q u i I ib r i u r n p o t e n t i a l
(volts)

Potassi urn K = K+ + e- - 2.922


Calcium Ca = Ca
++ + 2e- - 2.87
Sod i um Na = Na+ + e- - 2.712
Magnesium Mg = Mg
++ + 2e- - 2.34
Al urn in iurn ~t = AI+++ + 3e- - 1.67
Manganese Mn = Mn
++ + 2e- - 1.05
++ -
Zinc Zn = Z n + 2e- 0.762
C hrom i urn Cr = Cr
+++ + 3e- - 0.71
++ + 2e-
I ron Fe = Fe - 0.440
Coba I t Co = Co
++ + 2e- - 0.277
Nickel Ni = Ni
++ + 2e- - 0.250
Tin Sn = Sn
++ + 2e- - 0.136
Lead Pb = Pb++ + 2e- - 0.126
Hydrogen H2 = 2H+ + 2e- - 0.000 b y convention
Copper cu =
++
cu + 2e- + 0.345
Copper cu = cu+ + e- + 0.522
Silver Ag = A g + + e- + 0.800
++ + 2e-
P I a t inum Pt = P t + 1.2 a p p r o x .
Gold Au = Au
+++ + 3e- + 1.42
Gold AU = AU+ + e- + 1.68

Cathodic
area area
a-
Iron

Fig. 1.3 Corrosion c e l l on the surface o f i r o n in water


12

L
\
\
\

1 - \
\
%
\
\
\
\
Potential \
\

o f Metal - Oxidation
%
%

Eh - \
Corrosion
relative 0 - %

t o hydrogen

- Corrosion
-
-1 - Immunity due t o
- low i r o n p o t e n t i a l

F i g . 1.4 A s i m p l i f i e d form of the P o u r b a i x D i a g r a m f o r i r o n corrosion

If the, or some of the, iron oxides are present as protective layers


they may be eroded b y flowing water especially if sediment i s present.
C a v i t a t i o n can also erode the surface l a y e r s . The metal i s thereby exposed
a n d corrosion i s accelerated.
The e q u i l i b r i u m between i r o n and v a r i o u s compounds i n the presence of
water was studied by Pourbaix. He presented h i s results in a diagram
( F i g u r e 1 .4 ) which shows three zones:

A corrosion zone f o r low pH or high electrical potential relative to


l i q u i d solution.

A corrosion i n h i b i t i o n zone f o r h i g h pH due to passivation by a film


found on the surface

A cathodic protection zone f o r low i r o n p o t e n t i a l r e l a t i v e to a s t a n d a r d


elect rode.
13

Hydrogen is used as a reference electrode in the diagram. The


p o t e n t i a l of the i r o n w i l l depend on the reference system. Table 1 . 1 gives
the e q u i l i b r i u m p o t e n t i a l s of metals immersed in a normal solution of one
of i t s salts, r e l a t i v e to the s t a n d a r d hydrogen electrode at 25OC. There
a r e many texts on f a c t o r s a f f e c t i n g corrosion e.g. U h l i g (1963)

Types of Corrosion

There are many ways i n which corrosion can occur in the presence of
water. Corrosion i s commonly an electro-chemical phenomenon which occurs
a t an anode when electrons flow from an anode to a cathode, leaving a
positively charged anode to react with oxygen. The cathode does not
corrode. Ways i n which the electrons m i g r a t e for corrosion to occur, are
described below ( U h l i g , 1963).

G a l v a n i c Corrosion:
When e l e c t r i c a l l y dissimilar metals are in contact in o r through an
electrolyte, a p o t e n t i a l difference i s established. The more a c t i v e ( a n o d i c )
metal corrodes, as i t i s least noble.

Selective Lea china:


One element of an a l l o y can be corroded more r a p i d l y t h a n another.

Pitting:
A shell of permeable magnetite o r ferric h y d r o x i d e may form over an
iron surface. Salts may concentrate under the shell and the resulting
env ironmen t becomes i n c r e a s i n g I y corrosive.

Stress Corrosion :
Metals in stress may exhibit abnormal corrosive properties in a
corrosive environment. Once a c r a c k i s formed i t r a p i d l y deteriorates due
to s a l t b u i l d - u p s i m i l a r to p i t t i n g . Chlorides a n d amonia appear to be the
chief aggressors i n t h i s type of corrosion. Welding may also induce l i n e s
of corrosion unless stress r e l i e v e d .

A c i d Corrosion :
Acids, o r even carbon d i o x i d e i n solution, can increase the hydrogen
ion concentration. T h i s r e s u l t s i n general loss of metal b y corrosion. Some
chelants, e.g. NTA (nitrilotriacetic acid) may also become c o r r o s i v e as
they concentrate.
14

B a c t e r i a l Corrosion :
B a c t e r i a can cause biochemical a c t i o n which r e s u l t s i n corrosion. This
t y p e of corrosion i s often encountered i n s u l p h u r i c c o n d i t i o n s .

E l e c t r i c a l Corrosion
E l e c t r i c c u r r e n t s , d.c. i n p a r t i c u l a r , may cause severe corrosion. I f an
anode i s formed where the c u r r e n t leaves the conductor, corrosion occurs
there.

Reagent Corrosion :
Certain scale preventing agents such as acids and chelants and
complexing agents can promote corrosion

T h e effectiveness of a l t e r n a t i v e corrosion p r e v e n t i o n methods depends on


the p r e v a i l i n g circumstances a n d system to be protected. I n small closed
cooling systems relatively high concentrations of chemical dosage are
possible. I n l a r g e c i r c u i t s a n d c o o l i n g systems, i n o r d e r to be economic,
the treatment dosage must be less; sometimes little more than pH and
concentration control ( b y b l e e d i n g o f f a n d r e p l a c i n g w i t h f r e s h w a t e r ) can
be accomp I i shed.
I n c h i l l e d water circuits the circulating water may be consumed by
human beings. I n these circumstances i t i s imperative that any treatment
used i s non-toxic. T h i s requirement has the effect of severely l i m i t i n g the
number of chemical corrosion i n h i b i t o r s which can be considered.

Corrosion p r e v e n t i o n

Corrosion can be reduced b y c h a n g i n g the c h a r a c t e r i s t i c s of the water


o r c o a t i n g the metal. Metal i s sometimes i n f a c t coated n a t u r a l l y b y scale.
A u n i f o r m deposit of calcium carbonate can be created b y dosing the water
w i t h lime, soda ash o r c a u s t i c soda. The deposit i s f r e q u e n t l y non u n i f o r m
o r unstable, and cannot be r e l i e d upon f o r 100 percent protection.
Deaeration of water w i l l also reduce i t s c o r r o s i v i t y . I n closed systems,
vacuum deaeration if feasible. Oxygen and carbon dioxide which aid
corrosion a r e thus removed to some extent. Sodium s u l p h i t e can be used to
remove oxygen i n the water. The r e a c t i o n w i t h w i t h oxygen forms sodium
sulphate:
2Na2S03 + O2 = 2Na2S04
15

Corrosion i n h i b i t o r s a r e a v a i l a b l e commercially. One t y p e of inhibitor


passivates the surface by forming a protective oxide film such as
magnetite (Fe304). Other inhibitors react chemically to form insoluble
precipitates. I n t o the l a t t e r category fa1 I zinc, calcium carbonate, calcium
phosphate and ortho- and poly- phosphates.
Other i n h i b i t o r s act b y a b s o r b i n g o r b y p a s s i v a t i n g . The l a t t e r form a
protective f i l m and include chromate, nitrate, molybdate and tungstate.
Silicates also appear to work on s i m i l a r p r i n c i p l e s .
In general the corrosion rate is dependent on c o n d u c t i v i t y . pH and
oxygen. I t increases w i t h c o n d u c t i v i t y up to a limit, whereas it i s most
significant when the pH drops below 4 (see Fig. 1.5). Oxygen content
increases corrosion r a t e , e s p e c i a l l y a t h i g h e r temperatures.
Chromates a r e p a r t i c u l a r l y e f f e c t i v e corrosion inhibitors. Concentrations
up to 300 m i l l i g r a m s per litre in open circuits and 2000 m i l l i g r a m s per
I i t r e i n closed c i r c u i t s a r e used. I t i s therefore costly, and its toxicity is
a deterrent. Additives of zinc and phophate reduce the chromate
r e q u i remen t s.
To overcome the t o x i c i t y problem of chromates, s u i t a b l e ortho-phosphate
and polyphosphate m i x t u r e s have been developed. To p r e v e n t p r e c i p i t a t i o n
of calcium orthophosphate at orthophosphate concentrations above 5 to 7
milligrams per Iitre, an inhibitor such as phosphonate can be added.
Simultaneous p a s s i v a t ion of the anodic areas a n d p r e c i p i t a t i o n of calcium
salts a t the cathodic zone to form a p r o t e c t i v e layer i s thereby possible
( r e f e r r e d to as d i a n o d i c protection, a p r o p r i e t r y name), (Betz, 1980).
F i l m i n g amines such as octadecylamine act differently. They form a
p h y s i c a l b a r r i e r , often monomolecular i n n a t u r e .

POTABLE WATER STANDARDS

Although i n d u s t r i a l water i s not often intended f o r human consumption


the q u a l i t y should b e adequate to ensure no harm if it i s consumed. It
should be non-toxic, and i f d r u n k in l i m i t e d q u a n t i t i e s showld show no i l l
effects. The upper l i m i t s to dissolved s a l t s f o r p o t a b l e water a r e d i f f i c u l t
to f i x . They depend on the amount consumed and it should be b o r n in
m i n d t h a t men could d r i n k up to 2 l i t r e s a s h i f t .
M i c r o b i o l o g i c a l m a t t e r in the water can be more concern than dissolved
salts. After d i s i n f e c t i o n , normally with chlorine, bacteria and viruses are
not n o r m a l l y present i n mine service water, b u t r e g u l a r checks should be
made. Toxic substances include heavy metals, concentrated fluorides,
nitrates, some algae, organic phosphates and some poly-electrolytes (the
16

l a t t e r two a r e used in t r e a t i n g water sometimes)


Highly mineralized water possesses laxative properties. It may also
a f f e c t the sweating process, blood p r e s s u r e o r the c a r d i o - v a s c u l a r system.
Often human perception (taste, smell or colour) will identify the
possibility of unsafe water. Phenols, chlorine and organic matter are
e a s i l y detected b y taste.
Suggested l i s t of l i m i t s to c e r t a i n substances f o r p o t a b i l i t y i s given in
Table 1.2. Table 1.3 indicates the maximum allowable concentrations of
o t h e r t o x i c substances.

Fig. 1.5 The effect of pH on the corrosion r a t e .

TABLE 1.2 Recommended p o t a b l e water s t a n d a r d s .

Substance Concentrat ion


mg/e

A I k y I ben zenesu Ifona t e ( ABS ) , tast e-produc in g 0.5


Arsenic ( A s ) 0.1
C h l o r i d e ( C 1 1, taste-producing 250.0
Carbon chloroform e x t r a c t (CCE), taste-producing
possi b I y t o x i c 0.2
Cyanide (CN) 0.01
I r o n ( F e ) , taste- a n d colour-producing 0.3
Manganese ( M n ) , taste- and colour-producing 0.05
N i t r a t e (NO ) , p r o d u c i n g methemoglobinemia 45.0
3
P heno I s , t as t e-p r o d uc i n g a n d tox ic 0.001
Sulphate (SO)&),taste-producing and l a x a t i v e 250.0
Total dissolved solids, laxative 500.0
Zinc ( Z n ) , taste p r o d u c i n g 5.0
17

TABLE 1 . 3 Toxic concentrations in water

Substance Concentration,
mg/t

Arsenic (As) 0.5


Barium ( B a ) 1 .o

Cadmium (Cd) 0.01


6+) 0.05
Chromium (hexavalent, Cr
Cyanide ( C N ) 0.02

Lead ( P b ) 0.05
Selenium (Se) 0.01
Silver (Ag) 0.05

AGRICULTURE AND IRRIGATION

I r r i g a t i o n i s a major consumptive use of water. Crops cannot tolerate


h i g h s a l t loads and y i e l d s d e t e r i o r a t e unless remedial a c t i o n i s taken. The
f o l l o w i n g t a b l e shows levels of s a l t s which a f f e c t crops.

TABLE 1.4 Water Q u a l i t y which affect crops

Lower l i m i t Upper l i m i t

T DS mg/P 500 2000


Root a b s t r a c t i o n :
Chloride mg/P 150 350
L e a f water a b s t r a c t i o n
(sprinkling)
Chloride mg/P 100 1000
Nitrates mg/P 5 30

Rapid assessment of TDS i s often possible b y measuring conductivity.


The conductivity in mS/m is approximately equal to the TDS (total
dissolved s o l i d s c o n c e n t r a t i o n ) i n m g / t d i v i d e d b y 6.5.
F i g 1.6 shows the decrease in yield for some crops as a function of
s o i l moisture salinity.Some crops are more r e s i s t a n t than others to salts
due to t h e i r p u r i f y i n g a b i l i t y . For instance, vegetables a r e more r e s i s t a n t
than f r u i t , b u t a r e less p r o f i t a b l e .
There i s also the gradual deterioration in s o i l to contend with. Salt
b u i l d s up due to e v a p o r a t i o n a n d t r a n s p i r a t i o n of water. The s a l t s can b e
leached out b y a p p l i c a t i o n of excessive water, but, for instance, at least
25% more water would be r e q u i r e d to ensure good s o l i d c o n d i t i o n s w i t h the
TDS l e v e l s of 800 mg/e. More i r r i g a t i o n equipment i s also r e q u i r e d to cope
w i t h the h i g h e r flows.
The a l t e r n a t i v e i s to change the c r o p p i n g p a t t e r n . Crops r e q u i r i n g less
water o r a d a p t a b l e to s a l i n e water would h a v e to be developed.

\ Lucerne\ I

I I \ \ I

C o n d u c t i v i t y of g r o u n d w a t e r (mS/m)

Fig. 1.6 Crop y i e l d as a f u n c t i o n of water q u a l i t y


19

REFERENCES

Barenbrug, A.W.T., 1965. Psychrometry and psychrometric charts.


T r a n s v a a l a n d O.F.S. Chamber of Mines. Johannesburg.
Betz. 1980. Handbook of I n d u s t r i a l Water C o n d i t i o n i n g , 8 t h Ed., Betz,
Trevose, 440 pp.
Brownlow, A.H., 1979. Geochemisty, P r e n t i c e H a l l , N.J. 498 pp.
L a n g e l i e r , W.F. 1954. Journal America1 Water Works Assn., 46, 461.
Mrost, M. a n d L l o y d , P.J., 1980. B a c t e r i a l O x i d a t i o n of W i t w a t e r s r a n d
Slimes, I .A.E.A. Johannesburg.
P e l l e t i e r , R.A. 1964. M i n e r a l Resources o f South - C e n t r a l A f r i c a . O x f o r d
U n i v e r s i t y Press. Cape Town. 277 pp.
Porges, J. 1971. Handbook of Heating, V e n t i l a t i n g a n d A i r C o n d i t i o n i n g .
6 t h Ed., Newnes-Butterworths, London.
Ryzner, J.W. A p r i l 1944. A new index f o r d e t e r m i n i n g t h e amount o f
calcium carbonate scale formed b y water. JAWWA, 36, 472-473.
U h l i g , H.H. 1963. Corrosion a n d Corrosion Control. John Wiley a n d Sons,
N.Y.
Van Staden, C.M.V.H., 1970. Steps Taken b y t h e M i n i n g I n d u s t r y to
Prevent a n d Overcome Water P o l l u t i o n . Water f o r the F u t u r e Convention,
Pretoria.
CHAPTER 2

MATHEMAT I CAL MODELL ING O F WATER QUALI TY

A f i e l d to which many systems concepts can be a p p l i e d i s t h a t of water


q u a l i t y deterioration in industrial systems. Cooling and washing systems
are examples where quality will deteriorate in time. It is not easy to

predict the rate of build-up of dissolved salts or the equilibrium


concentrations i n water r e t i c u l a t i o n systems, even with an understanding
of the o r i g i n s and methods of concentration of salts. This i s because of
t h e complex n a t u r e of industrial water recirculation systems. One way of
accounting for all these effects in a real system appears to be by
m o d e l l i n g the system on a computer.
Once a model i s produced a n d v a l i d a t e d , i t may be used to improve the
operation of e x i s t i n g s e r v i c e water r e t i c u l a t i o n systems and f o r o p t i m i z i n g
the design of new systems.
The build-up of impurities in water can be simulated mathematically
together w i t h the water r e c i r c u l a t i o n cycle. The flows of water i n conduits
o r i n vapour form i n the a i r in and out of the system can be c a l c u l a t e d .
The processes of e v a p o r a t i o n , condensation, p o l l u t i o n a n d make-up can a l l
be modelled.

MASS BALANCES

For the purposes of mathematical simulation of water systems, the


system must b e described in terms of equations. One-stage systems can be
described in terms of a mass balance equation which can be solved
analytically. I n other more complex situations it is necessary to express
the equations in finite difference form and solve them numerically.
D i f f e r e n t types of models and the assumptions t h e r e i n a r e described below.
Parameters whereby pollution i s measured may e i t h e r be c o n s e r v a t i v e o r
non-conservative. I n a c o n s e r v a t i v e system i n p u t to a n y p a r t of the system
equals outflow. Thus, if the parameter studied is water flow then
evaporation will be neglected in a conservative model. Similarly if the
parameter is a chemical compound it is assumed there is no reaction,
deposition o r s o l u t i o n i n a c o n s e r v a t i v e model.
The model may be steady-state or time-varying. During the start-up
p e r i o d of a mine as concentrations build up the system is said to be
unsteady. After a w h i l e the system may reach e q u i l i b r i u m . That is, in the
case of s a l t s in s o l u t i o n , the increases i n mass of d i s s o l v e d s o l i d s in t h e
21

system due to leaching or evaporation equals the loss by pumping or


deposition.

MIXED AND PLUG FLOW SYSTEMS

In a plug-flow system, the water is assumed to travel through the


pipes a n d d r a i n s a t a c e r t a i n r a t e , conveying i m p u r i t i e s a t t h a t r a t e . The
s a l t s content a t any p o i n t can therefore be affected in a series of steps
as water with different concentrations arrives at that point. In a
completely mixed system, the concentration of salts will be the same at
every p o i n t . An input i s assumed to spread instantaneously through the
systems so that the concentration increases by the mass of salt input
divided by the total volume of water in the system. This simplified
mechanism i s often s a t i s f a c t o r y to describe systems which e x h i b i t gradual
r a t e s of change i n concentrations. Real systems w i l l probably be between
p l u g flow and completely mixed, as there w i l l be d i f f u s i o n and m i x i n g due
to tubulence and cross connections. In general salts are conveyed by
advection ( l a t e r a l t r a n s p o r t ) and dispersion.

Examples

The simplest i l l u s t r a t i o n of the use of the mass b a l a n c e equations i s f o r


a steady-state system. Q is flow rate in e/s or MP/d, C is the
concentration i n mg/e. I n f l o w of water a n d of salts per u n i t time e q u a l s
outflow r a t e :

F i g 2.1 Point Node

Flow Balance a, + a2 = a3
Mass Balance ale, + a2c2 = a3
.*.C = alcl+a2c2
3
Q1 +Q2
22

e.g. i f Q1 = 5 MP/d, Q2 = 10 M e / d ( w a t e r flow r a t e )


C 1 = 400 mg/P, C2 = 100 mg/P ( s a l t c o n c e n t r a t i o n )
then C3 = 200 m g / t
a n d the t o t a l mass of s a l t discharged p e r day
= Q3C3 = 15 x 200 = 3000 k g / d (2.4)

A completely m i x e d system can be described by differential equations:


Subscript i r e f e r s to i n f l o w , e to e x i t , s to i n i t i a l conditions:

Volume S
Conc. C

Fig. 2.2 M i x e d flow node

QiCi =
d (SC)
Q C + -
e dt

= Q C + SdC
x f o r constant S

.*. dt =
SdC
Qi C i-QeC

Fig. 2.3 Diffuse node

Integrating and e v a l u a t i n g the constant of integration from the fact


that C = C at t = 0 :
T = S en QiCi-QsCs (2.8)
- 1
( QiCi-QeC
Qe
23

QiCi QiCi/Qe- Cs (2.9)


or C = -- Qe e( Qet/S
1

e.g. a t t = 0, C = C s , and a t t = m , o r Be = - o r S = 0,
c = (ai/ae)ci
Observe that i f Q . does not equal Qe, there must be i n t e r n a l gains or
losses, e.g. due to evaporation.
The previous example could be studied numerically. Although this
r e q u i r e s specific numbers, it is often the o n l y practical way of solving
more complex problems.
Assume S = 1000 m 3 , Q . = lm’/s = Q C s = 0, Ci = 500 mg/P.
e’
Choose ~t = 100 5. The choice of ~t can affect the speed of solution,
the accuracy of r e s u l t s a n d the numerical s t a b i l i t y of the computations. It
must be determined by trial, from experience or from theoretical
considerations.
c 2 -c 1
NOW Q.C. - Q C = 5- (2.10)
I I At

.’. C2 = C
1
t
5
8 . ( C . -C
1 1 1
) = C1 + 0.1(500-C1) (2.11)

The computations can be set out in t a b u l a r form as follows:

t 500-C1 xo.l
c1 c2

0 0 500 50 50
100 50 450 45 95
200 95 405 40 135
300 135 365 37 172
0

1000 326 1 74 17 343 mg/t

Equation (2.9) would indicate C = 316 mg/e at t = lOOOs, which is


comparable with the result indicated by the numerical solution of 343
mde.
A plot of C versus t is called a pollutograph, and CQ versus t a
loadograph. T h e numerical computations f o r change of p o l l u t a n t load i n t h e
above table are very similar to flood routing calculations assuming for
example the Muskingum method.
Reservoirs a r e g e n e r a l l y assumed to be completely mixed, whereas r i v e r s
24

a r e sometimes assumed to be p l u g flow. I n fact i n b o t h there i s a degree


of m i x i n g due to:

a) Molecular diffusion due to Brownian movement (negligible in most


h y d r a u l i c systems).
b) Turbulent mixing, due to eddies i n the stream.
c) Short circuiting or t r a c k i n g e.g. in reservoirs where a track is
made across the water body by the flow. The stagnant water in
corners i s c a l l e d dead water
d) Wind m i x i n g
e) Thermal m i x i n g a n d i n v e r s i o n (e.g. Henderson-Sel lers, 1979).

The degree of m i x i n g can affect concentrations so much t h a t m o n i t o r i n g


systems need to account f o r i t (Sanders, 1983).

SYSTEMS ANALYSIS

A more sophisticated approach than the simulation method described


above i s the use of systems a n a l y s i s a n d o p t i m i z a t i o n techniques, with the
assistance of computers i f necessary. The methods a l l o w an optimum design
to be selected from numerous a l t e r n a t i v e s (Thomann, 1374).
The alternative standard engineering approach is to select the best
o p t i o n from a few selected designs. The l a t t e r approach is tedious where
there are many a l t e r n a t i v e s .
The design o p t i m i z a t i o n approach involves the creation of a general
configuration i n which the numerical v a l u e of independent variables has
not been f i x e d . A n o v e r a l l economic o b j e c t i v e i s defined a n d the system is
described i n terms of equations o r c o n s t r a i n t s .

TERMINAL CONCENTRATION I N A WATER C I R C U I T .

The t o t a l dissolved s o l i d s concentration i n a closed water recirculation


system w i l l b u i l d up due to e v a p o r a t i o n and a d s o r p t i o n o r leaching. The
c o n c e n t r a t i n g effect will continue indefinitely unless s a t u r a t i o n occurs, or
water is replaced. Make-up water will replace polluted water and the
relative proportion of rate of replacement to water in circulation will
control the e q u i l i b r i u m d i s s o l v e d s o l i d s concentration. Computation of the
e q u i l i b r i u m concentration i s performed a s follows:

F I :~ ai ~ = aD + ae (2.12)
25

c I c,

a. P l u g flow

b. Completely mixed system

c. D i f f u s e systea

Fig. 2.4 Comparison of p l u g f l o w a n d m i x e d systems.


26

S a l t s : QiTi + QiTe - (2.13)


- QPTP

where Qi i s the water input rate (.e.g. l i t r e s per' second or megalitres


per d a y ) ,
Q i s the d i s c h a r g e pumping r a t e
P
Q i s the e v a p o r a t i o n r a t e
T. i s the concentration of s a l t s i n the replacement o r makeup water
stream.

T is the concentration build up due to leaching, expressed in


e
terms of the incoming water flow r a t e here.
Tp i s the concentration in the pumped water which i s the same as
the c i r c u l a t i n g water f o r a mixed flow system.

I f Q i s i n m e g a l i t r e s a d a y and T i n mg/t then QT has the u n i t s of


k i l o g r a m s of s a l t p e r d a y . Solving f o r T the salt concentration in the
P'
system,

(2.14)

Thus f o r no l e a c h i n g ( T = 0 ) and a n e v a p o r a t i o n r a t e equal to 50% of


the pumping r a t e , T = 1.5 Ti i.e. the e q u i l i b r i u m s a l t concentration will
P
be 150% of that of the make-up water.

APPLICATION TO A MINE WATER CIRCUIT

South A f r i c a n g o l d mines use n e a r l y 2000 m i l l i o n l i t r e s of water a day


underground (Holton a n d Stephenson, 1983). The water is used p r i m a r i l y
f o r dust control a n d cooling. Owing to the great depths i.e. often over
3000 metres below surface, rock temperatures can r e a c h 65°C. The most
e f f i c i e n t method of c o o l i n g i s b y means of s p r a y i n g c h i l l e d water onto the
rock. The water i s also used f o r ore moving and to a l i m i t e d extent for
h y d r a u l i c emulsions in machinery.
The geological formations in which gold i s mined are in the Orange
Free State and T r a n s v a a l which s u f f e r water shortages. Water is i n fact
imported from a d j o i n i n g catchments such as i n Natal f o r domestic d r i n k i n g
purposes. The cost of water is therefore high and re-use of water is
encouraged both to conserve water and to minimize the discharge of
pol l u t e d wastewaters i n t o surface streams. The water requirements of the
g o l d mines a r e therefore l a r g e l y met b y r e c y c l i n g and only approximately
10% of the requirement is made up from surface water resources. Some
27

mines also have s u r p l u s underground water from infiltration and this is


used where possible.

#
Pure w a t e r

Remaining
concentrated
S a l i n e water

F i g . 2.5 Model of s a l t b u i l d u p due to e v a p o r a t i o n

The q u a l i t y of surface water is good and the total dissolved solids


content i s typically less than 500 mg p e r litre. The quality of ground
water where there i s any, i s also g e n e r a l l y good as the water o r i g i n a t e s
l a r g e l y from dolomitic a q u i f e r s in the upper s t r a t a . Although the water is
h a r d and contains magnesium and calcium carbonates, the dissolved salts
concentration i s r a r e l y above lOOOmg per litre. In the Orange Free State
on the other hand the natural water is known to contain high
concentrations of chlorides.

,Qi

Average
salinity I

Fig. 2.6 Model of s a l t e q u i l i b r i u m due to pumping


28

Despite the general p u r i t y of make-up water, concentrations of d i s s o l v e d


a n d o r g a n i c s a l t s u n d e r g r o u n d can t y p i c a l l y vary from 3000 to 10000 mg
per Iitre. T h i s water can therefore o n l y be used f o r l i m i t e d purposes. Care
has to be taken to ensure that it is not used for drinking, in certain
machines and with heat exchange apparatus. In many mines there is
s c a l i n g a n d f o u l i n g of m a c h i n e r y a n d p i p e w o r k because of the poor q u a l i t y
of water and in other mines there is corrosion of pipework and other
metal-work.
Reasons for the deterioration in mine service water can be attributed
primarily to l e a c h i n g from the mined ore. In a d d i t i o n c e r t a i n pollutants
are brought from the source of the water and there is a secondary
concentration effect due to the e v a p o r a t i v e loss of water underground and
i n cooling towers. The chemical dosing of water for purification purposes
and for corrosion and scaling inhibition can be ruled out as a major
contributor owing to the small dosage rates relative to the increase in
dissolved solids i n the water.

Evaporation

Groundwater

F i g . 2.7 Mine water r e t i c u l a t i o n system

After i d e n t i f y i n g p o s s i b l e sources of chemical salts appearing in the


water, l a b o r a t o r y a n d s i t e t e s t i n g were performed to e v a l u a t e the p o s s i b l e
r a t e of appearance of s a l t s in the water. The complicated n a t u r e of the
geochemical environment make% a n exact q u a n t i t a t i v e estimation of leaching
in any particular circumstance impossible. Nevertheless relative r a t e s of
l e a c h i n g can be assessed with these methods. Once pilot tests had been
29

conducted to i d e n t i f y the prime effects of water qua1 i t y d e t e r i o r a t i o n , these


parameters were s t u d i e d i n more d e t a i l . I t appeared t h a t the fineness of
the crushed ore originating from blasting or drilling was a prime
parameter i n a f f e c t i n g the r a t e of geochemical leaching. The composition of
the reef i.e. the ore b e a r i n g stratum, also i s a factor. The contact time
w i t h the water affects the amount of l e a c h i n g from any p a r t i c u l a r mass of
ore. Temperature affects the chemical reaction, as does the pH of the
water, a n d i n i s o l a t e d cases p o s s i b l y the presence of b i o l o g i c a l matter, in
particular thio-bacillus ferro-oxidans and thio-oxidans. The presence of
a i r appeared i n a l l cases s u f f i c i e n t to s a t u r a t e the water w i t h oxygen a n d
therefore was not a l i m i t i n g factor.

Conditions unless otherwise stated 209 fine, 3OoC, a i r b u b b l e d t h r o u g h 2e

/' 5-409 fine crushed ore

100 -

00-
E
s,E
-
-
.-f>
.-
"
a
60

c
P

40-

d0VS

F i g . 2.8 L a b o r a t o r y l e a c h i n g tests on crushed o r e

L a b o r a t o r y tests were performed b y immersing samples , o f ore crushed to


various finenesses in one to two litres of water. Temperature was
controlled b y means of a b a t h and a i r was b u b b l e d t h r o u g h the samples to
agitate and provide sufficient oxygen. Datum tests with pure distilled
water were performed simultaneously. Tests were run for longer than a
month and conductivity and various dissolved salt parameters were
measured r e g u l a r l y as well as pH a n d temperature.
30

The r a t e of l e a c h i n g of a t y p i c a l b a t c h of samples i s indicated in F i g .


2.8. I t will be observed that the l e a c h i n g r a t e s were most rapid during
the f i r s t day a n d then gradually decelerated as the s o l u b l e chemicals in
the ore were depleted. Confirmatory tests w i t h initial water concentrations
a t v a r i o u s levels i n d i c a t e d t h a t s a t u r a t i o n of the water was not the cause
for the reduction in leaching rate, The effects of different passes of
crushed ore, d i f f e r e n t sizes of particles and temperature, presence of air
a n d a g i t a t i o n were s t u d i e d i n d i f f e r e n t samples.
The increase i n t o t a l dissolved s o l i d s in the water v a r i e d from 5 to 30
grams of dissolved s o l i d s per k i l o g r a m of crushed ore.
The following inorganic salts were detected in the water samples
a n a l ysed : sulphates, chlorides, carbonates, nitrates, c a l c i urn magnesi urn,
sodium as well as o t h e r elements in the relative order indicated. The
concentration of sulphates in milligrams per Iitre (mg/P) of SO4 was
t y p i c a l l y one h a l f of t h e t o t a l d i s s o l v e d s o l i d s concentration i n mg/e. This
can be a t t r i b u t e d to the h i g h s u l p h i d e concentration in the ore (up to 8
per cent sulphur by mass). In the presence of oxygen and water some
s u l p h i d e forms in p a r t i c u l a r were o x i d i z e d to sulphates. The i r o n from the
r e a c t i o n was often precipitated as iron oxide and the chemical reaction
which is well known in both coal mining and gold mining pollution
problems i s i n d i c a t e d below:
+ O2 + H20 4 FeO(OH) + H2S04 (2.15)
The pH of the s o l u t i o n remained between 6 a n d 8 f o r the f i r s t week in
most cases. By the end of the second o r t h i r d week the pH often dropped
to below f o u r . As the pH dropped an a c c e l e r a t i o n i n the l e a c h i n g r a t e , as
i n d i c a t e d b y an increase i n conductivity and total dissolved, solids, was
evidenced. The presence of b a c t e r i a was noticed in i s o l a t e d samples a f t e r
a month of t e s t i n g , b u t not i n a l l samples in which the p H dropped o r the
r a t e of l e a c h i n g was noted to be p a r t i c u l a r l y high.
It is therefore concluded that the leaching reaction is primarily a
geo-chemical r e a c t i o n a n d b i o l o g i c a l r e a c t i o n can be s a i d to be small in
the environment studied.
The application of the laboratory results to the field conditions is
particularly complicated. It is not only the total mass of fine ore
generated by mining operations which is of importance, but also the
exposed surface of the fine which settles out rapidly in the horizontal
d r a i n s t a k i n g water back to the s h a f t . Only the surface l a y e r appears to
leach a t a high r a t e and this may explain the relatively low leaching
rate underground as compared with the maximums measured in the
laboratory. Tests in the field could only indicate increase in total
31

dissolved solids of the o r d e r of 100 to 300 mg/e per cycle as the water
r a n from the workings back to the s h a f t .
I t was therefore not possible to i n s e r t the complete chemical process in
equation form into the computer model of the system. Empirical
r e l a t i o n s h i p s were therefore used and these will have to be verified for
each mine and each o r e mined

COMPUTER SIMULATION MODEL

The r a t e s of use of water underground vary considerably during the


day and are highest d u r i n g the d r i l l i n g a n d ore moving s h i f t s . Water is
often stored in the cascade dams underground o r in surface dams at
various stages. Fluctuation in water quality is therefore difficult to
p r e d i c t unless the volumes of all the storage dams as well as the flow
rates in the various conduits can be modelled. External flows such as
evaporation, water removed w i t h the ore, seepage a n d i n t e r m i t t e n t make-up
additions also affect internal volume, flow rates and quality. The most
logical method of simulating the process was with a digital computer
model. This was adapted to a micro computer w i t h considerable success.
A general simulation program was developed for simulating specific
models of mine water systems. Models a r e constructed i n general form for
p a r t i c u l a r mines. The sizes of dams, the p o s i t i o n s and the c a p a c i t i e s of
conduits and the usage hydrographs can then be specified by the user.
The o p e r a t i n g r e l a t i o n s h i p s which, for example, define the salt leaching
rate, criteria for adding make-up water, starting pumps etc. are
programmed as p a r t of the model source code. The computer program will
then be used to simulate the model. Flow rates, volumes and dissolved
s a l t s concentrations a r e d i s p l a y e d a t specified time i n t e r v a l s as output.

Mathematical Basis of Model

The computer model was p r e p a r e d in a modular structured fashion for


easy updating and modification. The varying volumes and salt
concentrations a r e described i n models b y means of first-order ordinary
differential equations. Alternative methods of solving the equations
n u m e r i c a l l y are b u i l t i n t o the s i m u l a t i o n p r o g r a m and the methods can be
selected to s u i t the p a r t i c u l a r equations. I n many cases a fast algorithm
i s s u i t a b l e w h i l e i n o t h e r cases a more a c c u r a t e a l g o r i t h m i s r e q u i r e d to
solve the equations w i t h s u f f i c i e n t accuracy on a numerical basis.
In a mine water system many processes occur simultaneously and the
32

net effect i s e i t h e r to increase o r decrease the d i s s o l v e d s a l t concentration


of flows a n d water volumes w i t h time. The s a l t concentration of water in a
storage element, such a s dams, depends on the mass of salts and the
volume of water i n storage, a n d the s a l t c o n c e n t r a t i o n of a n y inflaw and
outflow. Denote Q1 a n d Q2 as the i n f l o w a n d outflow to a dam, a n d denote
Cl a n d C 2 as the corresponding s a l t concentrations. I f M i s the mass of
dissolved s a l t a n d V the dam volume a t a c e r t a i n time t, then the r a t e o f
change of water volume a n d s a l t mass w i t h time i s

(2.16)

and -
dM
dt
= Qi.Cl - Q2.C2 (2.17)

If perfect mixing is assumed to occur in the dam then the salt


concentration of the outflow is:
M
(2.18)
c2 = 0

A mine water model b a s i c a l l y consists of e q u a t i o n s ( 2 . 1 6 ) and ( 2 . 1 7 ) f o r


each storage element wherein the volume and s a l t mass change w i t h time.
Other r e l a t i o n s h i p s govern flow r a t e s a n d changes i n s a l t concentrations of
flows between storage elements.
S t a r t i n g w i t h known o r assumed initial v a l u e s f o r a l l M, V a n d C the
d i f f e r e n t i a l equations a r e n u m e r i c a l l y solved u s i n g E u l e r a n d Runge-Kutta
methods. Values of M, V and C are determined at each iteration
time-increment d u r i n g the s i m u l a t i o n a n d can be d i s p l a y e d as o u t p u t . The
s t a b i l i t y and accuracy of the solution depends v e r y much on the time step
a n d numerical method selected.
Considerable e f f o r t h a s to be expended i n g a t h e r i n g data for the model
it i s found. Owing to the u n p r e d i c t a b l e changes i n m i n i n g p a t t e r n s as the
characteristics of the Reef change, the water reticulation pattern is
continually b e i n g extended o r a l t e r e d . The c o n d u i t s and dams constructed
therefore form a complex storage distribution system which is often not
monitored as it i s designed to operate automatically. Flow rates, stored
volumes and times of makeup were therefore often difficult to ascertain.
The model can thus oe used to p r e d i c t the water q u a l i t y a t a n y time a t
any point in t h i s system for alternative operating conditions. Fig, 2.9
indicates a typical variation in flow rate at the workings underground
and Fig. 2.10 i n d i c a t e s the water q u a l i t y v a r i a t i o n in the water pumped
to the surface a t the same g o l d mine over a p e r i o d of a week.
33

12h00 24h00 l2hOO

Fig. 2.9 Flowrate from Coldwell to u n d e r g r u n d (M4/d)

T h e i n i t i a l conditions in starting up a n d running the model could be


varied to an extent. That i s the initial water quality could be varied
assuming that d i f f e r e n t make-up q u a n t i t i e s of surface water c o u l d be used
to replace poor q u a l i t y water i n the surface storage dams over a weekend
when m i n i n g a c t i v i t i e s were m i n i m a l . By comparing a l t e r n a t i v e management
p o l i c i e s i n t h i s manner i t i s possible to r e a c h a minimum cost procedure
f o r m a i n t a i n i n g the water q u a l i t y a t a c e r t a i n selected level. The r a t e of
usage underground was assumed fixed by the mining operation and
therefore o n l y storage dam c a p a c i t i e s and make-up r a t e c o u l d be v a r i e d in
t h i s way.
I t i s also possible t h a t m i n i n g methods c o u l d be v a r i e d to affect the
water q u a l i t y . I t was recognised t h a t the contact time between f i n e ore in
suspension and i n the r e t u r n water systems h a d a n important bearing on
the rate of deterioration in the water quality. Alternate methods of
returning the water were therefore investigated i n order to optimize the
water quality. In t h i s manner the effects of p o l l u t i o n can be minimized,
therefore r e q u i r i n g less surface make-up water a n d r e d u c i n g m i n i n g costs
i n e l i m i n a t i n g to a l a r g e extent s c a l i n g and erosion.

12000

10000

8000

5000

4000

2000

TueS Yed Thur Fri Sat Sun Hon Tues


F i g . 2.10 Salt concentration i n the settlers
34

REFERENCES

Henderson-Sellers, B. 1979. Re se rv o i rs , M c M i l l a n , 1 2 8 ' p .


Holton, M.C. a n d Stephenson, D . , 1983. A computer model of c i r c u l a t i n g
s e rv i c e wa t e r i n South A f r i c a n g o l d mines. I n t . J. M i n e Water , 2 ( 2 ) p
33-42.
Sanders, T.G. ( E d . ) , 1983. Design of Networks for M o n i t o r i n g Water
Q u a l i t y . Water Resources P u b l i c s . 328 p .
Thomann, R . V . , 1974. Systems A n a l y s i s a n d Water Q u a l i t y Management.
McGraw H i l l , 286 p .
35.

CHAPTER 3

NON CONSERVATIVE PARAMETERS

INTRODUCTION

Mass balances are not always possible. Many constituents in still


waters change concentration n a t u r a l l y . Some r e a c t chemically to r e s u l t in
d i f f e r e n t salts. I f a l l the s a l t s before and a f t e r r e a c t i o n a r e s o l u b l e the
total concentration of dissolved salt in mg/e in the water remains the
same. Sometimes oxygen i s taken out of the water to release hydrogen gas
which i s more v o l a t i l e and escapes.
Oxygen i n water i s the cause of many changes. For instance ammonia is
oxidized to nitrites, and these in turn are oxidized to nitrates. The
n i t r a t e s cannot be e l i m i n a t e d except by chemical replacement, absorption
or biochemically, as i s now done in some waste water treatment processes.
Absorption of oxygen and other chemicals i n water may occur due to
biological matter in water. Decay i s generally approximated by a first
o r d e r equation

- _ - KC
at

BASIC MASS BALANCE EQUATION

The one-dimensional balance equation allowing for dispersion, decay


and sources or s i n k s i s d e r i v e d below
Source

ISdtAdx

direct i o n

Decay Kc Adxdt

F i g . 3.1 Mass b a l a n c e
36

Net increase i n mass of C i n element in time d t i s


- KC Adx - C -
aa - as
dC.Adx = dt (SAdx a x dx Q
a x d x + a x ( A € aa sx ) d x j

For a u n i f o r m channel A = constant and t = constant a n d Q = constant

... ac
-a +
t k C + v s- a2c
E,,I-~=O (3.3)

1 ) i s r a t e of increase in concentration of p o l l u t a n t
2) i s decay r a t e
3) i s advection
4) is diffusion
5) i s source

E i s the t u r b u l e n t diffusion coefficient. It is similar to the k i n e m a t i c


viscosity which represents transfer of momentum between layers in a
f r i c t i o n model, e.g.
du
T = p E - against a wall, (3.4)
dy
where
(3.5)

where U=
, v'(T/P) = shear velocity
a n d k i s the von Karman constant, 0.4.
But i t i s not t h a t simple i n channels as not o n l y molecular diffusion
b u t macro turbulence, tracking, dead water, s t r a t i f i c a t i o n etc. complicate
the action, therefore one needs to c a l i b r a t e models.
E l d e r ( D e i n i g e r , 1973) suggests E = A h J(ghS) (3.7)
where h = depth a n d A = coefficient ( a v e r a g i n g 0.07).
Normally d i f f u s i o n i s n e g l i g i b l e in r i v e r s , except estuaries.
Thus one gets the Streeter-Phelps equation

-a_t - - v ac-_
ax
KC
( o m i t t i n g sources)
(3.8)

(3.9)

(3.10)

K ranges from 0.01 per day i n laboratory conditions ( a s found b y A r n o l d ,


1980) w i t h p u b l i s h e d f i g u r e s f o r r i v e r s a v e r a g i n g 0.1 per day.
37

or

F i g . 3.2 Decay curves

OXYGEN BALANCE I N R I V E R S

Oxygen concentration in a r i v e r i s measured in terms of DO ( d i s s o l v e d


oxygen). Shortage o f oxygen i s measured as a chemical oxygen demand
(COD) o r a biochemical oxygen demand (BOD). The long term BOO i s about
1.45 x BOD5 where BOD5 i s the BOD as measured in a laboratory over 5
days, a s t a n d a r d test (AWWA, 1965).

Coupled equations for DO a n d BOD

I f DO concentration i s designated C and BOD i s L then


a2c ac
-a _t ax - v~
- E 7 - K,L + K,(C - C) * S c
(3.11)
38

Dissolved Oxygen
Dernond Sag Curve
Dissolved
Oxygen
P-.L._-I n-:-A
Lriilcui ruin1
R e oxygenot ion C u r v e

Deoxygenalion Curve

Distance Downstream
E f f luen t or Time
Outtall

Fig. 3.3 The d i s s o l v e d o x y g e n sag curve

J,

Olcygen
0 Carbonaceous plus

F i g . 3.4 Carbonaceous and nitrogenous o x y g e n demand c u r v e s


39

where C = s a t u r a t i o n conc. of oxygen

aL a2L aL (3.12)
and - = E 7- v- - K,L, t S
a t ax ax L

These simultaneous equations can be solved a t p o i n t s a l o n g a r i v e r a n d


over time increments, K, = K~~~~ e (T-20) i.e. it is a function of
temperature.

charac t e r i s t i C

At
n - l

F i g . 3.5 Solution g r i d

A s a n example of t h e solution of these two equations a two-step explicit


method can be employed (Deininger 1973, p 122). One can get pseudo
diffusion where A x f v A t (3.13)
unless a c a r e f u l numerical procedure i s used.
Where the river is depleted of oxygen, the BOD equation must be
replaced b y
KIL = K ( C - C ) - Sc (3.14)
2 s
i.e. the q u a n t i t y of oxygen consumed i s equal to the quantity of oxygen
introduced i n the same time (Thomann, 1972).

Ana I y t ica I solution

dC
I f - = - K L + K (C - C ) (3.15)
dt 1 2 s
a n d oxygen d e f i c i t D = C - C (3.16)
2:
_ - - KIL - K2D (3.17)
Integrating gives
KILO -K,t -K,t -K, t
D = - (e -e 1 + Doe
Kz-Ki (3.18)
One can also e v a l u a t e K, and K2 a t t C ( D e i n i n g e r 1972 p 126).
40

CALIBRATION OF A MOVING BOD MODEL

As an a p p l i c a t i o n of the c a l i b r a t i o n of a r i v e r oxygen model, the K l i p


r i v e r in South A f r i c a was analyzed. The K l i p r i v e r h a s d i s c h a r g i n g into i t
effluent from major municipal sewage works and runoff from an
underdeveloped township. The stream i s also h i g h l y m i n e r a l i z e d a n d flows
t h r o u g h reed beds. Measurements of BOD a n d DO o v e r summer and winter
show r a p i d n a t u r a l self a e r a t i o n . The waters a r e e v e n t u a l l y recycled w i t h
o t h e r sources. A numerical model p r e d i c t s d a i l y v a r i a t i o n s i n BOD a n d DO.
The decay coefficient and sources and sinks were fitted by linear
programming opt im iza t ion.
The Klip river rises in the watershed of the Witwatersrand. On its
banks are three major municipal sewage works and a large residential
area. Separate s a n i t a r y sewers a r e p r o v i d e d g e n e r a l l y but a tendency to
l i t t e r i n g r e s u l t s in h i g h l y p o l l u t e d s u r f a c e runoff.
The population of the area is nearly 2 million. Of a total water
consumption within the watershed of the Klip r i v e r of approximately 500
m i l l i o n l i t r e s per day, n e a r l y 50 percent i s r e t u r n e d to the K l i p r i v e r v i a
sewage p u r i f i c a t i o n works o r separate storm sewers ( u n t r e a t e d ) i .e. 2m3/s.
The base flow of the r i v e r i n the reaches s t u d i e d amounts to o n l y lm’/s.

OXYGEN BALANCE

The dissolved oxygen content (DO) of water i s a useful i n d i c a t o r of its


ability to support life. A lower level of 4 m g / t i s r e g a r d e d as the limit
f o r f i s h l i f e i n the area studied.
The r a t e of which dissolved oxygen reduces the biochemical oxygen
demand i s dependent on the level of free oxygen concentration. The u p p e r
l i m i t i s the s a t u r a t i o n concentration, C s , estimated to be
C S (mg/e) = 14.6 - 0.41T + 0.008T‘ - 0.000778T’ (3.19)
where T i s i n “C
The DO i n a p o l l u t e d stream v a r i e s a l o n g the l e n g t h i n accordance w i t h
the r a t e of takeup a n d the r a t e of re-oxygenation (Fig. 3.3). I n addition
to b i o d e g r a d a t i o n of carbonaceous o r g a n i c matter, oxygen i s required for
nitrification, oxidizing i n o r g a n i c chemicals a n d p l a n t respiration. With a
high sulphur concentration in the waters, due to mining activity, the
oxygen requirement i s f a i r l y high. T h i s i s counterbalanced to some extent
b y the h i g h llime content, as the waters o r i g i n a t e from a dolomitic area.
Temperature and sludge deposits in winter also influence the oxygen
demand.
41

Owing to deposits of sludge in the slow moving stream v e l o c i t y (less


t h a n 0.2 m/s) d u r i n g w i n t e r months, BOD was observed to increase. After
the summer rains the deposists were scoured out and a more rapid
reoxygenation was observed. The sludge arose primarily from organic
matter, while benthal deposits were considered relatively inactive (Velz,
1970).
There are two primary biochemical oxygen abstractors; carbon and
nitrogen. The BOD removal c u r v e t y p i c a l l y e x h i b i t s a n i n i t i a l hump due to
carbon a n d a subsequent hump due to nitrogen (Fig. 3.4). The decay
equation smooths the c u r v e out.
The coupled d i f f e r e n t i a l equations d e s c r i b i n g the v a r i a t i o n of BOD a n d
DO a r e 3.11 and 3.12 r e w r i t t e n i n the form

(3.20)

(3.21)

A method of e v a l u a t i n g the coefficients K, and K2, a n d the source a n d


sink term, S and P, so that the equations represented the real river
system, would be to f i n d values f o r these parameters that would lead to
the minimum total difference between the concentrations predicted by
equations 3.20 and 3.21 and the actual concentrations observed in the
field. Linear programming may be used f o r m i n i m i s a t i o n of an objective
f u n c t i o n subject to c e r t a i n constraints provided the system is linear. In
the above case, the o b j e c t i v e f u n c t i o n would b e

Minimise { z I Predicted BOD - observed BOD I


+Z I Predicted DO - observed DO I} (3.22)

subject to the c o n s t r a i n t s formed by the system of equations and to the


c o n s t r a i n t that the e r r o r p l u s the p r e d i c t e d v a l u e must e q u a l the observed
value.
I n other words the calibration of the model can be carried out by
m i n i m i s i n g the sum o f the absolute errors.
Another method would be b y .means of least squares f i t t i n g techniques.
This has been attempted elsewhere, using the data from the sampling
survey on 21 March, 1979 (McPherson a n d S h a r l a n d , 1979) equations 3.20
and 3.21. The I inear ' programming method has been used be Kleinecke
(1971 f o r estimating geohydrologic parameters of groundwater basins.
O L - . ; ' . ' - . . . . I . . . . . 1 . . . . . .
6h00 l2hOO 18h00 24h00
1 1 . . ...
06h00 06h00 12h00
I . . . . . .

18h00
. . a . A

24h00
l . . . . . I I1 .....
06h00 06h00 12h00
I . . , . . . ..... .....
18h00
I
24h00
I ,

06h00

Fig. 3.6 Results of S i m u l a t i o n using minimum e r r o r c a l i b r a t i o n p a r a m e t e r s


43

1-1 I 1+1 x

F i g . 3.7 x - t grid

Considering the concentration-space grid in Fig. 3.7 i t can be shown


how the above coupled equations 3.20 and 3.21. can be formulated for
l i n e a r programming e v a l u a t i o n of the parameters as follows:
The BOD concentration a t a p o i n t P can be w r i t t e n in terms of implicit
f i n i t e differences as:

- (L 1
2At i,n + L i + l , n - Li,n-l + Li+l,n-l

For I i n e a r programming purposes two requirements must be met:


(i) a l l terms must be l i n e a r
(ii) a l I v a r i a b l e s must be non-negative
I n the above f i n i t e difference form these c o n d i t i o n s a r e not s a t i s f i e d .
F i r s t l y the term - K + L. + L. + Li,n-l) i s not
t , i /4 . ( L i - , , n 1-1 ,n-1 1In
l i n e a r since both the K1 and the L. a r e unknowns. Secondly the net
1,n
source/sink term may be e i t h e r p o s i t i v e o r negative.
To overcome these problems the prediced L. a r e r e p l a c e d b y the known
1,n
observed values b . a n d the source/sink term i s s p l i t i n t o a n i n p u t term
1,n
+ S and an output term - T where one of S a n d T w i l l be p o s i t i v e a n d the
other zero. The equation then becomes
44

( L 1. ,n + Li+l,n)/2 - (Li,n-l + L .i+1 ,n-1 ) / 2

-- -UAt
-
2 Ax
(Li+l,n + Li+l,n-l - L .i,n - L i , n - l 1

+ At Si - At T i (3.24)

T h i s can be r e w r i t t e n as

1 U At 1 UAt (-1 - UAt


-)-L. 1 UL!)
Li,n-l (2 + d + Li+l,n-l (Z - ~ x - ) i,n 2 2Ax i+~,n ‘Z + 2AX

- - *
K1 iAt
(bi+l,n + bi+l,n-l + bi,n + bi,n-l 1

+ AtSi - AtTi = O (3.25)

I n addition another set of equations can be w r i t t e n in terms of the


e r r o r b y which the p r e d i c t e d v a l u e of L . d i f f e r s from the actual value
1,n
of L .
1 ,n
.
L. + Ui,n - Vi,n = b. (3.26)
‘,n i,n

Again the requirements that the v a r i a b l e must be p o s i t i v e necessitates


the s p l i t t i n g of the e r r o r i n t o a p o s i t i v e e r r o r U o r a n e g a t i v e e r r o r -V,
one of which w i l l be zero i n the solution.
Similarly a set of equations can be written for equation 3.22. This
includes a reaeration term which i s also non-linear unless the observed
values a r e s u b s t i t u t e d f o r the p r e d i c t e d values.
These equations a r e g i v e n below

1 UAt
- ci+l,n (- + -126x
2

+ A t Pi - At R i = o (3.27)
C. + M. - N. = d. (3.28)
1,n 1,n 1,n 1,n
45

Equations 3.27 a n d 3.28 can be w r i t t e n f o r a l l p o i n t s i, except the last


point, a l o n g a study reach f o r which observed d a t a i s a v a i l a b l e over a
p e r i o d of time. The observed values b a n d di,n a t each v a l u e of n may
i ,n
have to be inerpolated from observations taken a t o t h e r times. O n l y
equations 3.26 and 3.28 can be w r i t t e n f o r the p o i n t f u r t h e s t downstream.
The o b j e c t i v e f u n c t i o n now becomes

Minim i se
{; (Ui,n; + V.
1,n
+ Mi,n + N.
1,n
subject to the c o n s t r a i n t s g i v e n b y equations 3.27 a n d 3.28.
(3.29)

F I ELD MEASUREMENTS

The length of stream modelled was 6 km. It was divided into four
reaches and two sets of samples were taken a s representative, one set in
mid w i n t e r and one i n mid summer. Samples were taken e v e r y hour f o r 24
hours of each section, which was p r o b a b l y a b i t sparse. DO was measured
w i t h a p o r t a b l e meter. The samples were tested f o r 5-day and 20-day BOD,
COD a n d pH, conductivity, ammonia, nitrate, nitrite, chloride, alkalinity
and suspended sol i d s were determined. Photosynthetic oxygen release was
estimated from light and dark bottle tests, and time of passage and
dispersion were determined w i t h f l u o r e s c i n dye,
Various methods were employed to calibrate the simulation model :
linear programming was used to minimize the absolute value of the
differences between observed and s i m u l a t i o n concentrations of BOD a n d DO.
The method is described elsewhere. In order to render the equations
I inear, the theoretical concentrations were approximated by observed
values whenever p r o d u c t s of two unknowns appeared in the equations. This
may have been the result of often apparently high decay rates and
unaccounted f o r sources a l o n g some of the reaches. The methods a r e b e i n g
extended to non-l i n e a r equations (McPherson and Sharland, 1979) with
encouraging results.

The i n p u t parameters f o r the p l o t s g i v e n in Table 3.2 were d e r i v e d b y


t r i a l a n d e r r o r f i t s i n the model.
Even then, there appeared i n e x p l i c a b l y h i g h BOD o r COD sources a l o n g
the r i v e r reaches. These were a t t r i b u t e d to b e n t h i c deposits o r r u n o f f from
adjacent sewage irrigation works, and seepage from the industrial and
other townships to the north.
The accuracy of the BOD measurements a t the levels observed (5 to 10
mg/4) i s questionable, due to the complex way of d e t e r m i n i n g i t . Various
46

researchers have proposed TOC (total organic carbon) or COD (chemical


oxygen demand) as i n d i c a t o r s of oxygen demand. Due to the high inert
f r a c t i o n of COD, the change i n COD may be a more a p p r o p r i a t e parameter
t h a n COD, and t h i s in f a c t gave b e t t e r r e s u l t s t h a n the BOD model.
The sampling frequency of 1 h o u r was r a t h e r coarse. Once r e s u l t s were
plotted it was realized that pollution loading varied rapidly. This was
more likely due to surface runoff than to the e f f l u e n t from the sewage
works.
The decay r a t e of the COD was estimated to be up to 3,O per day,
which i s h i g h i n comparison w i t h laboratory results and other published
data. This may be due to high turbulence, or the high salinity of the
water promoting reactions.
Photosynthesis was noticeable o n l y on v e r y overgrown reaches. A value
of 3 mg/P/day was t y p i c a l .
Oxygen s i n k s were found to be l a r g e in w i n t e r ( u p to 75 mg/P/day) but
n e g l i g i b l e i n summer ( t h e r a i n y season).
The dissolved oxygen content was found to be s u f f i c i e n t to support life
(above 3 mg/P) a t a l l stages.
T y p i c a l r e s u l t s a r e i n c l u d e d as Tables 3.1 to 3.3 a n d F i g u r e 3.6.

REFERENCES

American Water Works Association, 1%5. Standard Methods for the


Examination of Water a n d Wastewater.
Arnold, R.W., 1980. M o d e l l i n g Water q u a l i t y in the u p p e r K I i p r i v e r .
MSc(Eng) Dissertation, U n i v e r s i t y of the Witwatersrand.
Deininger, R.A., 1973. Models f o r Environmental P o l l u t i o n Control. Ann
Arbor.
Kleinecke, D., 1971. Use of linear programming for estimating
geohydrologic parameters of groundwater basins. Water Resources
Research, 7 ( 2 ) , p 367-374.
McPherson, D.R. and Sharland, P.J., 1979. River Quality Tests.
Undergraduate project, U n i v e r s i t y of the W i t w a t e r s r a n d .
Thomann, R.V., 1972. Systems A n a l y s i s a n d Water Q u a l i t y Management.
McGraw H i l l , N.Y.
Velz, C.J., 1970. A p p l i e d Stream S a n i t a t i o n . Wiley Interscience, N.Y.
TABLE 3.1 Results of c a l i b r a t i o n u s i n g d a t a of 21 M a r c h 1979 ( e n d of summer)

COD C a l i b r a t i o n BOD C a l i b r a t i o n

Value Value

Parameter jyrnbo Units ?each 1 3each 2 leach 3 teach 1 Reach Z Reach : Method of determination

Reaches 1 a n d 2 -
Tracer studies
Dispersion coeff i c i e n E 10.4 0.4 10.0 10.0 0.4 10.0 Reach 3 - C a l i b r a t i o n

Decay c o e f f i c i e n t 0.05 1.3 3.0 3.0 2.0 4.0 Ca I i b r a t ion


K1
Reaeration coeff i c i e n 2.0 2.7 3.0 2.0 2.7 7 .o Reaches 1 and 2 -
K2
Formula
Reach 2 - C a l i b r a t i o n

BOD source/sink (1) S -1 80 330 100 75 30 -50 Calibration

DO source/sink (2) 8 10 0 30 10 20 Calibration


R1
Photosynthetic DO ( 2 3 0 0 2 5 5 Bot t I e tests
p1
I n e r t source/sink (1 -1 50 175 80 Not applicable Calibration
) 1 -
Notes (1) A n e g a t i v e v a l u e i n d i c a t e s a source ( p o s i t i v e b e i n g a s i n k ) .
(2) A p o s i t i v e v a l u e i n d i c a t e s a source ( n e g a t i v e b e i n g a s i n k ) .
The i n e r t f r a c t i o n o f t h e i n p u t COD was t a k e n a s 60%
The BOD5/BOD20 r a t i o was t a k e n a s 0.69.
48

TABLE 3.2 Results of model f i t t e d to COD d a t a of 18 J u l y 1978


(mid-w in t e r

Value
-- - Method of
Parameter ymbol each each each Det erm i n a t ion
1 2 3
--
D i spers ion E 10.4 0.4 10.0 Assumed same a s
coefficient for March survey

Decay 0.1 1.3 1 .o Model f i t t i n g


K1
coefficient

Reaera t ion 2.0 2 .7 5 .0 Reaches 1 a n d


K2
coefficient 2 - formula
Reach 3 - model
fitting

BOD source/
sink ( 1 ) S -32 175 50 Model f i t t i n g

DO source/sinb
(2) -1 3 13 -1 5 Model f i t t i n g
R1
Photosynthesis
(2)
DO 3 5 5 Bottle tests
p1
~ n e r tsource/
(1)
sink -48 20 50 Model f i t t i n g

- - -
Notes: (1) A n e g a t i v e v a l u e i n d i c a t e s a source ( p o s i t i v e b e i n g a s i n k )
(2) A p o s i t i v e v a l u e i n d i c a t e s a source ( n e g a t i v e b e i n g a s i n k )

The i n e r t f r a c t i o n of the COD was taken as 60%


49

TABLE 3.3 Program output


(.I D.I. LISIIng (lor mol
K l l p R l v w Slmlatlm

ENn 1.7s bn

-.
I

Vn = 23.40 kruleay
DFFN l0.00kruld.y
aa*
. 0.050 1/*y

.-.
KFm 2.00 I/&Y
SINt -190.0 n p / l / d a Y
91RE 8.0 n p / l / d a Y
Po0
SINt I . 3.0 np/l/dmy
-150.0 np/l/dmy

5.30 bn
20.03 krulday
10.00 krulday
3.m I/&y
1.00 I/&Y
100.0 -/1/dmy
0.0 n p / l / d a Y
0.0 np/l/daY
m.0 np/i/ea~

lncrl F n c l l m : 0.00 0.00 -


O l u m l V a r l a l l m of Rolos,mlhesla used IFW VE IRO-ll
aLn - 0.007 days

--
O a X I l l * 0.203 bn
OELXIII 0.241 bn
OaXl3l 0,174 km
NO of pace 1n1ew.I~ NX
No of Ilm Inl.walsNT
- 27
123

IbI 1yplc.l Lonplludlnal Oulpul

ULIPRIVER SIMULATION - Run 364 Tlm. aO.00 h r s

DI.l.nc. Sirnulaled Ov..r".d


Ownsirearn 9 o o w moo
statim m E 27.0 2.8

0.0 27.0 2.9


0.250 35.6 3. 0
0.500 15.0 3.2
0,750 52.3 3.3
I .ooo 5a.A
.. _._
1.5
1.250 62.7 3.5
1.500 24.3 3.6 0
1.7% 62.4 3.5 0
0 0
SIallm M F 57.1 3.6 0 0
0 0
2.W 50.6 3.4 0
2 . m 54.6 3.1
0
0
2.503 49.7 i.1 0
0
1.750 44.0 3.4 0
3.W 17.6 3.5 0 0
3.250 30.9 1.7 0 0
0 0
Slallm M G 22.0 b.0

3.500 U.6 3.9


3.750 24.0 1.9
4.m 23.3 4.3
4.250 22.9 4.1
4.503 22.9 4.1
4.750 23.0 4.2
5.m 20.2 4.2
5.254 23.3 4.2

Slmllm No. HI41 11.9 4.0


50

TABLE 3.3 Contd.

( c l T y p i c a l Time V a r i a t i o n Output
KLlPRlVER SIMULATION - RUN

eOp a n d 00 VARIATION = I T N TIME AT STATION F

.
TIME SIMULATED OBSERVED

N0.5 m a , Kn 00 5 10 15 20 25 30 3 5 44 4 2 50 55 60 6: ?$ :7 80

.
0
0.0 57.56 3.38 55.0 3.52 ox 0.
1.04 59.69 3.45 59.0 3.51 on
2.09 57.55 3.52 56.00 4.00 ON 0.
2.49 53.24 3.65 54.00 4.02 on 0 .

.. ..
3.06 52.68 3.91 59.00 4.26 0% 0 .
5.00 52.24 4.21 59.00 4.50 a .
6.04 52.55 4.47 60.97 4.50 XO 0 .
7.07 M.55 4.56 60.02 4.57 SO
7.92 49.75 4.71 55.00 4.62 NO
8.% 49.32 4.76 61.23 4.62 NO 0 .

... . ....
10.00 47.41 4.73 57.62 2.37 LO 0 .
11.04 51.56 4.65 58.80 4.25 SO 0 .
12.05 55.93 4.48 60.95 4.13 *O 0.
12.92 61.39 4.19 60.17 3.66 SO
13.96 66.30 4.04 60.24 3.81 X

..
15.00 69.45 3.76 56.00 3.78 no
16.04 70.40 3.65 56.90 3.56 NO
17.09 69.30 3.53 56.36 3.34 no . I I
0

.. .
17.92 53.32 1.54 60.46 3.33
18.96 53.95 3.42 66.55 3.52 0 0 .
20.00 60.98 3.43 57.07 3.60 0 0 .

.
21.36 49.00 3.48 39.95 3.60 0
22.09 49.12 3.50 47.13 3.62 ON 0.
22.92 51.33 3.46 53.00 3.60 0 6
23.94 54.72 3.47 37.50 3.35 0 0
25.0 54.91 3.46 27.70 3.39 0
I
5 10 15 20
51

CHAPTER 4

NUMER ICAL METHODS

SIMULATION OF HYDRAULIC SYSTEMS

Simulation of systems described b y d i f f e r e n t i a l equations can be done in


a number of ways:

F i n i t e elements
Characteristics
F i n i t e difference - Implicit - Four p o i n t 0-0

F i n i t e difference - Explicit - Four p o i n t 7


'V

Leap f r o g

Diffusive

Backward centred

L a x - Wendroff = d i f f u s i v e / l e a p frog

E x p l i c i t schemes a r e simple b u t not as accurate o r stable as implicit


schemes. Problems which manifest w i t h e x p l i c i t schemes i n c l u d e numerical
i n s t a b i l i t y and numerical d i f f u s i o n . I n s t a b i l i t y can occur i f the time step
is too great. The accepted stability criterion for diffusive schemes is
(Deininger, 1973);

Ax2/At 2 2E (4.1

or At 5 A x2/2 E (4.2)

There is an additional problem, that of numerical diffusion i.e.


s p r e a d i n g of the p o l l u t i o n gradient due to successive c a l c u l a t i o n s using
concentrations a t adjacent points. From a second o r d e r Taylor expansion
the maximum numerical d i f f u s i o n is E max = A x 2 / 8 A t . ( D e i n i n g e r 1973)
P
Using the previous expression f o r At, we get the pseudo d i f f u s i o n cannot
be less than €/4.
52

Two-step method

The water q u a l i t y equation i n c l u d i n g the d i f f u s i o n term can be solved


in two steps to ensure c o r r e c t a d v e c t i o n a n d d i f f u s i o n . Thus s t a r t i n g w i t h

a‘c
_ _ EaxZ ac kC
- v - (4.3)
at ax
c. - Cidl
use aC
Ax
=
Ax
(4.4)

a2c - ci+l - 2ci + ci-l


(4.5)
ax7- A X2

then C .
i,n+l
= c.
i,n
+ EAt Ci-l,n+Ci+l,n-2Ci,n - vAt ‘i+l,n-‘i,n - kC.
1,n
Ax2 A X
(4.6)
The f i r s t and last two terms on the right hand side in the above
equation for advection and decay can be used to get the first
a p p r o x i m a t i o n to C . a n d then the d i f f u s i o n term.
I, n+l

F i g . 4.1 Basic r e c t a n g u l a r x - t grid

Demonstration of n u m e r i c a l i n a c c u r a c y

The convection term in the water quality equation will be used to

i I l u s t r a t e problems a n d i n a c c u r a c i e s d u e to a n i n c o r r e c t n u m e r i c a l scheme.
Neglecting the d i f f u s i o n a n d decay term, we h a v e

‘i ,n+l = C.
1,n
- v ~ t‘i+l,n - ‘i,n
A X (4.7)
53

We should have a wave of concentration move downstream a t a r a t e v,


unattenuated o r changed i n concentration.

C i ,n

i-1 i it1 it2


AX

F i g . 4.2 Theoretical advection

I f Ax = vAt
then u s i n g a f o r w a r d difference e x p l i c i t method
= c. - c.i , n )
‘i ,n + l 1,n - (‘i+l,n
= 1 - (0-1)

= 2 which i s wrong, i t should be 0


i.e. dont use a f o r w a r d difference ac/ax = (Ci+l - Ci)/~x
Instead use a b a c k w a r d difference ac/ax = (Ci - Ci-l)/Ax
Then ‘i,n+l = c .t,n - (‘i,n - ‘i-l,n 1
(4.7b)
= 1 - (1-0) = 0, correct.
on the other h a n d i f we use AX = 2vAt,

= 1 - 0-0 = 1, also wrong.


2

I f we continued w i t h t h i s scheme, the v a l u e of C o s c i l l a t e s (see below)

0.5

F i g . 4.3 O s c i l l a t i n g scheme
54

O n the other h a n d i f one uses a backward difference with Ax = 2vAt


n u m e r i c a l d i f f u s i o n occurs a s i n d i c a t e d below.

1 -

F i g . 4.4 Numerical d i f f u s i o n

I f At > Ax/v we get n u m e r i c a l i n s t a b i l i t y , e.g.


if At = 2 Ax/v,
C. = Ci,n - *Ax
( C .i,n - ‘i-1,n 1
(4.7d)
1 - 2 (1-0) = -1.

C o n t i n u i n g so, a n o s c i l l a t i n g c u r v e occurs:

I \

F i g . 4.5 Instability
55

I m p l i c i t f i n i t e d i f f e r e n c e schemes

i-1 i i +1 X

F i g . 4.6 I m p l i c i t scheme

(4.10)

becomes C. z c, - vAt ). A l l values a-t n + l a r e


~,n+l i,n ('i,n+l - 'i-1 , n + l
unknown and a set of i equations is established for i unknowns. The
method is unconditionally stable but solution of the i simultaneous
equations can be l e n g t h y , especially for non l i n e a r systems, e.g. if we
use the hydrodynamic equation with the term v -
ax
this is non-linear
since v . ( V i , n + l - ' i - l , n-
+l) i s parabolic.
I , n+l (4.11)
Ax
V.
So r a t h e r use 'i,n ('i,n+l - I - 1 , n t l ) which i s l i n e a r .
(4.12)
A X

Methods of solution of i equations include ( F r i e d , 1975)

i) Direct methods e.g. m a t r i x methods and Gauss e l i m i n a t i o n .


ii) I t e r a t i v e method - i.e. assume reasonable v a l u e s f o r all C's and
i t e r a t e the equations s u b s t i t u t i n g assumed values on the r i g h t h a n d
side until the left h a n d side agrees with assumed values. This
o n l y converges i f At < AX/V.
...
III) Relaxation methods (Timoshenko, 1951).
iv) A l t e r n a t i n g d i r e c t i o n i m p l i c i t procedure (Fried, 1975), i.e. compute
56

derivitive w i t h respect to x i m p l i c i t l y and y e x p l i c i t l y a n d then


v i c e versa ( s t a b l e ) .

One also gets combined explicit/implicit methods for more accuracy


(e.9. McDonnel a n d O'Conner, 1977).

Comments on f i n i t e difference methods

E x p l i c i t method:

1. T h i s must be designed to be s t a b l e i.e. a n y e r r o r s due to 2nd o r d e r


terms in the T a y l o r expansion (we took j u s t the f i r s t o r d e r ) must decay
d u r i n g comp u t ion.
The time interval must t h r e f o r e be smaller than for implicit method.
For e x p l i c i t hydrodynamic equation, using Fourier series it may be
-
shown to be s t a b l e if 2 Jgy = wave celerity i.e. speed of
computation g r e a t e r t h a n speed of a d i s t u r b a n c e i n the system.

2. I t must be accurate. Check with a few space a n d time intervals and


a g a i n s t an a n a l y t i c a l s o l u t i o n i f there i s one.

3. I t shohld minimize numerical d i f f u s i o n

4. One can use v a r y i n g g r i d s where g r e a t e r a c c u r a c y i s r e q u i r e d :

F i g . 4 . 7 V a r y i n g g r i d s p a c i n g (zooming)
57

NUMER I CAL METHODS FOR THE SOLUT ION OF SINGLE D IFFERENT IAL EQUATIONS

Numerical solutions appear i n the form of a t a b u l a t i o n o f the values of


the functions of v a r i o u s values of the independent time v a r i a b l e and not
as a f u n c t i o n a l r e l a t i o n s h i p . Numerical methods h a v e the a b i l i t y to s o l v e
p r a c t i c a l l y any equation b u t they h a v e the d i s a d v a n t a g e t h a t the e n t i r e
t a b l e must be recomputed i f the i n i t i a l c o n d i t i o n s a r e changed.
I f a f u n c t i o n f ( t ) can be represented b y a power series i n a certain
i n t e r v a l then i t can be represented b y the T a y l o r series expanded about a
p o i n t t = to, i.e. about the i n i t i a l value:

I
Y ( t ) = y (tO)+Y ( t o ) ( t - t O ) + y
II
-
( t o ) (t-t0)2+y
I1
-
( t o )(t-t0)3+ ... (4.13)
2! 3!
L e t t i n g n represent the p r e v i o u s step at time to a n d n + l represent the
next step a t t +h, the series can be w r i t t e n as:
0

-yn I I + c y n
Yn+l=~n+hyn l+h2
Ill+
*.. (4.14)
2 6

Consider the examp Ie p r o b lem

(4.15)

w i t h i n i t i a l conditions
Y(0) = 1 (4.16)
This is a linear time variant 1st order differential equation. The
a n a l y t i c a l solution to the problem,
y = 2e-t-1 w i l l be used to compare the
t
numerical r e s u l t s o f some of the methods a n d t o i l l u s t r a t e the e r r o r a t a n y
step.

The E u l e r Method

The E u l e r method i s the simplest b u t least accurate of a l l the methods


discussed. To o b t a i n a n exact numerical s o l u t i o n to the example problem
(4.151,
II
a l l the d e r i v a t i v e s y , y
I l l, y I V ...
must be e v a l u a t e d a n d
s u b s t i t u t e d i n t o the T a y l o r series (4.14). Knowing the i n i t i a l values of y
n'
I
yn , yn
II
..., Y,+~ c o u l d be e v a l u a t e d a f t e r a time increment h. The
values of a l l the d e r i v a t i v e s c o u l d then be c a l c u l a t e d a t n+l, and y
n+2
could be evaluated a f t e r the next time increment a n d so on. D e r i v a t i v e s of
a r b i t r a r y functions cannot e a s i l y be formulated in computer programs. The
derivatives y l ' , Y l I I , etc. a r e easy to e v a l u a t e f o r the example (4.14)
58

b u t t h i s i s not g e n e r a l l y the case. The E u l e r method t r u n c a t e s t h e T a y l o r


series b y e x c l u d i n g the terms a f t e r the f i r s t d e r i v a t i v e a n d e l i m i n a t e s the
problem o f h a v i n g to evaluaate the second and subsequent derivatives.
Then
yn+l=yn+hynl+O(h') error (4.17)

Neglecting h ' y n " / 2 and the subsequent terms in (4.14) results in a


t r u n c a t i o n e r r o r of order h' which i s denoted O ( h * ) . This is the l o c a l
e r r o r a n d r e s u l t s from one step o n l y , i.e. from n to n+l. I t can b e shown
t h a t the g l o b a l e r r o r accumulated over many steps becomes O(h), i.e. an
e r r o r of o r d e r h.
S u b s t i t u t i n g the example (4.15) i n t o the E u l e r a l g o r i t h m (4.17) gives:
Yn+l=Yn+h. (Yn+tn) (4.18)

The i n i t i a l c o n d i t i o n y ( O ) = l means that y=O at t=O. Choosing the time


increment h=0.02 a n d l e t t i n g the step number n=O a t t=O, the v a l u e s f o r y
can be evaluated a t successive time increments a s follows:

y =y + h ( y o + t O ) = 1+0.02(1+0) = 1.0200 (4.19)


1 0
y =y +h +t ) = 1.0200+0.02( .0200+0.02) = 1.0408 (4.20)
2 1 Yl 1
y =y +h y +t ) = 1.0408+0.02( .040+0.04) = 1.0624 (4.21)
3 2 2 2
= 1 .ow0 (4.22)
y4
= 1 .lo81 (4.23)
y5
etc.

Anal y t i c a1
solution

. c

F i g . 4.8 The Euler method

t
The numerical solution a f t e r 5 steps i s y(0.10)=1 .lo81 whereas y=2e -t-1
g i v e s the exact a n a l y t i c a l solution as y(0.10)=1.1103. Hence the a b s o l u t e
g l o b a l e r r o r i s 0.0022, i.e. two-decimal-place accuracy. Since the global
59

e r r o r of the E u l e r method i s p r o p o r t i o n a l to h, i.e. O(h), t h e step size h


must b e reduced at least 22-fold to g a i n four-decimal accuracy, i.e. h
<0.004. T h i s would increase the computational e f f o r t 22-fold. Fig. 4.8 shows
how the slope a t the b e g i n n i n g o f the i n t e r v a l ynl i s used to determine
the f u n c t i o n v a l u e a t the end o f t h e i t e r a t i o n i n the E u l e r method.
The slope a t the b e g i n n i n g of the i n t e r v a l i s a l w a y s wrong unless t h e
solution i s a s t r a i g h t l i n e . Thus t h e simple E u l e r method s u f f e r s from t h e
d i s a d v a n t a g e of l a c k of accuracy, r e q u i r i n g a n extremely s m a l l step size.

The M o d i f i e d E u l e r Method

F i g 4.8 a n d the subsequent discussion suggest how the E u l e r method c a n


be improved with little additional computational effort. The arithmetic
average o f the slopes a t the b e g i n n i n g a n d the end o f t h e i n t e r v a l i s used
( o n l y the slope a t the b e g i n n i n g i s used in t h e E u l e r method).

1 1
yn+l = Yn + h'n +',+I (4.24)
2
I
The E u l e r a l g o r i t h m must f i r s t be used to predict yn+l so that y
n+l
can be estimated. Applying the same example (4.15) as before a n d
s u b s t i t u t i n g y1 = x+t i n t o (4.24) gives

Y n + l --yn+h(Yn+tn) + (~,+~+t,+l) (4.25)


2
S u b s t i t u t i n g the E u l e r e q u a t i o n (4.18) f o r Yn+l gives

'n+l = yn+h('n +t n 1 + (yn+h(Yn+tn) + tn+l 1 (4.26)


2
Using h=0.02 a n d the i n i t i a l conditions: y =l,t =O
0 0

(4.27)

= 1 + 0.02 (1+0) + (1+0.02(1+0)+0.02) (4.28)


2

= 1.0204 (4.29)
60

y 2=
1.0204 + 0.02(1 .0204+0.02)+(1.0204+0.02(1.0204+0.02)+0.04)
2 (4.30)

= 1.0416 (4.31)

y5= 1.1104 c f a n a l y t i c a l s o l u t i o n 1.1103


The answer agrees to w i t h i n 1 in the f o u r t h decimal p l a c e . Nearly twice
a s much work was done a s in the E u l e r method b u t certainly not the 22
times more that would h a v e been needed with that method to attain four
decimal p l a c e a c c u r a c y . I t can be shown t h a t the l o c a l a n d g l o b a l errors
of the Modified Euler method are O(h3) and O(h2) respectively. The
Modified Euler and the simple Euler methods are often referred to as
second a n d f i r s t o r d e r methods r e s p e c t i v e l y .

Runge-Kutta Methods

The Fourth-Order Runge- K u t t a methods a r e amongst those which p r o v i d e


t h e greatest a c c u r a c y p e r u n i t o f c o m p u t a t i o n a l e f f o r t . The development of
t h e method i s a l g e b r a i c a l l y complicated a n d i s g i v e n completely in Stummel
and Hainer (1978) while Gerald (1980) derives the Second-Order
Runge-Kutta a l g o r i t h m a n d e x p l a i n s the p r i n c i p l e s b e h i n d t h e methods. Al I
t h e Runge-Kutta methods use the simple E u l e r method a s a f i r s t estimate.
Improved estimates a r e then made u s i n g p r e v i o u s estimates and different
time-values within the time interval h. A weighted average of all the
estimates i s used to c a l c u l a t e yn+l. The Fourth-Order Runge-Kutta methods
are the most widely used because of their power and simplicity. The
f o l l o w i n g i s a p a r t i c u l a r Fourth-Order method w h i c h i s commonly used a n d
w h i c h i s i n c l u d e d i n the s i m u l a t i o n p r o g r a m :

= y + I ( kl +2k2+2k3+k4) (4.32)
Yn+l n6

(4.33)

= h f ( t n + i h , yn+gkl ) (4.34)
k2

k3 = hf(t,+ih,yn+$k2) (4.35)

k4 = hf(tn+l,yn+k3) (4.36)
61

Again the problem g i v e n in (4.14) above is solved as an example:


dy/dt=f(t,y)=t+y,y(O)=l. T h i s time y ( 0 . 1 ) i s c a l c u l a t e d in one step (h=0.1)
whereas ~ ( 0 . 1 ) was c a l c u l a t e d i n f i v e time increments (h=0.02) using the
simple a n d modified E u l e r methods above.

kl =h(tn+yn)
=o. 1 (0+1 = 0.10000 (4.37)

k 2 =0.1 (0.05+1 .05) = 0.11000 (4.38)

k 3 =0.1 (0.05+1 .055) = 0.11050 (4.39)

k 4 =0.1(0.10+1.1105 = 0.12105 (4.40)


1 (4.41
y(0.1)=1.000+ -(0.10000+2x0.11000+2x0.11050+0.12105)
6
=1.11034 (4.42)

T h i s agrees to f i v e decimals w i t h the a n a l y t i c a l r e s u l t a n d i l l u s t r a t e s


a f u r t h e r g a i n i n accuracy w i t h less e f f o r t t h a n r e q u i r e d b y the p r e v i o u s
Euler methods. I t s computationally more e f f i c i e n t than the modified Euler
method because, while four e v a l u a t i o n s of the f u n c t i o n a r e r e q u i r e d for
each step r a t h e r than two, the steps can be many-fold l a r g e r for the same
accuracy. The simple E u l e r method would h a v e r e q u i r e d of the o r d e r of 220
steps to achieve five-decimal accuracy in y(O.1) b u t each step involves
only one evaluation of the function. The efficiency of the Euler and
Runge-Kutta methods can be r o u g h l y compared b y c a l c u l a t i n g the number of
function evaluations required for the same o r d e r of accuracy. In this
p a r t i c u l a r example the Runge-Kutta method i s a p p r o x i m a t e l y 50 times more
e f f i c i e n t than the simple Euler method (220/4). The local e r r o r term f o r the
5
Fourth-Order Runge-Kutta algorith (7.35) is O(h ) a n d the g l o b a l e r r o r
4
would be about O(h 1.

M u l t i s t e p Methods

The simple Euler, Modified Euler and Runge-Kutta methods are called
s i n g l e step methods because they use o n l y the i n f o r m a t i o n from the last
step computed. I n t h i s they have the a b i l i t y to perform the next step w i t h
a d i f f e r e n t step size a n d a r e i d e a l f o r b e g i n n i n g the s o l u t i o n where o n l y
the initial conditions are available. The principle behind a multistep
method is to utilize the past values of y and/or yl to construct a
polynomial that approximates the derivative function and to extrapolate
this into the next time interval. Most multistep methods have the
d i s a d v a n t a g e that they use a constant step size h to make the c o n s t r u c t i o n
of the polynomial easier. Another d i s a d v a n t a g e of m u l t i s t e p methods i s t h a t
62

several past p o i n t s a r e required whereas only the initial conditions are


a v a i l a b l e a t the s t a r t . The s t a r t i n g v a l u e s a r e g e n e r a l l y c a l c u l a t e d from
the i n i t i a l conditions using a single-step method such as a Runge-Kutta
method.

F I N I TE ELEMENTS

An imp1 i c i t method involving mass balance across element boundaries


(Connor a n d Brebbia, 1976) i s p o p u l a r i n f i x e d systems b u t has not g a i n e d
much popularity in hydraulic systems owing to changing boundaries
necessitating i t e r a t i v e methods. The steps a r e as follows:

D i v i d e body i n t o elements ( 2 o r 3- dimensional)


Define the nodal unknowns
The flow across an external boundary can be approximated as a
ma themat i c a l f u n c t i o n .

F i g . 4.9 F i n i t e elements

One sets up equations giving balance for each element and solve
simu I taneousl y .
Boundaries for numerical methods

Conditions on a b o u n d a r y may be e i t h e r constant p o t e n t i a l ( o r head o r


water level or cencentration) i.e. flow across boundary, stream1 ines (no
flow across) o r mixed (same f u n c t i o n ) .
One can use pseudo p o i n t s e.g.
hi-, = h. f o r no flow
h. -h. = h . - hi+, f o r flow p e r p e n d i c u l a r to b o u n d a r y .
1-1 I
S i m i l a r schemes may be used f o r concentration d e f i n i t i o n .
63

REFERENCES

Connor, J.J. and B r e b b i a , C.A. 1976. F i n i t e e l e m e n t s f o r f l u i d f l o w .


Newnes-Bu t t e r w o r t h s .
D e i n i n g e r , R.A., 1973. M o d e l s f o r E n v i r o n m e n t a l P o l l u t i o n C o n t r o l . Ann
A r b o r Science.
F r i e d , J.J., 1975. G r o u n d w a t e r P o l l u t i o n , E l s e v i e r .
G e r a l d , C.F., 1980. A p p l i e d N u m e r i c a l A n a l y s i s ; 2 n d Ed. A d d i s o n VJesley.
McDonell, D.M., O ' C o n n e r , B.A., 1977. H y d r a u l i c B e h a v i o u r of E s t u a r i e s .
Macmi I Ian.
Sturnel, F . a n d H a i n e r , K., 1978. I n t r o d u c t i o n to N u m e r i c a l A n a l y s i s ;
S c o t t i s h Academic P r e s s L t d .
Timoshenko, S. and Goodier, J . M . , 1951. T h e o r y of E l a s t i c i t y , McGraw H i l l .
64

CHAPTER 5

MASS BALANCE O F STORMWATER POLLUTANTS

I NTRODUCT ION

Pollution l o a d i n g s from two catchments i n Johannesburg (Green et al.,


1986) were i n v e s t i g a t e d . One, Montgomery Park, is a suburban catchment
a n d t h e other, Hillbrow, a densely built up city area. A comparison of
stormwater r u n o f f and d r y weather flows from b o t h catchments n a r r o w e d
down sources of pollutants and assisted in understanding the washoff
process. I t i s r e p o r t e d that non-point source pollution is responsible f o r
70% of the load i n u r b a n r u n o f f (Wanielista, 1979), and it is largely this
t y p e of contribution which i s detected here. Bradford (1977) attempts to
relate pollutant loads to landuse, and this paper contributes to his
hypothesis. The unpredictability of runoff qua1 i t y indicated by Simpson
a n d Kemp (1982) i s borne out though.

CATCHMENT DESCR I PT ION

F i g . 5.1 Montgomery P a r k Catchment


65

The Montgomery Park catchment is situated 6 km north-west of


Johannesburg and measures 10.53 km2 (1053 ha). The population is
estimated at 15000. The developed area is 75% of the total and the
remainder includes parks, a cemetery and undeveloped area. The
development i s housing and some commercial a n d l i g h t i n d u s t r y . There i s a
s o l i d waste t i p in the catchment f r o m w h i c h seepage occurs. The catchment
is fairly hilly, slopes ranging from 0.02 m/m to 0.15 m/m. The main
drainage system comprises natural and artificial channels (see Figure
5.1). Rainfall over the catchment is recorded at five locations by
a u t o g r a p h i c r a i n gauges. Runoff i s measured a t a g a u g i n g s t a t i o n a t the
catchment outlet i n which the measuring element i s a Crump weir with a
bubble type recorder. Electrical conductivity of the water was recorded
continuously since March 1983.
The Hillbrow catchment measures 67.2 ha and is a fully developed
u r b a n a r e a comprising h i g h - r i s e b u i l d i n g s , some h i g h d e n s i t y housing a n d
a school and is illustrated in F i g u r e 2. The population i s estimated at
12000. There are f o u r r a i n g a u g e s a n d a streamgauge for t h i s catchment.
Both catchments have separate stormwater drainage systems i .e.
seoarate from waste sewerage systems.

F i g . 5.2 H i l l b r o w Catchment
66

Q U A L I T Y OBSERVATIONS

Fa1lout measurement

An attempt was made to assess the level of TDS occurring as


atmospheric f a l l o u t on t h e Montgomery P a r k catchment. After a p e r i o d of 28
d a y s without any r a i n f a l l , the r a i n g a u g e s in the catchment were "washed
down" w i t h d i s t i l l e d water, t h i s water b e i n g collected in a sample bottle.
I t was found t h a t the TDS w i t h i n the f u n n e l s averaged 9.5 mg. Since t h i s
was deposited onto a funnel a r e a of 0.020 m2 i t was deduced that the
equivalent fallout loading on the Montgomery Park catchment was 4.75
kg/ha over 28 days. If washout was omitted this would represent 62
kg/ha/annum. Atmospheric f a l l o u t was collected i n a funnel with an area
of 0.72 mz a t a location n e a r the H i l l b r o w catchment over a p e r i o d of 18
d a y s w i t h no r a i n f a l l . I t was found t h a t the TDS w i t h i n the funnel in t h i s
case was 188 mg resulting in an atmospheric loading rate of 48
kg/ha/annum.
Stormwater r u n o f f qua1 i t y data were collected for selected storms and
analyzed to determine whether relationships could be e s t a b l i s h e d a n d to
o b t a i n the p r o p o r t i o n s of the d i f f e r e n t constituents.
C e r t a i n researchers h a v e observed a c o r r e l a t i o n between the number o f
dry d a y s preceding a storm and the level of p o l l u t i o n of the resulting
runoff (e.g. Sartor et al., 1974; Colwill et al., 1984) while others
maintain that no such relationship exists (e.g. Whipple et al., 1977;
Bedient, 1980).
An attempt was made to see whether the peak concentration of TDS
could be related to the number of antecedent dry days. A regression
a n a l y s i s was performed on a l l the d a t a a v a i l a b l e a n d the best f i t resulted
from a l i n e a r r e l a t i o n s h i p , viz.
C =568 + 68N (5.1 1
P
where C i s the peak TDS concentration in mg/l and N i s the number of
P
antecedent d r y d a y s w th a maximum v a l u e of 5. The c o r r e l a t i o n c o e f f i c i e n t
corresponding i s 0.12 which i s poor. There i s an increase in TDS with
l e n g t h of time between storms i t appears.
In a further tes , TDS was correlated with antecedent moisture
condition classes proposed by Terstriep and Stall (1974) used and the
following relationship emerged.
Cp = 1020 - 163 AMC (5.2)
w i t h a c o r r e l a t i o n coefficient of 0.29. C i s the peak TDS concentration in
P
mg/l a n d AMC i s the antecedent moisture c o n d i t i o n class.
67

Relationship between Total P o l l u t a n t Load and Runoff Volume

A regression analysis was performed on the pollutant load - flow


volume data to determine whether any definite relationship could be
established between these parameters. In all cases the best fit was
obtained from linear approximations with reasonably high correlation
coefficients.
Considering d a t a from Montgomery P a r k alone r e s u l t s in the equation
W = 3395 + 23V (5.3)
with a correlation coefficient of 0.84. W i s -the mass of transported
dissolved solids i n k g a n d V i s the volume of r u n o f f in m’.
With the i n c l u s i o n of the d a t a from H i l l b r o w , the e q u a t i o n becomes
W = 1186 + = 0.27V (5.4)
w i t h a c o r r e l a t i o n coefficient of 0.90.
Treated separately, the relationship between pollution load and low
flow volume i s
W = 3.24 + 0.55V (5.5)
w i t h a coefficient of c o r r e l a t i o n of 0.97

Chem ica I Const ituen t s

The r e s u l t s of t h e chemical analyses of the “ g r a b “ samples collected in


both the H i l l b r o w a n d the Montgomery P a r k catchments a r e l i s t e d i n Tables
1 to 5 .
These results were analyzed to quantify the presence of nitrates,
chlorides and bicarbonates as i t was considered t h a t these were the m a j o r
anions present in the water. The highest anion concentration was
bicarbonate, followed b y sulphates d u r i n g storm r u n o f f a n d c h l o r i d e in d r y
weather conditions. Sulphates a r e predominant i n Johannesburg and could
be wind blown from neighbouring mine waste tips which have high
s u l p h a t e concentrations. Sulphates also reach concentrations over 300 mg/l
i n water supplies f o r the area.
T h e p r o p o r t i o n of n i t r a t e s , sulphates, c h l o r i d e s a n d bicarbonates to the
total dissolved salts is much lower in storm runoff than in the dry
weather flow analyzed. I n the l a t t e r case 68.6% of the TDS consists of
these anions whereas this proportion is as low as 38.8% in the storm
runoff ( a v e r a g e d over both catchments) indicating probable washoff of
constituents t h a t do not n o r m a l l y apear in the d r y weather flow.
68

TABLE 5.1 Results o f chemical a n a l y s e s o n r a i n f a l l a n d r u n o f f


samples f o r H i l l b r o w on 03/01/85

- ~~

I l l OM conduct1V l t Y TDI
taman
ma/. .PI 1
--
2Ohl 1 6.20 14.31 138 1010 0.2 12 5.1 36
20hl4 6.30 13.12 112 242 0.3 10 4.1 31
20118 6.05 11.81 134 160 0.8 10 3.0 36
2ohz3 6.00 9.69 102 770 4.1 10 3.0 24
.?Oh26 5.55 9.91 100 512 8.6 11 5.1 10
2013 1 5.85 10.88 126 232 5.3 13 5.1 24
20150 5.45 13.37 120 110 15.0 16 1.1 10
2lho1 5.90 15.43 170 102 12.9 21 8.2 27

I
*/A 5.55 6.60 18 2.7 4 3.8 6

- 63

TABLE 5.2 Results o f chemical a n a l y s e s on r a i n f a l l a n d r u n o f f


samDles f o r H i l l b r o w o n 18/01/85

&gl* T l l pM ODMYCLlVltY UIICNt.


Yrh taWn
.s/*

1811 i ~ n 2 i 6.35 46.10 346 64 tO.l 6U 37.0 122


18/2 14132 6.35 35.50 265 84 <o. 1 57 30.0 90
18/3 lh138 6.15 2h.50 182 380 to. 1 a3 12.3 85
18/U 14144 6.35 13.40 104 130 a.1 17 10.3 S6
18/5 14hM 6.15 8.88 69 2Oh go.1 13 8.2 20
1016 14h52 5.05 8.23 65 \ u4 0.3 14 7.1 14
1811 1hh54 6.20 6.29 55 92 0.1 14 6.1 12
18/8 lhh57 5.60 7.03 65 56 0.2 12 10.2 7
18/9 15Mo 5.65 6.14 u 8 0.2 11 4.1 10
18/10 15hO4 5.10 5.95 60 (1 0.2 10 6.1 7
5.85 5.92 49 0.2 11 6.1 7
13/11 15hO9
15hl6 5.10 6.58 50
Y1 0.3 10 5.1 1
18/12
R */A b.07 P.ZO 18 41 8.2 6
69

TABLE 5.3 R e s u l t s o f c h e m i c a l a n a l y s e s o n r a i n f a l l and r u n o f f


s a m p l e s f o r M o n t g o m e r y P a r k on 07/03/83

LwIe T i n on CO n a W t l v l t Y roa auspewd ~~tntm Su1ph.t. chiorlam c. c.rbmet.


mark taken so1 id.
mSlm -1 I -1 1 9 1I -1 I -1 1 -1 1

50111 *** 6.25 13.26 104 95 eo. 1 10 6.3 30


sol12 5.85 10.31 86 200 0.1 16 5.2 20
so113 6.00 12.91 112 450 2.2 13 6. 3 31
Ylll
RFl/l*
6.20
1.25
22.30
10.86
166
52
44 1.5
0.4 ..
25
..
16.0
21
16

TABLE 5.4 R e s u l t s o f c h e m i c a l a n a l y s e s on d r y w e a t h e r f l o w s a m p l e s
from Montgomery P a r k

TOI 8usp.na.a
soilas
-11 -1 I

104 10 210.0 100 34 15


1625 4 91.0 15 480 456
314 12 10.0 16 45 175
544 12 11.8 64 101 202
446 10 30.4 kz 86 1@3
320 24 b.0 18 29 not aow
262 14 0.1 25 10 not aow
620 la 2.0 40 120 not row
70

TABLE 5.5 Results of chemical analyses o n d r y weather flow samples


from H i l l b r o w

As one would expect, the concentration of suspended solids in dry


weather flow i s much lower t h a n i n the storm r u n o f f , indicating a higher
t r a n s p o r t r a t e of sediments as well as p o s s i b l e erosion d u r i n g storms. For
the samples analyzed, the suspended solids in the dry weather flow
averaged o n l y 27 mg/l compared w i t h a n average of 236 mg/l for the storm
flows.
Comparing Tables 4 a n d 5 ( d r y weather flows) with Tables 1, 2 and 3
r e v e a l s t h a t the TDS concentrations a r e c o n s i d e r a b l y h i g h e r in d r y weather
flows than in storm flows. The average TDS for the dry weather flow
samples i s 644 mg/l while average values of TDS for the three runoff
events a r e 125 mg/l, 1 1 3 mg/l and 117 m g / l , indicating that the dry
weather flow has about f i v e times as h i g h a concentration as stormwater
runoff. The base load o f TDS from Montgomery Park appears to be largely
from a refuse tip, which averages 160000 kg/annum or 150 kg/ha/annum
averaged over the catchment ( B a l l , 1984).
I t was mentioned that samples of runoff were o b t a i n e d on the rising
l i m b of the h y d r o g r a p h of 18 J a n u a r y 1985 in H i l l b r o w , making i t possible
to detect a flushing effect at the start of the runoff. The high TDS
concentrations a t the e a r l y stages of the runoff, viz. 346 mg/l and 265
mg/l, followed b y a time-dependent decrease in TDS concentration to f i n a l
l e v e l s of about 60 mg/l indicate a "first flush" effect in accordance w i t h
the f i n d i n g s of many o t h e r s (e.g. Cordery, 1977; Helsel et a l . , 1979).
The p r o p o r t i o n s of n i t r J t e s are also much higher in the dry weather
flow than in the stormwater runoff. I n the case of the d r y weather flow
sampled in August and October 1982 (see T a b l e 5. 4 ) the l e v e l s of n i t r a t e
71

mar
PH
(mpnl
18 10

18 160- 14 6.5

14 140- 12 0.0

12 120- 10 5.5

10 100- 8 5.0

21
Fig. 5.3 Plot of p o l l u t a n t c o n c e n t r a t i o n VS. time f o r r a i n f a l l - r u n o f f
event on H i l l b r o w on 03/01/85

45 - \
-
----Suop.
HW+I+WI

-.-
Nllrole
Sulohale
TO9
Solids
-
--.-.
-
Conducllvlly
pn
cc CMorldoo
Flowale -.18

40- 6.1 -.18

35- 8.t -.14

30- 5.e

25 - 5.t

20 -

15-

10-

5-

14h25 30 35 40 45 50 55 15hoo 5 10 15 T h

Fig. 5.4 P l o t of p o l l u t a n t c o n c e n t r a t i o n vs. time f o r r a i n f a l l - r u n o f f


event on H i l l b r o w on 18/01/85
72

a r e so h i g h as to suggest p o s s i b l e blockage o f a s a n i t a r y sewer w i t h the


resulting overflow entering the stream. It was observed for all three
r u n o f f events t h a t the n i t r a t e concentrations increased over the d u r a t i o n of
each hydrograph, reaching their maximum on the recession limbs of the
respective h y d r o g r a p h s . A possible e x p l a n a t i o n f o r t h i s phenomenon i s that
lighting activity w i l l increase the n i t r a t e l e v e l s in t h e r a i n f a l l d u r i n g the
course of the storm, r e s u l t i n g i n i n c r e a s i n g n i t r a t e concentrations in the
r u n o f f w i t h time. There were however large differences in magnitudes of
these concentrations between events. The initial and final nitrate
concentrations from the H i l l b r o w catchment were 0.2 mg/l a n d 12.9 mg/l in
the r u n o f f on 3 J a n u a r y 1985 w h i l e the maximum n i t r a t e concentration in
the r u n o f f on 18 J a n u a r y 1985 d i d not exceed 0.3 mg/l. A maximum n i t r a t e
concentration of 2.2 mg/l was recorded i n the r u n o f f from the Montgomery
P a r k catchment on 7 M a r c h 1983. The recommended n i t r a t e l i m i t i n domestic
water i s 6.0 mg/l w i t h a n upper l i m i t of 10.0 mg/l (SABS, 1984).
There does not apear to be any definite time-related decrease or
increase in the l e v e l s of the o t h e r c o n s t i t u e n t s i n the r u n o f f . For example
s u l p h a t e concentrations increase w i t h time i n the r u n o f f from H i l l b r o w on 3
J a n u a r y 1985 w h i l e the converse i s t r u e f o r the r u n o f f on 18 J a n u a r y 1985
from the same catchment.
Plots of p o l l u t a n t concentrations w i t h time f o r the H i l l b r o w events a r e
presented i n F i g u r e s 5.3 and 5.4.

Mass Balance for event of 18 J a n u a r y 1985 on H i l l b r o w Catchment

A rainfall d e p t h of 6 mm was measured for t h i s event and the TDS


concentration in the r a i n f a l l was 18 mg/l (see Table 2). T h i s can a l s o be
expressed as a rainfall l o a d i n g r a t e of 0.18 kg/ha/mm of rain or 1.08
kg/ha in total. For a catchment size of 67.2 ha this depth of rainfall
corresponds to 4030 rn3 of rainfall over the catchment mass of 73 k g of
pollutants.
For t h i s event a r u n o f f volume of 475 m’ a n d a t o t a l load of 121 k g of
pollutant were estimated. There was thus a net washoff of 48 kg of
p o l l u t a n t from the catchment. E x p r e s s i n g the p o l l u t a n t load i n the runoff
i n terms o f catchment a r e a a n d r a i n f a l l g i v e s 0.30 kg/ha/mm or 1.8 kg/ha
total.
The sources of these p o l l u t a n t s have not been identified, but in a
densely developed a r e a l i k e H i l l b r o w , the most l i k e l y sources a r e washoff
of deposits from wind and motor vehicles and s o l u b l e f r a c t i o n s of litter
w h i c h i s u s u a l l y present.
73

Since the r u n o f f was o n l y 12% of the r a i n f a l l and the catchment still


experienced a net washoff of pollutants, with 66% more pollutant being
washed o f f than was deposited b y the r a i n f a l l , i t i s conceivable t h a t this
washoff may reach even h i g h e r percentages f o r events where the p r o p o r t i o n
of runoff to rainfall is greater. Such events would result from storms
h a a v i n g a greater depth of h i g h e r i n t e n s i t y r a i n f a l l . It i s also p o s s i b l e
t h a t i n p u t d u r i n g one storm i s stored and released a f t e r loss of moisture,
to be washed o f f d u r i n g a subsequent storm.
A p l o t of hydrograph, p o l l u t o g r a p h a n d TDS v a r i a t i o n with time for
t h i s event i s presented i n F i g u r e 5.5.

Mass Balance for event of 7 March 1983 on Montgomery Park Catchment

On 7 March 1983 a t o t a l depth of 14 mm of rainfall was recorded on


the Montgomery P a r k catchment. T h i s event was preceded b y a time p e r i o d
exceeding f i v e days of no rain, so it is not surprising that the TDS
concentration of the rainfall is much higher than that measured in
Hillbrow on 18 October 1985 when only two dry days had passed. The
measured TDS of the r a i n f a l l was 52 mg/l (see Table 5.31, r e s u l t i n g in a
rainfall loading rate of 0.52 kg/ha/mm. The total mass of soluble
p o l l u t a n t s deposited on this 10.53 km’ catchment was thus 7666 kg in
147420 m3 of r a i n f a l l .
A r u n o f f volume of 5508 m3 w i t h a corresponding c u m u l a t i v e r u n o f f load
of 1479 k g of dissolved p o l l u t a n t s was measured. I n terms of r a i n f a l l this
pollutant load can be expressed as 0.10 kg/ha/mm. The runoff volume
represents o n l y 4% of the rainfall and the TDS washed o f f 19% of that
deposited b y the r a i n f a l l . I n t h i s case the catchment therefore experienced
a net gain of 6187 kg of pollutant, or 81% of that deposited. This
corresponds to a net g a i n of 5.87 k g / h a o r 0.42 kg/ha/mm of rain-borne
pollutant i.e. net deposition of pollutant occurred in the peri-urban
catchment w h i l e net washoff occurred i n the densely developed catchment.
Since there i s a deposit ( l o s s of m a t t e r ) from r a i n as indicated b y the
Montgomery P a r k catchment i t can be expected t h a t a s i m i l a r deposit would
occur in H i l l b r o w , so the l i t t e r load must be h i g h e r .

Once again it is difficult to attempt to identify the sources of


p o l l u t a n t s washed o f f t h i s catchment. R e f e r r i n g to Tables 5.2 a n d 5.3 it
w i l l be seen that n i t r a t e levels in the r u n o f f a r e h i g h e r f o r t h i s catchment
than for the Hillbrow catchment on 18 January 1985, signifying the
possible washoff of decaying vegetation, animal faeces and garden
fertilizers. This seems a reasonable deduction as the Montgomery Park
74

,/-•
r--- L*'
--.----.

-
14hOO 16h00
TIME
lEhoO

Fig. 5.5 Hydrograph, p o l l u t o g r a p h a n d TDS f o r H i l l b r o w


f o r event on 18/01/85

I
1,20
- /--'

E 0,90
"
-
E
W
I-
U
d 0,60
s -
s
LL

0,30 -

15h00 17h00 19hOO 2lhoo 23h0


TIME
Fig. 5.6 Hydrograph, p o l l u t o g r a p h a n d TDS f o r Montgomery P a r k
f o r event on 07/03/83
75

catchment consists of p r e d o m i n a n t l y s u b u r b a n r e s i d e n t i a l developments w i t h


gardens. Another source in Montgomery Park c o u l d be leachate from the
ground (either previously deposited by rain seeping in or from soil
minerals). I t i s noted that the p r o p o r t i o n of sulphates a n d carbonates i n
r u n o f f i s s i m i l a r to the r a i n , b u t c h l o r i d e s increase.
It appears that sulphates and chlorides are unaffected by the two
d i f f e r e n t land-uses, the respective levels b e i n g of the same o r d e r f o r b o t h
catchments which also i n d i c a t e s they may be a i r - b o r n e i n t o the catchment.
I t has also been observed that there a r e ( i l l e g a l ) discharges of industrial
wastes i n t o the separate stormwater system in H i l l b r o w .
The hydrograph, pollutograph and TDS variation with time for this
event a r e i l l u s t r a t e d i n F i g u r e 5.6.
I n the mass b a l a n c e of p o l l u t a n t s o u t l i n e d above i t was found p o s s i b l e
i n both the H i l l b r o w and the Montgomery P a r k catchments to relate the
pollutant load in the runoff to the load in the rainfall causing that
runoff. To determine whether the catchment h a s experienced a net loss o r
g a i n of p o l l u t a n t s i t i s also necessary to know the TDS concentration of
the r a i n f a l l as well a s the r u n o f f . In the present p r o j e c t r a i n f a l l q u a l i t y
was o n l y analyzed f o r three events, TDS levels in the r a i n f a l l being 18
mg/l ( H i l l b r o w ) , 52 mg/l (Montgomery P a r k ) a n d 78 mg/l ( H i l l b r o w ) . A TDS
concentration of 118 mg/l i n r a i n f a l l was observed b y Madisha (1983) a t a
location near the H i l l b r o w catchment.
Assuming a r a i n f a l l l o a d i n g r a t e of 0.52 kg/ha/mm f o r Montgomery P a r k
and an average r a i n f a l l l o a d i n g r a t e of 0.71 kg/ha/mm for Hillbrow, the
total weight of dissolved solids deposited on the two catchments was
computed f o r twelve rainfall-runoff events f o r which both discharge and
e l e c t r i c a l c o n d u c t i v i t y d a t a were a v a i l a b l e . These r e s u l t s a r e presented in
Table 5.6.
I t can be deduced from Table 5.6 that the average pollution load of
runoff expressed in terms of rainfall is 0.40 kg/ha/mm of rainfall for
Montgomery P a r k and 1.54 kg/ha/mm of r a i n f a l l f o r H i l l b r o w . This f i n d i n g
i s i n accordance with the f i n d i n g s of other researchers (e.g. Polls and
Lanyon, 1980; Mikalsen, 1984), v i z . that i n general the level of p o l l u t i o n
of stormwater is higher from commercial and downtown land-use
developments than from r e s i d e n t i a l developments.
Another i n t e r e s t i n g deduction from Table 6 i s t h a t more p o l l u t a n t was
deposited on the Montgomery P a r k catchment t h a t was washed o f f f o r five
out of the seven events w h i l e t h i s was o n l y the case f o r two out of five
events in the H i l l b r o w catchment. The h i g h e r percentage imperviousness in
the H i l l b r o w catchment i s p o s s i b l y the reason f o r t h i s phenomenon.
76

TABLE 5.6 Comparison of p o l l u t i o n loads i n r a i n f a l l a n d r u n o f f


w i t h r a i n f a l l depths

-
wation lainfall U l p h t of U i p h t 01 Ratio O f ?Oi I U t l O l

and dapth dap0alt.d 106 In runofr load in


data 106 rumrr I M d to wnorr
rainfall
(-1 Ihal tho1 laad I hg/h./rl
-
mntwuw
Pa~h
01/03/03 14 7666 1479 0.19 0.10
W/I2/01 11 7118 7356 1.03 0.54
12/12/83 17 9309 13086 1.41 0.73
21/01/85
30/10/85
3 I / 10185
46
55
24
25100
30116
13141
23451
15872
7680
0.93
0.53
0.59
0.40
0.27
0.30
-
Avarmpa 10s for
71 -/I
ni I IbIOr

0111 1/85 67 36607 26391 0.72 0.37

nlllb-

13/09/011 1 48 247 5.15 3.60


16/09/04 14 669 217 0.32 0.21
H)/lO/04 2 96 01 0.ou 0.60
21/10/04 1 48 193 4.02 2.07
18/01/85 6 73 121 1.66 0.30

H a v i n g established r e l a t i o n s h i p s between d e p t h o r r a i n f a l I a n d amount


of pollutant washed off a catchment, annual pollutant loads can be
compu ted .
Considering the H i l l b r o w catchment for example and assuming a mean
annual precipitation of 763 mm (Adamson, 1981), the total mass of
p o l l u t a n t s washed o f f t h i s catchment will be of the o r d e r of 80000 k g p e r
annum or 1190 kg/ha/annum. For the Montgomery Park catchment the
amount of a n n u a l p o l l u t a n t loading will be a p p r o x i m a t e l y 320 000 k g o r
305 kg/ha/annum.
Assuming a n average d r y weather flow of 0.0015 m’/s or 130 m’/day in
H i l l b r o w a n d 310 d r y d a y s p e r annum results i n an annual dry weather
flow volume of a p p r o x i m a t e l y 40300 m’. This results in an annual dry
weather p o l l u t a n t load of a p p r o x i m a t e l y 22100 k g o r 330 kg/ha/annum. The
average d r y weather flow i n Montgomery P a r k i s about 0.004 m’/s so the
annual dry weather flow off this catchment is approximately 110000 m’
which corresponds to a total pollutant load of 60500 kg or 57
kg/ha/annum. Therefore i t can be deduced t h a t the a n n u a l p o l l u t a n t load
due to d i r e c t stormwater r u n o f f i s about 3.6 times t h a t due to d r y weather
77

flow for the Hillbrow catchment and about 5.3 times that due to dry
weather flow f o r the Montgomery P a r k catchment.
The pollutant loading rates derived from the different sources are
summarized i n Table 5.7.

TABLE 5.7 Summary of dissolved loads i n kg/ha/mm

CONCLUSIONS

Despite the l i m i t e d monitoring, the f o l l o w i n g tentative conclusions c a n


be drawn.
The t o t a l dissolved p o l l u t i o n load i n stormwater a n d surface drainage
from H i l l b r o w , a densely populated c i t y a r e a i s about 15000 kg/ha/annum
which i s about 3 times as great from a s u b u r b a n catchment, Montgomery
Park. The m a j o r i t y (70%-80%) occurs d u r i n g storm r u n o f f in both cases.
Only about 430 kg/ha/annum f a l l s o r i s washed out of the atmosphere. The
m a j o r i t y i s therefore l i t t e r and from vehicles in the case o f H i l l b r o w , and
decaying vegetable matter o r leachate from Montgomery P a r k .
There i s a net g a i n of p o l l u t a n t s from H i l l b r o w b u t in Montgomery P a r k
the t o t a l washoff i s about the same o r d e r as the t o t a l deposited from the
atmosphere. A s a l a r g e p r o p o r t i o n of r a i n seeps i n t o the ground, i t could
store TDS to be released i n f u t u r e r u n o f f . There i s a net g a i n of n i t r a t e
in Montgomery P a r k however.
Dry weather concentrations are higher in both catchments, due to
seepage from a p o l l u t e d l a n d f i l l in the case of Montgomery Park, (Ball,
1984) and illegal waste discharge in Hillbrow. Concentrations in storm
r u n o f f increase w i t h number of p r e v i o u s d r y days, signifying that street
sweeping would reduce loads.
78

The m a j o r i t y o f d i s s o l v e d s a l t s i s washed o f f d u r i n g the r i s i n g l i m b of


the storms except n i t r a t e s w h i c h e x h i b i t a l a g . Release from the g r o u n d o r
alternativley the i n f l u e n c e of atmospheric l i g h t i n g could b e the cause of
this. Before p r e d i c t i o n b y m o d e l l i n g c a n be u n d e r t a k e n , intensive further
i n v e s t i g a t i o n w i l l be r e q u i r e d .

REFERENCES

Adamson, P.T., 1981. Southern A f r i c a n Storm R a i n f a l l . D i r e c t o r a t e o f Water


A f f a i r s , Department o f Environment A f f a i r s , Technical Report TR 102.
B a l l , J.M., 1984. D e g r a d a t i o n o f g r o u n d a n d s u r f a c e w a t e r q u a l i t y in
r e l a t i o n to a s a n i t a r y l a n d f i l l . MSc(Eng) D i s s e r t a t i o n , U n i v e r s i t y o f t h e
W i twatersrand.
Bedient, P.B., Lambert, J.L. and Springer, N.K., 1980. Stormwater
p o l l u t a n t load-runoff r e l a t i o n s h i p s . Jnl. Water P o l l u t i o n Control Fed.,
52, 2396-2404.
B r a d f o r d , W.J., 1977. U r b a n stormwater p o l l u t a n t l o a d i n g s : a s t a t i s t i c a l
summary t h r o u g h 1972. J n l . Water P o l l u t i o n Control Fed., 49, 613-622.
C o l w i l l , D.M., Peters, C.J. a n d P e r r y , R., 1984. Water q u a l i t y o f motorway
runoff. Transport and Road Research Laboratory, Dept. of the
Environment a n d Dept. o f T r a n s p o r t , TRRL Supplementary Report No.
823.
Cordery, I., 1977. Q u a l i t y c h a r a c t e r i s t i c s o f u r b a n stormwater in Sydney,
A u s t r a l i a . Water Resources Research, 13, 197-202
Green, I.R.A., Stephenson, D. a n d Lambourne, J.J., 1986. Stormwater
p o l l u t i o n a n a l y s i s . U r b a n H y d r o l o g y a n d D r a i n a g e Research Contract,
Water Research Commission Report No. 115/10/86.
Helsel, D.R., Kim, J.I., G r i z z a r d , T.J., R a n d a l l , C.W. a n d Hoehn, R.C.,
1979. L a n d use i n f l u e n c e s o n m e t a l s in storm d r a i n a g e . Jnl. Water
P o l l u t i o n Control Fed., 51, 709-717.
Madisha, J.L., 1983. I n v e s t i g a t i o n p r o j e c t o n u r b a n stormwater p o l l u t i o n in
Braamfontein. Department o f Civil Engineering, University of the
W i twatersrand.
Mikalsen, K.T., 1984. Assessment of water q u a l i t y changes r e s u l t i n g from
urbanization, a g r i c u l t u r e and commercial f o r e s t r y in the s t a t e of
Georgia, U.S.A. Proceedings o f t h e T h i r d I n t . Conf. "Urban Storm
Drainage," Goteborg, Sweden, 801-810.
Pol Is, I . a n d L a n y o n , R., 1980. P o l l u t a n t c o n c e n t r a t i o n s from hogeneous
l a n d uses. Jnl. Environmental Eng. Div., ASCE, 106, 69-80.
S a r t o r , J.D., Boyd, G.B. a n d A g a r d y , F.J., 1974. Water p o l l u t i o n aspects
of street s u r f a c e contaminants. Jnl. Water P o l l u t i o n Control Fed., 46,
458-467.
Simpson, D.E. a n d Kemp, P.H., 1982. Q u a l i t y a n d q u a n t i t y o f stormwater
r u n o f f from a commercial land-use catchment in N a t a l , South A f r i c a .
Water Sci. Tech., 14, 323-38.
South A f r i c a n Bureau o f S t a n d a r d s (SABS), 1984. S p e c i f i c a t i o n f o r w a t e r f o r
domestic supplies. SABS 241.
Stephenson, D. and Green, I.R.A., 1987. Mass b a l a n c e o f stormwater
p o l l u t a n t s . Water S.A.
T e r s t r i e p , M.L. a n d S t a l l , J.B., 1974. The I l l i n o i s u r b a n d r a i n a g e a r e a
s i m u l a t o r , ILLUDAS. I l l i n o i s State Water Survey, U r b a n a , B u l l e t i n 58.
Wanielista, M.P., 1979. Stormwater Management Q u a n t i t y a n d Q u a l i t y . Ann
A r b o r Science P u b l i s h e r s I n c . M i c h i g a n .
Whipple, W., Hunter, J.V. a n d Yu, S.L., 1977. E f f e c t s o f storm f r e q u e n c y
on p o l l u t i o n from u r b a n r u n o f f . Jnl. Water P o l l u t i o n Control Fed., 49,
2243-2248.
79

CHAPTER 6

OPT I MUM ALLOCAT ION OF WATER RESOURCES SUBJECT TO QUAL I TY CONSTRA I NTS

I NTROOUCT ION

We a r e r e a c h i n g an age of compromise. The 1960's h a i l e d the e r a of


economic benefit cost analysis, the 1970's saw the appearance of the
environmentalists and i d e a l i s t s a n d the 1980's appear to be p r o d u c i n g more
realistic planners. Multiple objectives including economic, sociological,
p o l i t i c a l and environmental w i l l be considered b u t h o p e f u l l y i n the correct
perspectives. High i d e a l s can o n l y r e s u l t i n slow-down of growth and t h i s
may h a v e detrimental effects on development of underdeveloped countries.
The w a t e r i n g down of e n g i n e e r i n g projects to meet high i d e a l s can also
stagnate the engineering industry and lose v a l u a b l e b r a i n power to o t h e r
profess ions .
Water resources a r e regarded b y many as a never d i m i n i s h i n g asset. On
account of annual replenishment it is assumed the resource cannot be
mined. This i s a fallacy, for apart from over-exploitation and drainage
basin deterioration, the n a t u r e of the resource can be a l t e r e d . As more
and more usage occurs so there will be g r e a t e r waste water discharges
and poorer q u a l i t y water i n our r i v e r s . New g r o w t h can o n l y be met from
these r i v e r s o r from water f u r t h e r a f i e l d i f surface waters a r e to be r e l i e d
on. We can often not a f f o r d the l u x u r y of p u r e mountain waters p i p e d from
many hundreds of kilometres away. It will be necessary to purify waste
water to acceptable s t a n d a r d s i n some cases. The cost of d e m i n e r a l i z a t i o n
and n u t r i e n t removal i s particularly high. T h i s cost may not be w a r r a n t e d
f o r a l l uses. I n many countries p o t a b l e water i s t r a n s p o r t e d separately or
obtained from containers while poorer quality water i s used for general
domestic and i n d u s t r i a l purposes. Although separate piped water supplies
of different quality water will be expensive there may be some areas
which a r e predominately h i g h density residential and could j u s t i f y high
quality water. Other areas r e q u i r i n g lower q u a l i t y c o u l d receive separate
supplies. This is particularly the case i n m i n i n g areas in South Africa
where these studies were i n i t i a t e d .
P a r a l l e l studies a r e i n v e s t i g a t i n g the cost o f d e m i n e r a l i z a t i o n and h i g h
q u a l i t y p u r i f i c a t i o n b u t that i s o n l y one of the options. The others a r e to
seek f r e s h surface o r groundwater resources f u r t h e r away, to make do w i t h
poorer q u a l i t y of local resources o r to a l l o c a t e i n an o p t i m a l manner as
i n d i c a t e d here.
80

Methods of research c o u l d e i t h e r adopt the g l o b a l systems approached o r


a more s i m p l i s t i c b u t p e r h a p s easier understood methodology. There a r e a
number of sophisticated techniques f o r o p t i m i z a t i o n of linear and nonlinear
water systems subject to various constraints. The use of computers is
readily, in fact sometimes too r e a d i l y , adopted b y eager students. Whereas
these methods may form ideal subject matter for dissertations the o u t p u t
from a computer program i s not easy to e x p l a i n to r e g i o n a l p l a n n e r s a n d
politicians. Simple g r a p h i c a l displays or tabular results a r e much easier
to describe a n d present. Simple h a n d c a l c u l a t i o n s o f t e n enable the a n a l y s t
to follow the alternatives and bear in m i n d marginal costs or multiple
objectives. By f o l l o w i n g the e n t i r e p l a n n i n g process t h r o u g h the a n a l y s t is
also able to c l a r i f y a l t e r n a t i v e objectives a n d allocate priorities. It is
t h i s approach which i s adopted i n the s i m p l i s t i c s t u d y below (Stephenson,
1982).

THE SYSTEM

Consider the t r a n s p o r t a t i o n problem depicted i n Fig. 6.1. A number of


sources of water a r e a v a i l a b l e (A, B a n d C ) a n d they each have limited
resources indicated as 10, 20 and 15 megalitres per day (Ml/day),
r e s p e c t i v e l y . T h e t o t a l a v a i l a b i l i t y may exceed the requirements of demand
of users W, X, and Y though, which r e q u i r e 8, 12 a n d 16 M l / d a y in t h i s
example. The cost of transport a l o n g each r o u t e i s i n d i c a t e d i n F i g . 6.1,
a n d a g a i n i n the t r a n s p o r t a t i o n t a b l e a u Table 6 . 2 . A s f a r as i t has been
described the system i s a simple t r a n s p o r t a t i o n example w h i c h could e a s i l y
be optimized, i.e. the flow a l o n g each r o u t e to r e s u l t i n a minimum t o t a l
t r a n s p o r t a t i o n system c o u l d be d e r i v e d r e l a t i v e l y e a s i l y .

F i g . 6.1 Supply requirements a n d a l t e r n a t i v e s


81

The s i t u a t i o n i s complicated b y the Fact t h a t the i n d i v i d u a l consumers


have c e r t a i n water q u a l i t y requirements. The measurements of the r e l e v a n t
i m p u r i t y , e.g. TDS ( t o t a l dissolved s o l i d s ) i s in mg/l a n d the requirements
of W, X and Y are that the TDS s h a l l not exceed 10, 11 and 8 mg/l
respectively. Note t h a t Ml/day m u l t i p l i e d b y mg/l g i v e s k g / d a y of s a l t s , a
mass flow rate. The TDS of the source waters from A, B a n d C a r e 6, 11
and 8 mg/l respectively.
T h e lower l i m i t on TDS may be achieved by selecting correct sources,
b l e n d i n g d i f f e r e n t sources o r , i f economic, p u r i f y i n g p a r t o r a l l of a n y of
the resources. The l a t t e r option, namely p u r i f i c a t i o n , could be h a n d l e d b y
assuming any source i s p u r i f i e d and adding the cost to the conveyance
cost.
Often the r e l a t i o n s h i p between cost of p u r i f i c a t i o n a n d r a t e of flow is
nonlinear such as w i t h desalination be reverse osmosis and the system
becomes more complex. I n such case, separable programming methods a r e
possible (Stephenson, 1978). Alternatively a gradient method may be
employed to seek a n optimum. I f only par t ( a v a r i a b l e p a r t ) of a source
need be p u r i f i e d , the d e s c r i p t i v e equations a r e more numerous b u t linear
programming methods may be employed to optimize the system.
I n the present example (Fig.6.1) the resource a n d demand c o n s t r a i n t s
may also be w r i t t e n a s l i n e a r c o n s t r a i n t s :

Demand
8
aAW + aBW + aCW=
Q~~ + aBX + aCX = 12
aAy + aBy + aCY = 16

The q u a l i t y c o n s t r a i n t s may be w r i t t e n :
6QAW + l l Q B W + 8QCW 1 1 0 x 8
6QAx + llQBx + 8QCX 511 x 12,
6QAy + llQBy + 8QBx 5 8 x 16

Provided there i s a feasible solution the set of m + 2n c o n s t r a i n t s c o u l d


be analyzed b y l i n e a r programming methods. m i s the number of sources
and n the number of demand p o i n t s . Alternatively the availability and
82

demand c o n s t r a i n t s c o u l d be considered i n a t r a n s p o r t a t i o n m a t r i x a n d the


qua1 i t y c o n s t r a i n t s h a n d l e d separately u s i n g t h e p r i n c i p l e of decomposition
of l i n e a r programmes ( D a n t z i g , 1963; Stephenson, 1969). A t h i r d method i s
described below. This is based on the transportation method with
additional c o n s t r a i n t s on the d i s t r i b u t i o n s . The r e s u l t i n g advantages o v e r
the l i n e a r programming a l t e r n a t i v e s a r e s i m p l i c i t y , rapidity, no necessity
for computers a n d more interaction between the water resources planner
a n d the system.
It is assumed the reader is familiar with linear programming and
t r a n s p o r t a t ion programming techniques.

SOLUTION METHOD

The d a t a a r e a r r a n g e d in a t a b l e a u s i m i l a r to a transportation tableau


(Table 6.1). Each demand is represented by a row, including a row
l a b e l led "slack" since resources exceed demand here. A column represents
each source and there i s an additional column labelled "artificial slack"
since the i n i t i a l assignment may not s a t i s f y q u a l i t y c o n s t r a i n t s w i t h o u t it.
I n f a c t , a s there a r e three a d d i t i o n a l c o n s t r a i n t s , one would expect u p to
three a d d i t i o n a l v a r i a b l e s in the f i n a l programmme. The cost coefficients
of the artificial slack flow variables should be very large, but not
necessarily so f o r s a l t mass flow s l a c k s since they are of the type. It
is not necessary to assign artificial cost coefficients in the following
method.

TABLE 6.1 Transportation matrix with shuffling to eliminate artificial


slack

Y
83

An initial assignment i s made i n Table 6.1 u s i n g the Northwest corner


r u l e (Loomba, 1964). At each assignment, two types of c o n s t r a i n t must be
satisfied - water flow and salt balance. Thus i n most blocks, e.g. AW,
flow l i m i t s the number, but i n b l o c k CY, TDS b a l a n c e l i m i t s flow to 2.25
Ml/day ( a flow of 6 would otherwise have been assigned to t h i s b l o c k ) .
The f i r s t step a f t e r m a k i n g the i n i t i a l assignment should be to evacuate
flows from the artificial slack column. Note each re-assignment must
s a t i s f y flow c o n s t r a i n t s and TDS limits. Observe that the water flows in
Ml/day a r e w r i t t e n i n the bottom l e f t of each b l o c k followed b y / a n d the
TDS flow in k g / d a y . Thus 8/48 indicates 8 Ml/day at a source TDS o f 6
mg/l r e s u l t i n g in TDS flow of 48 k g / d a y . I t i s r e l a t i v e l y easy to check at
each corner of a closed circuit whether water or TDS limit the
re-allocation, and select the lowest p e r m i s s i b l e flow allocation. I n Table
6.1, two re-allocations are necessary to evacuate block LY. a feasible
(non o p t i m a l ) solution r e s u l t s .
Now the o p t i m i z a t i o n proceeds a s f o r any t r a n s p o r t a t i o n exercise, except
for the additional constraint on each re-al location. After calculating
column and row cost coefficient and comparing implied costs in each
vacant block w i t h a c t u a l costs it i s decided to re-allocate to block BY.
Although flow consideration would l i m i t the a l l o c a t i o n to 1 M l / d a y , quality
c o n s t r a i n t s l i m i t i t to 0.4 M l / d a y . Then a l l the slack i n TDS f o r row Y is
eliminated. Table 6 . 2 r e s u l t s .

TABLE 6.2 T r a n s p o r t a t i o n m a t r i x step two


04

I t w i l l be observed t h a t there i s more than one p o s s i b l e cost coefficient


f o r some blocks, depending on which block i s used as a p i v o t . This arises
because the number of occupied blocks i s g r e a t e r t h a n n + m - 1 where n
a n d m a r e the number of rows and columns i n the t a b l e a u excluding the
a r t i f i c i a l slack column. Each possible combination should be i n v e s t i g a t e d .
Where the p i v o t sequence AW, AX, B X , B Y , BS, CY i s used, the maximum
difference between i m p l i e d cost a n d a c t u a l cost c o e f f i c i e n t appears i n block
C X , and i s 8 v s 4 ( T a b l e 6.3). It will be found t h a t b y proceeding f i r s t
a r o u n d the closed p a t h CX-AX-AY-CY-CX a n d then CX-EX-BY-CY-CX that f i r s t
1.4 and then 0.9 ml/day can be a l l o c a t e d to block CX without violating
flow and q u a l i t y c o n s t r a i n t s .

TABLE 6. 3 T r a n s p o r t a t i o n m a t r i x optimized

Dunand :

Subsequent calculations will reveal there is no further cost


improvement, i.e. no more i m p l i e d costs exceed a c t u a l costs once new cost
coefficients a r e ca Icu I a ted .
The optimum p l a n , w h i c h s a t i s f i e s q u a l i t y requirements, i s indicated in
F i g . 6.2.
85

D iscussionl

All water users do not require the same high quality water. Where
poorer quality is tolerable, allocation of alternative sources may be
considered. O v e r a l l economy of p u r i f i c a t i o n a n d d i s t r i b u t i o n r e s u l t .

F i g . 6.2 Optimum a l l o c a t i o n subject to c o n s t r a i n t s

A technique f o r a l l o c a t i n g water resources b y a form of transportation


programming has been demonstrated with an example. The technique is
simple and not computer orientated. The resulting distribution system is
depicted a n d can r e a d i l y be updated as q u a l i t i e s of the sources v a r y .
T h e method i s therefore of use f o r management and o p e r a t i o n of water
distribution systems as well as design. In fact even more so, since
construction costs a r e not as a r u l e e a s i l y l i n e a r i z e d whereas pumping
costs a r e g e n e r a l l y p r o p o r t i o n a l t o the r a t e of flow.

L I NEAR PROGRAMM I NG SOLUT ION

The p r e v i o u s sections high1 ighted the shortcomings of the t r a n s p o r t a t i o n


and transportation extended techniques. The techniques require that
onerous simp1 i f y i n g assumptions be made about parameters. They thus o f f e r
an approach which do not model the d i s t r i b u t i o n comprehensively enough.
The most severe shortcomings are:
a ) Qua1i t y constraints cannot be considered in Transportation
Programming.
b ) Optimization of the amount of water to be desalinated and hence
blended cannot be achieved in e i t h e r of the T r a n s p o r t a t i o n techniques.
86

c) Non-Linear cost f u n c t i o n s h a v e to be a p p r o x i m a t e d b y l i n e a r functions


in both techniques.
Linear programming techniques provide means whereby al I these
shortcomings can be overcome. In linear programming both quality and
q u a n t i t y c o n s t r a i n t s , a n d any o t h e r l i n e a r c o n s t r a i n t s , can be m a n i p u l a t e d
to yield an optimum solution. However, neither a non-linear objective
f u n c t i o n n o r c o n s t r a i n t s c a n be used unless they a r e converted to a linear
or piecewise linear form. This can be achieved by using linear
programming in c o n j u n c t i o n w i t h separable programming.
This section describes how the d i s t r i b u t i o n problem i s transformed into
a representative mathematical model suitable for analysis using linear
programming (Grosman, 1981).

The sets of d a t a r e q u i r e d a r e summarized below;

Qua1i ty Quantity
Sources :-
Water Board 500 mg/P 100.0 Me/d
Groundwater 600 mg/e 11.5 Me/d
Wastewater 1750 mg/l Variable
Desalinated wastewater 175 mg/P Variable
Demands:-
Transfer 1750 mg/e 7.0 Me/d
System 1 700 mg/e 9.5 Me/d
System 2 700 mg/e 0.7 Me/d
System 3 700 mg/e 0.5 Me/d
Waste ( S l a c k ) 1750 mg/P unused

The water b o a r d s u p p l y i s assumed to be 100 Me/d. I n r e l a t i o n to the


o t h e r sources, this is high, a n d consequently o n l y a p o r t i o n thereof may
b e used. The p o r t i o n to be used w i l l b e optimized.
Wastewater y i e l d s 10 Me/d, of which a v a r i a b l e p o r t i o n i s d e s a l i n a t e d
to y i e l d a n improved qua1 i t y available from the d e s a l i n a t e d wastewater.
This v a r i a b l e portion i s an unknown and hence it should be optimized.
The recovery r a t i o n of feed flow to product flow in a d e s a l i n a t i o n p l a n t is
0,69 f o r t h i s case. Expressing the above i n mathematical terms,
U + D /0.69 = 10

U + 1.45 D = 10 (6.10)
where U = Used MSW i n Me/d
D = Desalinated Wastewater in MP/d
87

Two new variables (U and D) and a new constraint Eq. 6.10 are
i n t r o d u c e d to c a t e r f o r t h e d e s a l i n a t i o n of a v a r i a b l e p e r c e n t a g e o f waste.
U was assumed t o b e 7.5 Mk‘/d, h e n c e f r o m Eq. 6.10, D w a s 1.73 Me/d. The
m a g n i t u d e o f t h e s l a c k a l l o c a t i o n t o w a s t e (W) w i l l c o n s e q u e n t l y v a r y . The
v a r i a t i o n i s a c c o r d i n g to E q . 6.11 where the sources a r e b a l a n c e d a g a i n s t
the demands:
W + 7.0 + 9.5 + 0.7 + 0.5 = 100 + 11.5 + U + D
but D = (10 - U ) 0.69 f r o m Eq. 6.10 (6.11)
hence W = 93.8 + U + ( 1 0 - U ) 0.69
W = 100.7 + 0.31 U (6.12)
From a T r a n s p o r t a t i o n E x t e n d e d a n a l y s i s U w a s 7.5 Me/d, a n d t h e 100
Mt/d was o n l y 1 MP/d, hence W was 4.03 MP/d. However, U varies now
with a maximum value of 10 Me/d, (that is without desalination).
T h e r e f o r e from Eq. 6.12
W L 103.8 Me/d (6.13)
The a c c e p t a b l e q u a l i t y assumed f o r t h e System 1 (S), System 2 ( V ) a n d
System 3 ( M ) i s s t i l l 700 m g / Q . T h e n e x t s e c t i o n r e v i e w s a n a n a l y s i s of a
r a n g e of acceptable q u a l i t i e s .
The only unrealistic assumption necessary in this analysis, using
linear programming, i s that the total costs are linearly r e l a t e d to feed
flow. The cost coefficients are summarized below in Table 6.4.
Ab b r e via tio n s f o r the sources a n d demands a r e a l s o i n d i c a te d .

TABLE 6.4 Cost c o e f f i c i e n t s (c/m’) u s e d in l i n e a r p r o g r a m m e

DEMANDS

I WASTE I TRANSFER I SYSTEM 1 I SYSTEM 2 1 SYSTEM 3


(c/m’

WB
1
R
w l O / S
0.0 0.0 24.0
I
32.5
v l M 39.5
-_
GROUND
cn WATER F 5.0 2.5 3.0 11.5 14.0
W
u WASTE
K
3 WATER u 7.5 6.0 0.0 1 .o 0.0
SI
DESAL.
WASTE D 44.0 43.0 38.0 46.0 51 .O
WATER
88

The f o r m u l a t i o n of t h e m a t h e m a t i c a l model b e g i n s b y e x p r e s s i n g the


o b j e c t i v e f u n c t i o n i n terms of t h e e x p r e s s i o n

n
z = c c. x (6.14)
J J
Hence

Z = O.O(RW) + O.O(RB) + 24.O(RS) + 32.5(RV) + 39.5(RM) + 5.0(FW) +


2.5(FB) + 3.0(FS) + 11.5(FV) + 14.O(FM) + 7.5(UW) + 6.0(UB) + O.O(US)
+ l.O(UV) + O.O(UM) + 44.O(DW) + 43.O(DB) + 38.O(DS) + 46.O(DV) +
51.5(DM) + O.O(U) + O.O(D)
(6.15)
where t h e terms in brackets represent the f l o w of water from a specific
source ( f i r s t l e t t e r ) to a s p e c i f i c demand (second l e t t e r ) a n d c o n s t i t u t e t h e
unknowns. The term U and D r e p r e s e n t the yields of the wastewater and
Desalinated wastewater respectively. Consequently they have zero cost
c o e f f i c i e n t s a n d a r e a l s o unknown.
The o b j e c t i v e function 2 must b e m i n i m i z e d s u b j e c t to the following
l i n e a r constraints.

Source C o n s t r a i n t s :-
RW + RB + RS t R V t RM 5 100.0 (6.16)
FW + FB + FS + FV + FM = 11.5 (6.17)
UW + UB + US + UV + UM = U (6.18)
DW + DB + DS + DV + DM = D (6.19)

Demand C o n s t r a i n t s : -
R W + FW + UW + DW 103.8 ( f r o m 6.14) (6.20)
RB + FB + UB + DB = 7.0 (6.21)
RS + FS + US + DS = 9.5 (6.22)
R V + FV + UV + DV = 0.7 (6.23)
RM + FM + UM + DM = 0.5 (6.24)

Quality Constraints:-
500 R W + 600 + FW + 1750 UW + 175 DW I 103.8 (1750) (6.25)
500 RB + 600 + FB + 1750 UB + 175 DB 5 7.0 (1750) (6.26)
500 RS + 600 + FS + 1750 US + 175 DS 5 9.5 ( 700) (6.27)
500 RV + 600 + FV + 1750 UV + 175 DV 5 0.7 ( 700) (6.28)
500 RM + 600 + FM + 1750 UM + 175 DM 5 0.5 ( 700) (6.29)
89

B I end i n g Constraints:-
U + 1.45 D = 10 ( from 6 . 1 0 ) ( 6 .3 0 )
Non-nega t i v i t y constra i n ts:-
A l l unknowns 0 (6.31 )
It is necessary to convert Eq. 6.18 and 6.19 to a form with no
unknowns on the r i g h t h a n d side. Hence:
UW + UB + US + UV + UM - U = 0 (6.32)
DW + DB + DS + DV + DM - D = 0 (6.33)
The set of constraints (Eq. 6.16, 6.17 and 6.20 to 6.33 can be
optimized, subject to the objective f u n c t i o n of Eq. 6.15, u s i n g the manual
Simplex Technique. Since the r e s u l t a n t m a t r i x i s r a t h e r l a r g e , a computer
programme to solve I i n e a r programming problems b y the Simplex Technique
i s preferred. The f o l l o w i n g i s noted :

a ) D i a g r a m m a t i c a l l y the solutions i s as i n d i c a t e d i n F i g . 6 . 3
b ) The water obtained from the water board is not required, and is
consequently a l l o c a t e d to waste as slack. This is interpreted to mean
that i s i s not necessary to o b t a i n water from the water b o a r d .
c ) Most of the ground water i s allocated to system 1 and most of the
wastewater i s a l l o c a t e d to the t r a n s f e r .

I WASTE J
1750mg/L

11 .5MP/d /
/f

Fig. 6.3 Optimal solution to the distribution problems using linear


programming
90

The amount of water d r a w n from the d e s a l i n a t e d wastewater source is


zero. T h i s implies t h a t no d e s a l i n a t i o n i s required. The importance o f
being able to optimize this variable is enhanced by comparing the
change in the values of the o b j e c t i v e f u n c t i o n .
The v a l u e of the o b j e c t i v e function i s now $1043/day, a decrease of
$453/day compared with the Transportation Programming Extended
Solution, solely as a result of allowing an unknown variable
desal i n a t ion f r a c t i o n to be opt imized.
The optimum solution was reached, since a l l the shadow v a l u e s of the
unknowns were n e g a t i v e o r zero. The lowest increase in the o b j e c t i v e
function would occur if a unit of water was allocated from the
groundwater to waste. The cost would increase by only 1 c/m' or
$lO/day/unit of water. The highest increase would be $336/day/unit of
water, i f water was a l l o c a t e d from the d e s a l i n a t e d wastewater to the
transfer. Table 6.5 summarizes the shadow values in ascending o r d e r .

TABLE 6.5 Shadow v a l u e s of empty c e l l s u s i n g l i n e a r programming.

SOURCE DEMAND SHADOW VALUE


( $ / d a y /un i t
-
Groundwater Waste 10
Water Board Transfer 15
Desal. waste System 2 225
Water Board System 2 238
Desal. Ww System 3 242
Water Board System 1 244
Desa I. WW System 1 257
Water Board System 3 280
Desa I. WW Waste 331
Desa I. WW Transfer 336

g ) The assumption that cost i s l i n e a r l y r e l a t e d to feed flow is clearly not


correct. Hence the r e s u l t s o b t a i n e d a r e not a ' t r u e ' optimum. Initially,
the expected feed flow r a t e s from each source to each demand were
assumed, and hence a cost was derived. The inaccuracies of the
assumptions ( a n d consequently of the cost coefficients) are borne o u t
b y the comparisons i n Table 6.6. I t may, however, be a r g u e d t h a t the
overestimates cancel the underestimates.
91

TABLE 6.6 Errors incurred by using linear cost functions instead of


non-l i n e a r cost functions.

SOURCE DEMAND ASSUMED FLOWS OPT I MUM FLOWS PERCENTAGE


AND CORRESPONDING FROM FIG. 6.3 DIFFERENCE
COSTS AND CORRESPONDING IN COSTS
COSTS

FLOW COST FLOW COST


(Ml/d) (c/m’) (Ml/d) (c/m’) (c/m3 1

Ground-
water System 1 3.0 3.0 8.7 1.6 47%
Waste-
water Waste 2.0 7.5 3.8 5.2 31 %

Waste-
water Transfer 2.0 6.0 5.3 4.5 25%
Ground-
water Transfer 5.0 2.5 1.7 3.1 -24%

THE LINEAR PROGRAMM ING TECHN IQUE WITH SEPARABLE PROGRAMM I NG


APPL I ED

The assumption of a linear cost function, and hence of objective


function, has been made throughout the discussion on transportation
programming, transportation programming extended and Iinear
programming, a n d the a p p l i c a t i o n of these techniques. I t has consistently
been mentioned that the assumption is onerous by the nature of the
cost-flow g r a p h s and t h a t in fact the technique of s e p a r a b l e programming
could be employed to avoid this assumption. Separable Programming
approximates n o n - l i n e a r functions b y piecewise l i n e a r approximations. The
accuracy depends on the d e v i a t i o n of the linear a p p r o x i m a t i o n from the
curve. For non convex separable functions, as i n t h i s case, the technique
does not guarantee a g l o b a l optimum.
This section employs the separable programming technique, in
conjunction w i t h l i n e a r programming, to p r o v i d e a n o p t i m a l solution. The
mathematical model of the system i s described in the next section, except
f o r the objective function.
92

T h e model i s thus expressed as below:-


Minimize the o b j e c t i v e f u n c t i o n Z subject to the l i n e a r o r piecewise linear
constraints
Z = CRW f CRB t CRS f CRV f CRM
+ CFW f CFB + CFS f CFV f CFM
+ CUW f CUB + CUS f CUV + CUM
f CDW t CDB t CDS + CDV + CDM (6.34)
where CXY represents the total cost of supplying water from source X to
demand Y in dollars/day, and represents a functional equation of
separable programming. DJXY represents the J t h increment D or S f o r the
XY combination.
CRW = 0
CRB = 0
CRS = 117 DORS + 107 D l R S + 223 D2RS + 565 D3RS f 1658 D4RS
CRV = 124 DORV + 108 D l R V t 227 D2RV + 565 D3RV f 1670 D4RV
CRM = 117 DORM + 107 D l R M + 223 DZRM + 565 D3RM f 1658 D4RM
CFW = 37 DOFW + 25 D l F W f 52 D2FW f 113 D3FW + 351 D4FW
CFB = 13 DOFB + 19 D l F B + 4 0 D2FB + 80 D3FB + 156 D4FB
CFS = 42 DOFS f 20 D l F S + 24 D2FS + 29 D3FS + 89 D4FS
CFV = 4 9 DOFV t 22 D l F V t 26 D2FV + 30 D3FV + 101 D4FV
CFM = 42 DOFM + 20 D l F M + 24 D2FM +
f 29 D3FM
89 D4FM
CUW = 6 2 DOUW + 38 DlUW + 64 D2UW f 131 D3UW + 407 D4UW
CUB = 42 DOUB + 34 D l U B + 5 7 DZUB + 109 D3UB + 236 D4UB
cus = 0
CUV = 4 DOUV + 1 DlUV f 2 D2UV f 3 D3UV + 6 D4UV
CUM = 0
CDW = 251 DODW + 229 DlDW + 583 D2DW + 1594 D2DW + 4595 D4DW
CDB = 231 DODB + 227 D l D B + 574 DZDB + 1572 D3DB + 4424 D4DB
CDS = 189 DODS + 193 D l D S + 517 D2DS + 1464 D3DS + 4287 D4DS
CDV = 193 DODV + 194 D l D V +
+ 1466 D3DV + 4194 D4DV
519 DZDV
CDM = 189 DODM + 193 D l D M + 517 D2DM f 1464 D3DM + 4287 D4DM

(6.35)
subject to source c o n s t r a i n t s 6.16, 6.17, 6.32 a n d 6.33,
demand c o n s t r a i n t s 6.20 to 6.24,
qua1 i ty c o n s t r a i n t s 6.25 to 6.29,
b I end in g c o n s t r a i n t s 6.30,
U + 1.45 D = 10 (6.30)
Non-nega t i v i t y c o n s t r a i n t s ,
adjacent c o n s t r a i n t s f o r separable v a r i a b l e s : -
93

I f a n y DJXY i s non-zero, all t h e preceding DJXY values must take on


the v a l u e 1 , and a l l the succeeding DJXY values must take the v a l u e 0.
where
RS = 0.3 DORS + 0.6 DlRS + 1.6 D2RS + 4.5 D3RS + 13 D4RS
R V = 0.3 DORV t 0.6 DlRV + 1.6 DZRV + 4.5 D3RV + 13 D4RV
RM = 0.3 DORM + 0.6 DlRM + 1.6 DZRM + 4.5 D3RM + 13 D4RN
FW = 0.3 DOFW + 0.6 DlFW + 1.6 DZFW + 4.5 D3FW + 13 D4FW
FB = 0.3 DOFB + 0.6 D l F B + 1.6 D2FB + 4.5 D3FB + 13 D4FB
FS = 0.3 DOFS + 0.6 D l F S + 1.6 D2FS + 4.5 D3FS + 13 D4FS
FV = 0.3 DOFV + 0.6 D l F V + 1.6 DZFV .t 4.5 D3FV + 13 D4FV
FM = 0.3 DOFM + 0.6 D l F M + 1.6 DPFM + 4.5 D3FM + 13 D4FM
UW = 0.3 DOUW + 0.6 DlUW + 1.6 DZUW + 4.5 D3UW + 13 D4UW
UB = 0.3 DOUB + 0.6 D l U B + 1.6 DZUB + 4.5 D3UB + 13 D4UB
UV = 0.3 DOUV + 0.6 DlUV + 1.6 DZUV + 4.5 D3UV + 13 D4UV
DW = 0.3 DODW + 0.6 DlDW + 1.6 DZDW + 4.5 D3DW + 13 D4DW
DB = 0.3 DODB t 0.6 D l D B + 1.6 DZDB + 4.5 D3DB + 13 D4DB
DS = 0.3 DODS + 0.6 DlDS + 1.6 DZDS + 4.5 D3DS + 13 D4DS
DV = 0.3 DODV + 0.6 DlDV + 1.6 DZDV + 4.5 D3DV + 13 D4DV
DM = 0.3 DODM + 0.6 DlDM + 1.6 DZDM + 4.5 D3DM + 13 D4DM
(6.36)
and where XY represents the unknown q u a n t i t y of water i n MP/d, supplied
from source X to demand Y. These set of equations i n 6.36 represents the
g r i d equa t ions of sepa r a b I e programm i n g .
S u b s t i t u t i n g the set of equations of 6.35 i n t o 6.34 a n d r e t a i n i n g the set
of equations of 6.36 a s independent equations, the system i s solved u s i n g
linear programming in conjunction with separable programming. The
solution was obtained using the IBM Mathematical Programming System
Extended/370 (MPSX/370) Software Package.
I n o r d e r to ensure the solution obtained i s close to a global optimum
(as opposed to a local optimum) i t i s necessary to complete two computer
runs. The f i r s t , with the control programme, the second with the line
XSETLB = -1 after the l i n e BCDOUT in the control programme. Separable
programming f o r non-convex separable functions, as in t h i s case, does not
guarantee a g l o b a l optimum. The reader i s r e f e r r e d to the MPSX/370 IBM
Program Reference Manual (1976) f o r f u r t h e r information. The s a l i e n t r e s u l t s
and conclusions apear below:
a ) Diagrammatically t h e solution i s i n d i c a t e d below i n F i g . 6.4.
b ) The 100 MP/d taken from the water board is in fact not used. If
lOMP/d is assumed to have been taken from the board, it is all
returned. Hence the s u p p l y from the b o a r d i s also optimized.
94

F i g . 6.4 Optimal solution using l i n e a r programming in conjunction with


separable programming

Most of the ground water i s allocated to system 1 and most of the


wastewater i s transferred.
The amount of water d r a w n from d e s a l i n a t e d wastewater i s zero. This
implies that no d e s a l i n a t i o n i s required. The m a i n reason f o r this is
the abundance of f a i r l y good q u a l i t y g r o u n d water w i t h a TDS of 600
mg/e. I n t h i s case the acceptable s t a n d a r d of 700 mg/e i s only just
above the q u a l i t y of the groundwater s u p p l y .
The v a l u e of the o b j e c t i v e function is now $691/day, a decrease of
$352/day compared with the linear programming solution. This comes
about solely as a r e s u l t of the i n t r o d u c t i o n of a more r e p r e s e n t a t i v e
cost f u n c t i o n , u s i n g separable programming.
Since the o b j e c t i v e f u n c t i o n i s of the non-convex type, the solution
obtained is not necessarily the global optimum. When using the
95

XSETLB=-1 command ( s e a r c h i n g from lower bound to u p p e r bound) the


v a l u e of the o b j e c t i v e f u n c t i o n was $714/d. When d e l e t i n g t h i s command
( s e a r c h i n g from upper bound to lower b o u n d ) the v a l u e was reduced to
$691/d. Fig. 6.5 i n d i c a t e s the $691/d is still a local optimum. The
g l o b a l optimum i s around $650/d.
g ) There is no difference in allocation between this solution, and the
solution u s i n g l i n e a r programming ( w i t h o u t a n o n - l i n e a r cost f u n c t i o n ) .
h ) The accuracy of the solution i s w i t h i n 5% - 10% of the true global
optimum. It would be improved by a refinement of the grid and
f u n c t i o n a l equations w i t h i n the v i c i n i t y of the c u r r e n t a l l o c a t i o n s .

S e n s i t i v i t y Study f o r v a r i o u s acceptable TDS v a l u e s

The case studies presented were based on a n acceptable Total Dissolved


Solids ( T D S ) of 700 mg/P for water used on the System 1, System 2 a n d
System 3. It i s necessary to examine the optimal solutions for various
possible acceptable TDS values i n o r d e r to e s t a b l i s h the best q u a l i t y . This
would lead to the establishment of an optimum TDS value, in terms of total
combined costs, r e q u i r e d a t the demand zones mentioned above.
The r a n g e of acceptable TDS values examined here i s between 500 mg/O
and 1200 m g / P i n discrete steps of 50 mg/O from 600 mg/P upwards, and in
closer steps between 500 mg/O and 600 mg/e. The only modification
r e q u i r e d to the mathematical model i s the s u b s t i t u t i o n of the r e l e v a n t TDS
values f o r 700 mg/P i n Eq. 6.27 to Eq. 6.29. Results a r e also given for
the e x i s t i n g system, i g n o r i n g a l l q u a l i t y aspects.
Table 6.7 presents the optimal a l l o c a t i o n of water of water from each
source to each demand, the amount of wastewater used a n d the amount of
desalinated wastewater. I t also i n d i c a t e s whether the s o l u t i o n i s a g l o b a l
or local optimum, and shows the value of the objective function for
v a r i o u s selected TDS values.
Fig. 6.5 indicates graphically the variation in the total cost of
procuring, desalinating and distributing for various acceptable TDS
values.
From Table 6.7 a n d F i g . 6.5 the f o l l o w i n g conclusions a r e a p p a r e n t :
a ) The g r a p h i s c h a r a c t e r i z e d b y two p a r t s : one w i t h acceptable q u a l i t i e s
g r e a t e r than 600 mg/e a n d the other w i t h q u a l i t i e s less t h a n 600 mg/O.
The former has a f l a t slope of 0.2, i n d i c a t i n g small cost decreases f o r
l a r g e TDS increases. The latter has a slope of 8.5, displaying the
reverse tendency.
96

TABLE 6.7 Comparison of optimal solutions f o r v a r i o u s TDS v a l u e s

RW
RB 0.651 4.339 3.409 7.000
RS
RV
RM
FW 3.318 4.522 3.800
FB 2.059 0.800 1.730 2.661 3;591 7,000
FS 7.265 8.382 9.500 8.674 7.848 7.022 6.196 0.700
FV 0.535 0.618 0.700 0.639 0.578 0.517 0.467
FM 0.382 0.441 0.500 0.457 0.413 0.370 0.326
uw 3.326 3.800 3.800 8.139 7.209 6.278
UB 6.349 4.941 6.200 5.270
us 0.826 1.651 2.478 3.304 8.800
uv 0.061 0.122 0.183 0.243 0.700
UM 0.043 0.087 0.130 0.174 0.500
DW
DB
DS 2.235 1.118
DV 0.165 0.082
DM 0.118 0.059
U 6.349 8.175 10.000 10.000 10.000 10 .ooo 10.000
D 2.518 1.259

TYPE G L O B A L G L O B A L G L O B A L LOCAL GLOBAL GLOBAL LOCAL GLOBAL

OBJ.
FN .1573 1213 698 691 633 61 2 64 9 359
(R/d)
97

I1
--1

FIG. 6.5 V A R I A T I O N I N T O T A L COST FOR


V A R I O U S A C C E P T A B L E TOS V A L U E S

LLL
1 6
u
7 800 900 1000 1100 1200
TOTAL D I S S O L V E D S O L I D S (mg/e)
The a b r u p t change in slope a t 600 mg/e i s a r e s u l t of the necessity f o r
p r o c u r i n g Desalinated wastewater for lower acceptable TDS values. As
water of improved q u a l i t y i s required, the amount of water requiring
d e s a l i n a t i o n incrases, a n d consequently the a l l o c a t i o n s to System 1, 2
and 3 increase.
The p r e v i o u s argument v a l i d a t e s the selection of a quality within the
reach, slightly to the r i g h t of 600 m g / t . The exact quality would be
determined a f t e r a g r a p h was d r a w n .
Each successive increase i n acceptable TDS should cause a decrease in
the t o t a l least-cost.
Not a l l the successive TDS increases manifest a decreased cost f o r the
p r e v i o u s TDS. This occurs as a result of the non-convex separable
o b j e c t i v e function. However, on inspection of Fig. 6.5, it is evident
t h a t 700 mg/4, 1000 mg/e a n d 1100 mg/e a r e in f a c t local optima, since
they do not follow the expected trend. Consequently the allocations
produced i n Table 6.7 f o r these TDS v a l u e s a r e also not true optima.
These problems may be overcome be p e r f o r m i n g a sensitivity analysis,
o r p o s s i b l y b y r e v i s i n g the g r i d a n d f u n c t i o n a l equations.
The absolute lowest cost occurs when all quality aspects a r e ignored
completely. The resulting cost is $359/d, a decrease of $34/d in
comparison w i t h the s o l u t i o n u s i n g t r a n s p o r t a t ion programming a lone.
Trends of increases, decreases a n d changes i n a l l o c a t i o n s as the TDS
v a r i e s a r e e v i d e n t from Table 6.6. Two t y p i c a l forms are:
1) An increase in allocation from 500 mg/Q to 600 mg/e and a
decrease thereafter - Groundwater to System 1 .
2) No allocation until after 600 mg/e and a steady increase
thereafter - Wastewater to System 2 .
Most of the groundwater is allocated to System 1, and most of the
wastewater t r a n s f e r r e d a t low TDS values d i s c h a r a g e d to waste.

REFERENCES

Dantzig, G.B., 1963. L i n e a r Programming a n d Extensions. P r i n c e t o n U n i v .


Press, Princeton.
Grosman, D.D., 1981. Optimum a l l o c a t i o n of mine s e r v i c e water subject to
q u a l i t y c o n s t r a i n t s . C i v i l Eng. in S.A.
Lcomba, N.P., 1964. L i n e a r Programming. McGraw-Hill, NY.
Stephenson, D., 1969. Optimum allocation of water resources by
mathematical programming. J . H y d r o l . 9, 20-33.
Stephenson, D., 1978. Optimum p l a n n i n g of r e g i o n a l waste water treatment.
I n : Modelling the Water Q u a l i t y of the H y d r o l o g i c a l Cycle (Proc. Baden
Symp., September 1978), 351-360. IAHS P u b l . No. 125.
Stephenson, D., 1982. Optimum a l l o c a t i o n o f water resources subject t o
q u a l i t y c o n s t r a i n t s . Proc. Exeter Symp. IAHS, P u b l i c . 135, 299-305
99

CHAPTER 7

ECONOMICS OF DESALINATION OF WASTEWATERS

I NTRODUCT ION

M u n i c i p a l wastewater i s generally treated to remove suspended matter


and to n e u t r a l i s e the b i o l o g i c a l activity. It i s disinfected and rendered
innocuous before b e i n g discharged i n t o streams. The m i n e r a l content of the
water is, however, not affected noticeably by treatment either at the
wastewater treatment works a n d a t the water p u r i f i c a t i o n works. Thus not
o n l y i s there a m i n e r a l b u i l d - u p in the water due to i n d u s t r i a l p o l l u t i o n
and to some extent domestic pollution, but also this mineral content is
contributed to by natural sources. Thus in addition to the nitrates,
phosphates and other nutrient m i n e r a l s coming from the residential type
areas, we also h a v e n a t u r a l m i n e r a l s such as calcium a n d s u l p h a t e b e i n g
contributed to the system from stormwater runoff. The total mass of
dissolved solids thus discharged by wastewater works into the rivers
averages m i l l i o n s of tons p e r day.
The magnitude of the problem of removing the dissolved m i n e r a l s i n the
water i s enormous. There a r e many options open, however, for optimum
reuse of t h i s wastewater. Some of the p o s s i b i l i t i e s a r e suggested below:

ALTERNATIVES FOR OPTIMAL REUSE OF WASTE WATER

Present p o l i c y for many affected water s u p p l i e s i s e f f e c t i v e l y to d i l u t e


p a r t l y treated and r e t u r n e d wastewaters w i t h f r e s h water from r i v e r s a n d
other upstream sources. Provided that the water quality is at an
acceptable limit, for example 500 mg/e per litre total dissolved solids
according to world health organisation standards, then there is little
concern. I n o r d e r to achieve t h i s d i l u t i o n , i t may be necessary in future
to discharge some of the wastewater downstream where other users will
have s i m i l a r o r more concentrated problems. Alternative to t h i s i s the use
of f u r t h e r sources of f r e s h water (Stephenson a n d Corbetis, 1984).
I t may be more p r u d e n t to adopt more o p e r a t i n g i n t e n s i v e schemes a n d
less c a p i t a l intensive schemes i n the l i g h t of economic risks involved in
capital i n t e n s i v e water supply schemes. ,In p a r t i c u l a r where conjunctive
use i s thought of then high capital cost schemes should be used on a
steady base load s u p p l y b a s i s whereas o p e r a t i n g i n t e n s i v e schemes would
g e n e r a l l y be reserved f o r times of drought i n s u r f a c e resources which a r e
100

Clarl-
flcatloi

Grit

500

Sand
50

10

1
0.1

Silt
0,Ol
SAND FILTERS
Clay

0.00

EVAPORATION

Collolc

R E V E R S E OSMOSIS

,001 -

0001 I I
1 10 100 1000 10000 too

F i g . 7.1 Selection of p u r i f i c a t i o n method based on water q u a l i t y


101

capital intensive. Demineralisation and desalination processes are more


operative intensive than surface resources development. Desalination
procedures i.e. those suitable for desalination of sea water are often
highly operating intensive as they use large amounts of power, for
example d i s t i l l a t i o n processes. On the o t h e r h a n d the t o t a l d i s s o l v e d s o l i d s
content of sea water i s n e a r l y 35 000 mg/t per litre, whereas we are
t a l k i n g of dissolved s o l i d s contents of less than 1 000 mg/P per Iitre in
wastewaters f o r a r t i f i c i a l recharge of g r o u n d water a q u i f e r s .
The p o s s i b i l i t y of l i m i t e d treatment before d i s c h a r g i n g into aquifers is
now under consideration. It is possible that by trickling the water
t h r o u g h the a q u i f e r s there w i l l b e n a t u r a l a e r a t i o n which would reduce the
b i o l o g i c a l oxygen demand as well as p r o v i d e a degree of f i l t r a t i o n a n d of
great interest, natural ion exchange resulting in demineralisation or
n e u t r a l i s a t i o n of some of the dissolved s o l i d s content. Alternatively, the
wastewaters c o u l d be discharged to lower levels o f the a c q u i f e r thereby
l i f t i n g the fresher waters which have seeped there b y n a t u r a l means such
as from r a i n w a t e r and i n f i l t r a t i o n from surface streams.
Not o n l y will this t y p e of artificial recharge have the a d v a n t a g e of
r e d u c i n g pumping costs b y keeping a h i g h water table, but it may also
solve the problem of dewatering of dolomitic compartments which is linked
to geotechnical problems. Previous dewatering exercises have r e s u l t e d in
collapse and dire consequences in residential areas and at mining
development so p a r t i e s concerned would be nervous about d e w a t e r i n g even
i f o n l y i n t e r m i t t e n t l y to s u p p l y i n times of drought.
Other possibilities include the local recycling of wastewater from
particular areas to other selected areas. In this way there is a
possibility of minimal treatment if water is used for successive lower
q u a l i t y - r e q u i r i n g uses. Yet another p o s s i b i l i t y i s the r e c y c l i n g w i t h f r e s h
water pumped from r i v e r s instead of the n a t u r a l recycling. In this way
pumping costs a n d p i p i n g costs, as well as storage costs, would be saved
f o r the water would be recycled a n d not have to be pumped.

SELECT ION OF OPT I MUM DESAL I NAT ION METHODS

Although h i g h d e s a l i n a t i o n costs a r e a deterrent to the general use of


desalination for water supply, an optimised system may in fact be
considerably more economic than may first appear. The location, scale,
type and a d a p t a b i l i t y of a d e s a l i n a t i o n o r d e m i n e r a l i s a t i o n p l a n t can a l l
be p u t to use i n r e d u c i n g t o t a l water costs. Thus the location o f in-house
d e s a l i n a t i o n p l a n t may a v o i d the necessity of d i s p o s i n g of effluents into
102

collecting sewers, then through municipal wastewater treatment works.


Distribution and pumping costs are thereby partly reduced, w h i c h offset
d e s a l i n a t i o n costs.
Studies f o r optimum location of d e s a l i n a t i o n p l a n t s have been conducted
with the assistance of computer simulation programmes. The depiction of
the reticulation system and alternative locations for desalination plants
p l u s a n a l y s i s h a v e p r o v e d t h a t such p l a n t s c a n be economically installed.
There may be s a v i n g s in pumping costs in particular for high heads if
plants are underground. There is also the saving in purchase of raw
water and piping costs. Desalination has also been shown to give a
c o n s i d e r a b l y b e t t e r water q u a l i t y than the use of river water. This has
further imp1 i c a t i o n s in reducing costs of corrosion and deterioration to
pipework due to other chemical a c t i v i t i e s such as s c a l i n g .
The method of d e s a l i n a t i o n may also depend to a large extent on the
cost of energy. Whereas d e s a l i n a t i o n methods such as e v a p o r a t i o n require
l a r g e amounts of energy and a r e therefore n o r m a l l y u n d e r t a k e n u s i n g coal
fired boilers, low energy consumption systems such as reverse osmosis
f r e q u e n t l y use e l e c t r i c i t y from the g r i d systems. There a r e many industries
which have some form of heat exchange or heat generation. Thus
oil-from-coal systems generate l a r g e amounts of s u r p l u s heat and may be
more suitable for evaporation type methods. Some industries require
cooling. Some mines c h i l l water before sending it underground and have
even contemplated the distribution of ice underground. In such cases
freezing desalination may prove most viable. The vapour compression
method i s also r e c e i v i n g close a t t e n t i o n i n t e r n a t i o n a l l y a t the moment.
The scale of the d e s a l i n a t i o n p l a n t a n d the q u a l i t y of the raw water
h a v e a considerable effect on the optimum method of d e s a l i n a t i o n . Scale o r
size of p l a n t w i l l i n f l u e n c e the o p e r a t i n g costs a n d these can be expected
to reduce the l a r g e r p l a n t a n d c a p i t a l costs such a s housing w i l l reduce
p e r k i l o l i t r e of water t r e a t e d the l a r g e r the p l a n t .
For low total dissolved sol i d s contents membrane-type processes such a s
reverse osmosis a n d e l e c t r o d i a l y s i s h a v e p r o v e d most economical. For very
low concentrations ion exchange i s economical. There a r e many problems
associated w i t h the membrane t y p e processes in p a r t i c u l a r where there i s a
h i g h s u l p h a t e content. I n such cases seeded systems h a v e often p r e v e n t e d
the c r y s t a l i s a t i o n of the sulphates on the membrane a n d kept the total
dissolved s o l i d s in suspension. The number o f stages in such p l a n t s cai
a l s o effect the f i n a l water q u a l i t y . The cost per ton of dissolved solid
removed may be a minimum f o r one p a r t i c u l a r method whereas the cost pel
k i l o l i t r e treated, i f one i s not concerned w i t h the amount of s a l t s removed
103

may be cheaper f o r another system. F i g u r e s 7.7 a n d 7.8 compare the costs


on these b a s i s f o r d i f f e r e n t methods.
Multi-stage demineralisation may also be suitable for particular
applications. Thus if a high water quality is required, multi-stage
methods are usually the most efficient and the final effluent can be
brought to a low total dissolved solids concentration, for example, less
than 10 m i l l i g r a m s per Iitre, most economically t h r o u g h successive stages.
It is possible that succession stages employ different techniques e.g.
reverse osmosis coupled w i t h ion exchange.
The dispossl of the brine is also a problem when it comes to
demineralisation of waste waters. Whereas the volume of concentration of
the b r i n e i s of l i t t l e concern in the case of desalination p l a n t s on the
sea, this i s not the case where the brine has to be disposed of. It is
most d e s i r a b l e that the b r i n e be b r o u g h t to a v e r y h i g h concentration and
even possibly to solids form before disposal inland. This minimises the
cost of transport and the cost of disposal sites if it is stored. In such
cases the b r i n e may have to be concentrated t h r o u g h successive stages of
the p l a n t . These concentration stages would b e designed d i f f e r e n t l y to the
e f f l u e n t p u r i f i c a t i o n stages a n d may work on a d i f f e r e n t process a g a i n .
Heat exchange p l a y s a n important part in the o p e r a t i n g costs of many
processes. Thus i f e v a p o r a t i o n techniques a r e used then the e f f l u e n t which
is a t a h i g h temperature can be used to heat the incoming stream to b r i n g
it to nearly evaporation temperature such as in multi-stage flash
evaporation methods. In the case of freezing processes the ice product
could be used to cool the incoming stream of water to near freezing
temperature. T h i s would be possible if the f i n a l p r o d u c t temperature were
immaterial but i t may not be wise i f c o l d water i s a product a s well as
p u r e water.
I t i s thus evident that desalination and d e m i n e r a l i s a t i o n cannot easily
be bought in the form of a package plant. The economics and
practicability of the d e m i n e r a l i s a t i o n process can only be selected when
considering the p l a n t and a l l factors a s a whole. The e n t i r e process must
be designed in conjunction with the plant of the factory if optimum
desalination costs are to be achieved. In this case effective costs less
t h a n 50 cents p e r k i l o l i t r e , comparable w i t h r a w water, can be achieved.

RELEVANT D E S A L I N A T I O N METHODS

The p o t e n t i a l for desalination is internationally recognized and as a


r e s u l t there was a n increase i n water d e s a l i n a t i o n c a p a c i t y of 40% d u r i n g
104

the l a s t 5 years (65% of which a r e m u l t i - s t a g e flash distillation seawater


p l a n t s a n d 25% reverse osmosis seawater a n d b r a c k i s h water p l a n t s ) .

Membrane methods appear promising for future development. Reverse


osmosis was restricted to small size p l a n t s for brackish water until 10
y e a r ago a n d now moving to l a r g e scale plants. Some examples are the
250000 m’/d p l a n t in Jeddah f o r sea water treatment a n d the 400000 m’/day
p l a n t i n Yuma ( U S A ) , f o r d e s a l i n a t i o n of d r a i n a g e .
Membrane methods are competitive with thermal processes, comparing
favourably i n energy consumption as well as susceptibility to corrosion
and scaling. Where brine disposal is a problem however (for example
i n l a n d ) the cost of a d d i t i o n a l f a c i l i t i e s may d e t r a c t from the membrane
processes.

I n d u s t r i a l Wastewater treatment

The increased energy costs during the last decade have directed
research and development work for al I desalination methods towards
r e d u c i n g energy consumption. Different methods of energy recovery have
been investigated and their applicability depends on the costs and the
size of the p l a n t (Binnies, 1981; Larson, 1979).
Some examples a r e :
1) RO p l a n t energy can be recovered b y installing a turbine on line in
the ( h i g h p r e s s u r e ) b r i n e stream.
2) Underground installation of RO plant can be justified based on
u t i l i s a t i o n of the s t a t i c pressure instead of h i g h p r e s s u r e pumps.
T h i s can be a p p l i c a b l e to the m i n i n g industry f o r f r e s h water production
underground. The energy consumption costs w i l I n o r m a l l y be h i g h .
In the USA it has been suggested that advanced treatment methods
(demineral i z a t i o n ) for domestic and municipal wastewater is the best
a l t e r n a t i v e f o r s o l v i n g the problems of water supply. This is applied iri
Denver (USA) where the RO for demineralization is included in a single
plant, a n d i s now b e i n g contemplated elsewhere.

Reverse Osmosis

T h e n a t u r a l phenomenon of osmosis occurs when salt water and fresh


water a r e separated b y a semi-permeable membrane a n d f r e s h water flows
through the membrane to dilute the saline water. This water flow stops
when equilibrium i s established and the p r e s s u r e d i f f e r e n c e between the
two solutions i s c a l l e d osmotic pressure, the m a g n i t u d e of which depends
105

on the s a l i n e solution concentration. If however, pressure i s exerted in


the s a l t solution greater than osmotic, fresh water diffuses through the
membrane free of s a l t (DSS, 1980; L u d w i g , 1980).

Membrane Description

The cellulose acetate membrane which is currently in general use is


approximately 1 0 0 ~ thick, of which only one layer is active and is
approximately 0,2p thick (2000A) on top of the membrane surface w i t h the
r e s t a c t i n g as p h y s i c a l support f o r the exerted pressure. This t h i n layer
acts as a f i l t e r to r e t a i n the ions such as Na' and Cl-.

E Iectrod ia I y s i s

I n the E l e c t r o d i a l y s i s process, water flows between alternately placed


c a t i o n and a n i o n permeable membranes.
A direct electric current i s the driving force for the ion migration
through the membranes. A series of a l t e r n a t i v e c a t i o n a n d a n i o n membranes
w i t h a p l a s t i c spacer between i s assembled i n t o membrane stacks. Several
hundred membranes and their separating spacers are usually assembled
between a s i n g l e set of electrodes to form a membrane stack.
The ion selection membranes a r e b a s i c a l l y ion exchange r e s i n s i n sheet
form w i t h s e l e c t i v i t i e s g r e a t e r than 90%. Normally E l e c t r o d i a l y s i s systems
consists of one to s i x stages, w i t h removal p e r stage v a r y i n g from 30 to
60% ( n o r m a l l y 50%).
Energy consumption i s based on F a r a d a y ' s Law, according to which f o r
100 mg/e removal of dissolved ionised solids from 5m3 of water, 200
amper-hours ( D C ) a r e r e q u i r e d w i t h voltages 1 - 2 V. Therefore about 0,3
kWh i s needed f o r 5m' of water treated in addition to which 2kWh is
r e q u i r e d f o r pumping.
Several hundred Electrodialysis plants have been i n s t a l led
internationally for process water treatment or portable water from feed
waters n o r m a l l y of less t h a n 3000 mg/t salinity.
Reverse polarity (electrodialysis reversal) configuration has been
introduced commercially to reduce polarization and scaling in the
membranes.

I o n Exchange

The ion exchange process has been used f o r many years for softening
106

iter

ter

'1 10 100 1000 10 000 100 1


Product Solinity mg/l

Fig. 7.2 S u i t a b l e feed s a l i n i t i e s a n d product s a l i n i t y


f o r v a r i o u s d e s a l i n a t i o n processes. Recovery
r a t i o s also i n d i c a t e d
107

of water and d e m i n e r a l i z a t i o n f o r v a r i o u s i n d u s t r i a l uses. It i s normally


r e s t r i c t e d to waters of not more than 1000 mg/4 t o t a l dissolved solids.
The process i s based on the c h a r a c t e r i s t i c of some neutral minerals
called zeolites which were found to exchange ions suitable for the
softening of water, l i k e exchange of Mg' a n d Ca
++ f o r Na
+. Artificial ion
exchange r e s i n s have s u p e r i o r exchange c h a r a c t e r i s t i c s a n d be d e f i n i t i o n
are insoluable solids containing fixed cations or anions capable of
r e v e r s i b l e exchange with mobile ions of the opposite sign in solutions.
Resins normally absorb Na+ ions and other cations and release H+ or
absorbs CP- and other anions and release OH-; these are called acid
r e s i n s and base r e s i n s respectively. The ions release H
+ a n d OH- in the
solution which combine to form H20.
Ion exchange is unlikely to prove economical for water treatment of
s a l i n i t y higher than 1000 mg/4. However, it can b e used in c o n j u n c t i o n
w i t h a membrane process.

COST ANALYSIS

The cost of desalination techniques i s often expressed in terms of


cents/m' of water produced. This approach can be m i s l e a d i n g as it fails
to take i n t o account a l l the v a r i a b l e s a f f e c t i n g t h e cost s t r u c t u r e . Before
proceeding w i t h cost estimation the f o l l o w i n g parameters h a v e to be f i x e d :

1) Product requirement, p l a n t load f a c t o r a n d recovery.


The v a r i a b l e s affect the q u a n t i t y a n d q u a l i t y of the f i n a l product, the
r e l a t i o n s h i p between product water quantity, feed water quantity and
the plant operational efficiency. Values assumed f o r the cost estimate
i n t h i s p a p e r a r e ; recovery r a t i o (Rc) 70% - 65% a n d p l a n t load f a c t o r
90%.
2 ) Rate of interest which for present purposes is taken as 10% w i t h a
redemp t ion p e r i o d of 20 years.
3) P l a n t l i f e i s taken a t 30 years.
The c a p i t a l and r u n n i n g costs a r e affected b y the above parameters.

C a p i t a l Costs

Capital costs include Civil Engineering and plant as well as site


development (roads, electricity and water etc. 1. They also include
equipment and controls as well as installation of intake and brine
d i sposa I .
108

I n d i r e c t C a p i t a l Costs

10 - 12% long term interest r a t e i s included d u r i n g p l a n t construction


a n d also labour costs w h i c h amount to 5 - 6% of the t o t a l c a p i t a l costs.

R u n n i n g Costs

Running costs a r e g e n e r a l l y d i r e c t l y p r o p o r t i o n a l to p r o d u c t t h r o u g h p u t
and include energy costs, chemical costs, labour for operation and
ma i n tenance, membrane r e p lacemen t , o p e r a t i n g a n d maintenance costs.
For membrane plants it is reasonable to assume electricity cost with
100% load factor of 2c/kWh. Chemical cost and treatment costs vary with
feed water characteristics, the process used and the plant's recovery
ratio.

L a b o u r Costs

These depend on the requirements o f the p l a n t w i t h respect to o p e r a t i o n


and control. It also depends on the reliability of the plant, as this
affects i t s maintenance l a b o u r cost. Costs g i v e n here a r e f o r p l a n t s u p to
10 000m3/day c a p a c i t y (medium s i z e ) .

Membrane Replacement

For E l e c t r o d i a l y s i s (ED) 20% of capital cost is spent on membranes


which on the average have a 7 year life. For reverse osmosis treating
b r a c k i s h water the l i f e of the membrane i s o n l y 3 years.
In F i g u r e 7.3 the capital cost for both Electrodialysis and Reverse
Osmosis include the site development costs, equipment costs and indirect
capital costs. Figure 7.4 shows the operating costs (running costs)
i n c l u d i n g l a b o u r , energy a n d costs f o r both membrane processes. This data
i s based upon a feed water w i t h a s a l i n i t y in the r a n g e of 1 000 t o 3 000
mg/e t o t a l dissolved sol ids.
The capital cost for the reverse osmosis process varies from
approximately $700/m3/day for a 120rn3/day feed capacity plant to
$170/m3/day f o r a 10 000m3/day feed c a p a c i t y plant, based upon 3 stages.
The cost a l l o w s f o r d i f f e r e n t equipment designs and manufacturer's prices
a n d s i t e costs which a r e dependent on t h e s i t e .
The c a p i t a l cost for E l e c t r o d i a l y s i s depends l a r g e l y on the s a l i n i t y of
the feedwater hence the number of stages involved i n the process a s well
109

D E S R L I N R T I O N PROCESSES
CFTXTRL L RUNNING co- ro11 nLcmtornmysxs L IINLRSL osnos~s

c
8

Fig

n
\

I
Fig. 7.4 Running cost versus P l a n t size (Feed) f o r ED & RO

ED process ( -.-.d
RO process ( )
I10

D E S A L I N A T I O N PROCESSES
ENERGY R f O U I R M W S (hl4hk) FOR CLCCTRODIRLYSIS h REVERSE OSMOSIS

Fig. 7.5 Energy Demand versus Feed Salinity for ED & RO

F i g . 7.6 Energy Demand versus P l a n t size (Prod) for ED & RO

ED process (-.-.- )

RO. p r o c c s d )
111

as the p l a n t size. The costs v a r y from $760/m3/day f o r two 4-stage plant


of 120m3/day feed capacity to a $200/m3/day for a 2 stage plant of
10000m3/day feed c a p a c i t y . I t can be seen i n F i g u r e 7.3 t h a t c a p i t a l costs
f o r E l e c t r o d i a l y s i s a r e g e n e r a l l y h i g h e r than other methods.
In running costs however electrodialysis is cheaper than reverse
osmosis, varying from 25c to 7c as opposed to 45c - 1Oc for reverse
osmosis from the same p l a n t c a p a c i t y ( F i g u r e 7.5). The costs in Figures
7.3 and 7.4 account for different water characteristics and p l a n t design
factors.
The water temperature, pH, turbidity, suspended matter etc. can
influence energy and pretreatment costs. Also recovery ratio (Rc) v a r i e s
between 70 a n d 85% f o r E l e c t r o d i a l y s i s a n d 65 a n d 85% f o r reverse osmosis
which can affect the energy requirement.
The t o t a l theoretical cost of the water v a r i e s from 40c to 20c per m3
for different size p l a n t s f o r E l e c t r o d i a l y s i s and 52c to 22c p e r m3 i n the
case of reverse osmosis (1983 f i g u r e s ) .
In Figure 7.5, it can be seen that Electrodialysis is less
energy-consuming f o r a feed s a l i n i t y r a n g e of 500 - 3 000 mg/Q. This is
due to the fact that energy demand i s d i r e c t l y p r o p o r t i o n a l to removal f o r
the E l e c t r o d i a l y s i s . The energy costs account for different plant sizes,
assumed of the o r d e r of 1 000m3/day,
F i g u r e 7.6 shows the v a r i a t i o n in energy requirements f o r d i f f e r e n t size
p l a n t s . Energy f o r reverse osmosis changes w i t h recovery v a r y i n g from 70%
to 85%. The e l e c t r o d i a l y s i s c u r v e assumes a recovery r a t i o of 75%.

CONCLUSIONS

It is evident that there are many variables affecting the costs of


demineralization. From the point of view of the applicability of a
p a r t i c u l a r process and the optimum process f o r a p a r t i c u l a r a p p l i c a t i o n
t h e f o l l o w i n g factors a r e r e l e v a n t :

Load factor - r e l a t e s c a p i t a l to r u n n i n g costs.


Locality - affects power cost, construction a n d l a b o u r costs.
C a p i t a l cost a n d interest r a t e .
Power and other o p e r a t i n g costs.
TDS of r a w water.
Q u a l i t y requirements f o r e f f l u e n t
Method of b r i n e disposal.
112

I f the costs a r e expressed a s cents p e r k i l o l i t r e , one p a r t i c u l a r method


may p r o v e most economic, whereas i f they a r e expressed a s cents p e r kg
of dissolved s o l i d s removed, another method may be optimal.
Methods r e l a t i v e l y i n s e n s i t i v e to the TDS o f the incoming steam include
the thermal methods. In the case of e v a p o r a t i o n methods, however, scaling
i s d i r e c t l y r e l a t e d to t h e TDS. There i s a tendency to use thermal methods
f o r h i g h TDS waters as the cost per k g of salt removed is low. On the
other h a n d the cost of membrane and ion exchange processes are more
r e l a t e d to the amounts o f s a l t removed and they are therefore lowest for
low TDS waters. F i g u r e 7.7 and 7.8 depict the costs of different method
p l o t t e d ( a ) versusflow r a t e a n d ( b ) versus r a t e of TDS removal.
It appears that membrane type processes are most economical for
industrial wastwaters (provided adequate pre-treatment is practical).
Recent advances i n membranes now i n c l u d e those c a p a b l e of passing high
rates at low pressures p r o v i d e d o n l y limited solids removed i s required.
Using present d a y p r i c e s i t i s f e a s i b l e t h a t costs w i l l be competitive w i t h
b u l k water supplies from a f a r and in addition there is the i n c e n t i v e of
reduced r e t i c u l a t i o n costs, a n d the p o s s i b i l i t i e s of use of the e f f l u e n t s f o r
specific a p p l i c a t i o n s o r f o r g r o u n d water recharge.

REFERENCES

B i n n i e a n d P a r t n e r s , 1981. D e s a l i n a t i o n methods a n d costs, Water Systems


Research Programme, U n i v e r s i t y of the Witwatersrand.
DSS Engineers Inc., 1980. Data Collection a n d A n a l y s i s of Commercial
Membrane D e s a l i n a t i o n P l a n t s .
Larson T.J. and L e i t n e r G., 1979. D e s a l i n a t i o n of sea water a n d b r a c k i s h
water, A cost update. D e s a l i n a t i o n , 30, p525-539.
L u d w i g , L., 1980. Reverse osmosis i n the d e s a l i n a t i o n of b r a c k i s h water
a n d sea water, Desalination, 36, p 153-178.
Stephenson, D. a n d Corbetis, S.S., 1984. Economics of D e s a l i n a t i o n of
Wastewaters f o r the W i twatersrand. Water We1 I Assn. Conf. Johannesburg.
113
\
\
114

Fig. 7.8 Desalination costs per k g of dissolved solids removed


as a function of f e e d s a l i n i t y , f l o w r a t e a n d method
115

CHAPTER 8

COMPUTER ANALYSIS JUSTIFIES DESALINATION

INTRODUCTION

Industry uses water for cleaning, air conditioning, dilution and


t r a n s p o r t amongst other purposes. Water may be r e c y c l e d a n d d e t e r i o r a t e s
in q u a l i t y due to contact w i t h contaminants a n d discharges i n t o t h e system
(Holton and Stephenson, 1983). The total dissolved solids (TDS)
concentration i n some waters can exceed 10000 mg/l.
The poor water quality can lead to corrosion and deterioration of
pipework and machinery. In a particular system, corrosion of the
refrigeration plant was extremely severe. Other problems associated with
the poor water quality are the health hazard, and surface effluent
discharge regulations.
The problem has been aggravated by deterioration in surface water
q u a l i t y over the last few years.
The approach adopted in this study, namely to optimize the use of
resources, can be a p p l i e d on a much wider scale than herein described
(Stephenson, 1986). The computer program developed f o r the purpose of
re-distributing water in the most economical way to meet quality
constraints i s universal. I t could be a p p p l i e d on a r e g i o n a l b a s i s o r on a
smaller scale f o r i n d u s t r i a l water systems. The b a s i s f o r justifying high
operating-cost sources such as desalination lies in the fact that other
costs may thereby be saved, namely p u m p i n g costs o r storage dam costs.
Less water may be r e q u i r e d in total, as the water may be of a higher
quality. Effluent discharge problems may be reduced as the volume of
effluent ( b r i n e ) i s c o n s i d e r a b l y less i f d e s a l i n a t i o n i s performed.
There are alternative ways of justifying high operating-cost systems
such as d e s a l i n a t i o n in comparison w i t h more conventional sources such as
r i v e r water. They l i e i n c o n j u n c t i v e use - namely use of c a p i t a l - i n t e n s i v e
schemes to s u p p l y the base load, and r e s o r t i n g to s t a n d y low c a p i t a l - c o s t
sources when there is a shortfall in surface resources e.g. during a
drought. A l t e r n a t i v e l y d e s a l i n a t i o n could proceed on a r e g u l a r small scale,
d i s c h a r g i n g i n t o a r e s e r v o i r (e.g. a q u i f e r ) which c o u l d be tapped i n times
of shortage elsewhere.
The f i t t i n g i n o f a d e s a l i n a t i o n u n i t w i t h a system i s thus more l i k e l y
to j u s t i f y desalination than a straight comparison of u n i t costs of water
from a plant or from more conventional sources. The optimization of
116

multi-source systems requires careful analysis and this is where the


computer is helpful. There a r e various ways in which computer systems
analysis can be used. Simulation of the water distribution system will
reveal builup of TDS over time. A simultaneous solution of the mass
balance equation at nodes will give steady state TDS loads, and
o p t i m i z a t i o n can produce the best ( g e n e r a l l y the most economic) p l a n .
The type of d e s a l i n a t i o n p l a n t most s u i t a b l e f o r a n y g i v e n system w i l l
also depend on the circumstances (Stephenson a n d Corbetis, 1983). The TDS
values mentioned here could be described as requiring desalination or
demineralization (usually confined to lower TDS water). Membrane type
processes, e.g. reverse osmosis, are often most suitable, although
freeezing d e s a l i n a t i o n i s now r e c e i v i n g a t t e n t i o n .
Various attempts a t o p t i m i z i n g complex systems i n v o l v i n g water quality
have been reported (Rinaldi et al, 1984). In general, the problem is
non-linear so l i n e a r programming packages (Loucks et al, 1967) are of
little use. Search methods (Smeers a n d Tyteca, 1981) must generally be
used. The technique described i n t h i s p a p e r i s one such method which uses
known c h a r a c t e r i s t i c s o f the system to speed the search. The method a n d
the cost of desalination are not e x p l a i n e d in detail here although such
studies have been r e p o r t e d ( A b u l n o u r et a l , 1983).

APPL I C A T ION OF OPTIMIZATION O F WATER SUPPLY

The options studied, in view of the poor quality of the water on an


i n d u s t r i a l system were:

1. Accommodate the poor quality at the expense of higher maintenance


costs of pipework a n d machinery.
2. Prevent d e t e r i o r a t i o n of the water a t the source o f p o l l u t i o n .
3. Remove the TDS i n the water b y means of a d e s a l i n a t i o n p l a n t .
4. Use g r e a t e r q u a n t i t i e s of f r e s h water.
5. Re-distribute the water i n a way w h i c h m a i n t a i n s h i g h e r q u a l i t y a t key
points.

Alternative 1 , namely c o n t i n u i n g w i t h the present poor quality water,


was quickly r u l e d out by assessing the cost of regular replacement of
pipework and machinery. Research was also i n p r o g r e s s to reduce l e a c h i n g
a t the source of p o l l u t i o n b u t no p o s i t i v e recommendations c o u l d b e made.
The most economical combination of a l t e r n a t i v e s 3-5 was investigated u s i n g
the computer program developed for the purpose.
117

No.3 Shaft

Flsrure water

Mlm u r v l c a water

-
-L

... . .....
C b r water return Ilne
-
-. -
Carcmd. rvstem overflow
F a d rate;
Sludge llmr
----- MIM wwklngs u r d water
6 nrsure water

chamber water system


dam

F i g . 8.1 Section through a t y p i c a l mine

The water d i s t r i b u t i o n system i s often complex (Stephenson, 1983) a n d


not accurately documented. I n general, water i s p i p e d from one w o r k i n g to
another and it i s collected i n d r a i n s and pumped to waste or recycled
after s e t t l i n g to remove suspended p a r t i c l e s . The water may be used f o r
dust suppression and cooling. Minimum flows are required but the
d i s t r i b u t i o n p a t t e r n a n d the amount r e c i r c u l a t e d o r pumped can b e v a r i e d .
118

Water from d i f f e r e n t locations i s often blended. Water i s cooled by spray


evaporation or draught towers before refrigeration in an indirect
heat-exchange p l a n t i n c o o l i n g systems. The e v a p o r a t i o n i n the warm a r e a s
as well a s i n e v a p o r a t i o n dams a g g r a v a t e s the TDS concentration in the
water. F i g . 8.1 i l l u s t r a t e s a section t h r o u g h a t y p i c a l mine w o r k i n g s w i t h
water taken from a fresh water source at the surface, returned from
underground f o r cooling a t surface cooling towers, and p a r t l y replaced b y
f r e s h water before r e c y c l i n g .

SYSTEMS ANALYSIS

One s h a f t i n the complex d i s t r i b u t i o n system depicted in F i g . 8.1 can


be reduced to the flow d i a g r a m of F i g . 8.2, reproduced by the computer
program described here. The h y d r a u l i c p r i n c i p l e used for analyzing the
system i s the c o n t i n u i t y equation, t h a t is, net i n f l o w minus outflow a t a n y
node must equal zero. It is possible on this basis to make up any
d i s t r i b u t i o n system out of a number of closed loops. I t may be necessary
to close the system b y means of a dummy node i.e. a n y loose i n p u t to a n d
outflow from the system can be taken from o r to the dummy node. Such a
dummy node is handled differently from other nodes as no quality
c o n s t r a i n t o r mass TDS b a l a n c e a p p l i e s to i t .

F i g . 8.2 Graphic of system a n a l y z e d


119

I t may be shown t h a t the minimum number of closed loops in a network


of pipes and channels i s (Smith et a l , 1981):
L = P - N + l
where P i s the number of conduits connecting N nodes. Additional loops
may be defined b y i n c l u d i n g pieces of o t h e r loops a n d the selection of the
best loops can speed a n a l y s i s i f manual h y d r a u l i c a n a l y s i s i s performed.
A n algorithm for f i n d i n g loops i n a system described o n l y in terms of
nodes comprised p a r t of the program. For p r a c t i c a l purposes i t i s easiest
to identify conduits as flowing from one node to another and not as
s e p a r a t i n g adjacent loops, i.e. loops a r e not i d e n t i f i e d a t the d a t a input
stage. The d e f i n i t i o n of loops a t d a t a i n p u t stage i s also irksome when it
comes to r e v i s i n g , adding to o r subtracting conduits. D u r i n g r e v i s i o n of
flows i n t h e network, however, flows should a l w a y s b a l a n c e a t nodes. The
simplest way of e n s u r i n g t h i s i s to a d j u s t flows b y a d d i n g a flow around
closed loops.

MAIN PROGRAM SUB PROGS


~

IDENTIFY SYSTEM
READ CONDUIT DATA
IDENTlfY NODE CONS
SEEK ML LOOPS
HYDRAULIC NETWORK

. I
I I
ANALYSIS FOR CONDUITS
I
I I I
I TRYMTERNATNEDESM
PUNT LOCATIONS AND-SCALE I
I
I

SOLVE SIMULTANEOUS
MASS W C E TDS EOS
AT ALL NODES XPT 0
ADJUST FLOW IN LOOPS
IN BEST WAY UNTIL
AU TDS's WITHIN LIMITS

RETC COST

F i g . 8.3 Program flow d i a g r a m


1 20

A computational a l g o r i t h m f o r selecting t h e minimum number of closed


loops was developed. The algorithm can include unclosed branches by
inserting dummy conduits and nodes. Special h a n d l i n g of these conduits
was r e q u i r e d as o n l y flow balance, not TDS b a l a n c e occurs a t some dummy
nodes. For instance, e v a p o r a t i o n can be represented a s a n e g a t i v e flow f o r
a dummy node w i t h zero TDS. A flow d i a g r a m f o r the procedure i s g i v e n in
Fig. 8.3. To ensure t h a t each p i p e i s i n c l u d e d in a loop, the procedure
s t a r t s w i t h each p i p e in t u r n i n the system. It proceeds in the positive
flow direction (from top node to bottom node) from one pipe to another
l e a d i n g from i t . Where there i s more t h a n one p i p e e x i t i n g from a node, a
new loop i s created f o r each b r a n c h . At each step a check i s made f o r
loop closure and then a n y p i p e s in the series not in a closed loop are
dropped. Whenever a closed loop i s formed a check i s made, pipe b y pipe,
w i t h o t h e r loops to ensure no d u p l i c a t i o n . Loops w h i c h h a v e n e g a t i v e f l o w s
are ignored. A negative flow from node zero (which has zero TDS) is
p r e s c r i b e d to represent e v a p o r a t i o n since node 0 i s a l w a y s a t zero TDS.

START WITH W C H PIPE IN TURN


GO FROM TOP NODE TO BOTTOM OF
SUCCESSIVE PIPES
Where there Is a branch-off store plpedata
up to branch for another loop
CHECK FOR LOOP CLOSURE
ELIMINATE REDUNDANT PIPES
CHECK FOR LOOP DUPLICATION
BY COMPARING PIPE FOR PIPE
AND ELIMINATE DUPLICATES

GO BACK TO CHECK FOR BRANCH LOOPS

F i g . 8.4 Loop seeking a l g o r i t h m


121

Once a l l the loops a r e i d e n t i f i e d , a n increment i n flow i s added to one


loop a t a time to determine the corresponding change in TDS a t each node
and the corresponding cost increase. The node with the worst TDS, in
comparison w i t h t a r g e t TDS, i s i d e n t i f i e d a n d the improvement in TDS a t
the node p e r u n i t increment i s established. T h i s i s repeated f o r each loop
a n d the one w i t h the maximum of d(TDS)/dC i s selected where C i s the cost
of c i r c u l a t i o n . The maximum necessary increase in flow i s made a r o u n d the
loop consistent w i t h quality constraints. The procedure i s repeated until
a l l TDS's a r e w i t h i n specified l i m i t s .
I t i s necessary to r e c a l c u l a t e TDS's a t each node i n the system each
time an increment is made to flows around a loop. This is done by
i t e r a t i n g f o r each node in succession the mass b a l a n c e equation
T' = z Q ( T + P)/YQ
where Q i s the flow from an upstream node to the node a t w h i c h T I i s to
be determined. T i s the TDS a t the upstream node, and P i s the p i c k u p
( o r decrease i n the case of desalination) of TDS a l o n g the conduit. The
summation i s over a l l c o n d u i t s l e a d i n g to the node.

General o p t i m i z a t i o n problem

The problem to be solved can be described as follows: g i v e n minimum


flow requirements a n d maximum TDS requirements a t certain points in a
network, as well as q u a l i t y of r a w water a v a i l a b l e , r a t e of d e t e r i o r a t i o n
in water a t c e r t a i n p o i n t s a n d the network layout, what amount of water
should be purchased from the raw water source and what level of
d e s a l i n a t i o n should be performed? At the same time the program yields
flows in each conduit and quality (TDS) at each node. Costs in the
objective function to be minimized include cost of raw water, cost of
transport per Iitre along any conduit, e.g. pumping and cost of
d e s a l i n a t i o n in c / l f o r a l t e r n a t i v e l e v e l s of treatment (mg/l removal).
A t r i a l a n d e r r o r process i s r e q u i r e d to e s t a b l i s h the best p o s i t i o n of
the desalination plant or plants, and the best level of treatment
( p r o p o r t ion of TDS removed ) .
The costs were assumed to be linearly proportional to flow rate
a l t h o u g h non-linear o b j e c t i v e functions could be h a n d l e d b y the program.
C a p i t a l costs of c o n d u i t s were not i n c l u d e d a s the c o n d u i t s a l r e a d y exist
a n d can h a n g l e l a r g e r flows t h a n w i l l n o r m a l l y occur ( t h e y a r e based on
emergency c o n d i t i o n s ) . Flows can be c o n t r o l l e d b y valves in the case of
g r a v i t y lines and pumping power in the case of p u m p i n g lines.
122

PROGRAM APPL I CAT ION

The program was run on a 128 kb x 16 bit m i c r o computer which


handles up to 100 conduits. It is often run in conjunction with a
s i m u l a t i o n a n d g r a p h i c s program. The s i m u l a t i o n p r o g r a m a l l o w s for time
v a r i a t i o n in d r a w o f f s f o r meeting s p e c i f i c w a t e r demands, a n d f o r storage
which can fluctuate over a period. It does not, therefore, assume
c o n t i n u i t y of flow a t nodes a n d i s useful f o r more r e f i n e d studies t h a n the
o p t i m i z a t i o n program such as time varying quality (see Fig. 8.5). The
graphics system is useful for visualization of the system, although it
r e q u i r e s co-ordinates of nodes as i n p u t . A common d a t a f i l e can b e used
f o r a l l programs a n d t h i s f i l e can be amended ( c o n d u i t s a l t e r e d , added o r
s u b t r a c t e d ) as a n a l y s i s proceeds.
The o p t i m i z a t i o n p r o g r a m commences w i t h minimum flow d a t a . Flows can
o n l y be increased to reduce TDS a t a n y p o i n t . The location of d e s a l i n a t i o n
p l a n t s must be selected by trial, as well as the level o f desalination,
e.g. i f s l i p stream desalination i s resorted to, it i s equivalent to p a r t
removal. Although i t would be a simple m a t t e r to a u t o m a t i c a l l y investigate
desalination plants along successive conduits, the number of physical
c o n s t r a i n t s l i m i t i n g d e s a l i n a t i o n p l a n t s to s p e c i f i c locations o r l e v e l s does
not warrant automatic repositioning of plants. Brine disposal, space
requirements and access a r e of p a r t i c u l a r concern i n mines.

0 ,DAY
1 2 3 4 5 6 /

F i g . 8 . 5 Deterioration of TDS at a strategic point as indicated by


simulation
123

OPTIMIZATION O F MINE WATER SYSTEM

Mines s u f f e r from poor water q u a l i t y due to poor r i v e r q u a l i t y , large


concentrations of c h l o r i n e i n ground water, a n d a severe shortage of water
which necessitates extensive recycling. Baker-Duly (1985) r e p o r t e d on a
particular mine's method of solving the problems using a simulation
program.
The p a r t i c u l a r mine presented as a n example of the a p p l i c a t i o n of the
c u r r e n t o p t i m i z a t i o n p r o g r a m suffered the v a r i o u s problems out1 i n e d above.
The s i m p l i f i e d system a n a l y z e d here i s depicted i n Fig. 8.2, whereas the
r e a l mine i n v o l v e d 3 s h a f t s w i t h cross flows from one shaft to another. In
the system analyzed there i s one m a i n shaft from the surface down to
2140m below surface, and two sub-shafts down to 2430 and 2620m below
surface, respectively. There a r e s e t t l e r s f o r removing suspended sol i d s a t
the lowest level and at the bottom o f the main shaft. Water from the
s e t t l e r s i s d i v e r t e d i n t o sumps a n d thence pumped to higher levels. Part
of the water i s recycled with 'fresh' water purchased from a regional
water board. P r i o r to the study 23 I/s was thus purchased, the makeup
from groundwater totalled 22 I/s and evaporation amounted to 16 I/s
resulting i n a d i s c h a r g e to waste on the surface o f 29 I/s. Owing to the
h i g h heads, cost of p u m p i n g to waste was h i g h a n d the purchase p r i c e of
f r e s h water was also h i g h . The q u a l i t y of ' f r e s h ' water was in f a c t poor,
making marginal improvement to the system f o r the p r i c e p a i d . As water
q u a l i t y was of p a r t i c u l a r concern a t the r e f r i g e r a t i o n p l a n t s , f r e s h water
was directed s t r a i g h t to them. U n f o r t u n a t e l y , water from c o o l i n g dams was
also sent to the r e f r i g e r a t i o n p l a n t s , a n d t h i s water was of particularly
poor q u a l i t y . That was because the dams received warm water from the
mine workings where the water h a d come i n t o contact w i t h ore, with all
i t s contamination, and e v a p o r a t i o n a t the w o r k i n g s a n d the s p r a y i n g on
the dams had concentrated the dissolved s o l i d s even more b y e v a p o r a t i o n .
T h e fact t h a t water i n the s p r a y dams was of p a r t i c u l a r l y poor q u a l i t y
a n d that i t h a d to be improved c o n s i d e r a b l y before re-use, indicated the
most a p p r o p r i a t e p o s i t i o n f o r a d e s a l i n a t i o n p l a n t .

RESULT OF ANALYSIS

Table 8.1 i n d i c a t e s flows a n d TDS's a t a l l p o i n t s in the system p r i o r


to the analysis a n d Tables 8.2 a n d 8.3 present the optimum flows and
TDS's r e s u l t i n g from the a n a l y s i s f o r the best p o s i t i o n of the d e s a l i n a t i o n
plant.
1 24

Table 8.1 was obtained from the simulation program, where water
quality was arbitrarily assumed to start at 500 mg/P before increasing
over l i t t l e more than a week to e q u i l i b r i u m TDS's of over 4000 mg/l at
p I aces.

TABLE 8.1 Water Flows a n d TDS without d e s a l i n a t i o n

Cost, $/a = 1.056825E+6

~~ ~

Pipe Node 1 Node 2 Water Flow Increase TDS Cost Equilibrium TDS
number (W (will (CPU (WlU
1 0 1 23 800 80 i91
2 1 3 23 0 0 193
3 3 6 8 0 0 183
4 6 I 8 0 0 3296
6 0 11 -8 0 0 4133
6 0 I -8 0 0 3296
I 2 4 36 0 0 2419
8 4 6 61 0 0 2298
9 0 6 18 1800 0 2298
10 5 8 69 400 0 8180
11 8 7 62 0 6 8295
12 8 9 67 0 0 3174
13 9 2 67 0 20 8188
14 I 10 36 0 0 8138
16 10 12 40 200 0 8826
16 I 11 26 0 0 4183
17 11 12 18 200 0 3826
18 0 10 4 1800 0 3138
19 12 8 60 0 10 3180
20 12 0 8 0 60 0
21 2 0 21 0 16 0
22 3 4 16 0 0 2419

Table 8.2 summarizes the TDS's, with a desalination plant between


nodes 8 a n d 7, removing 1000 mg/l. The h i g h e s t TDS in the system i s 1334
mg/l a n d no a d d i t i o n a l f l o w s above minimum were r e q u i r e d . The net system
operating cost assuming a d e s a l i n a t i o n cost of 100 c / k l , would be $2.9
million/y. I n f a c t a maximum TDS of 2000 mg/l anywhere in the mine was
specified f o r this run whereas a maximum o f 1334 mg/l resulted without
any re-distribution or increase of flows.
For the r u n w i t h no d e s a l i n a t i o n i t was found t h a t i t was impossible to
achieve a maximum TDS as low as 1334 mg/l everywhere purely by
r e c i r c u l a t i n g more f r e s h water. T h i s i s p a r t l y because the TDS of the r a w
water was relatively high but even with better quality raw water (if
a v a i l a b l e ) considerable geochemical d e t e r i o r a t i o n o c c u r r e d u n d e r g r o u n d , so
125

there were limits to what could be achieved. Anyway, even with a


maximum TDS of 1603 mg/l, the cost of c i r c u l a t i n g water was $6.5 m i l l i o n / y
which exceeds the d e s a l i n a t i o n a l t e r n a t i v e f o r b e t t e r water. On the o t h e r
hand, if the TDS limit was relaxed to 2000 mg/l, the cost of the
recirculation solution became comparable with the desal i n a t ion solution,
namely $2.5 mi 1 I ion/year.
It should be noted that the costs only include pumping, raw water
purchases a n d d e s a l i n a t i o n . B r i n e disposal i s not costed as the b r i n e is
used to pump to the m e t a l l u r g i c a l plant where it i s used. The cost of
maintenance, especially replacement of corroded pipes, i s omitted as the
c o n s t r a i n t s (maximum TDS) were set to e l i m i n a t e corrosion problems. The
d e s a l i n a t i o n system cost i s more than j u s t i f i e d in s a v i n g s i n replacements
and it was only a question of whether to use more raw water or to
desalinate.

TABLE 8.2 New Water Flows and TDS w i t h d e s a l i n a t i o n

Max TDS 1344 mg/l, Cost, $/a = 2.912175E+6

~~

Pipe Node1 Node2 WatcrFlow lncmrrTD8 Cod BpuilibriumTD6


number (Ud (=g/l) (CW) cwn,
1 0 1 23 800 30 797
2 1 3 2s 0 0 798
3 3 6 8 0 0 783
4 6 7 8 0 0 448
6 0 11 -8 0 0 843
6 0 7 -8 0 0 448
7 2 4 36 0 0 1179
8 4 6 61 0 0 1339
9 0 6 18 1800 0 1839
10 6 8 69 400 0 1844
11 8 7 62 -1000 100 448
12 8 9 67 0 0 1342
1s 9 2 67 0 20 1348
14 7 10 36 0 0 681
16 10 12 40 200 0 799
16 7 11 26 0 0 643
17 11 12 18 200 0 799
18 0 10 4 1800 0 681
19 12 8 60 0 10 1344
20 12 0 8 0 60 0
21 2 0 21 0 16 0
22 3 4 16 0 0 1179
126

One factor omitted at this stage, however, is the requirement for


additional cooling of the recirculated water, if less raw water i s used.
The r a w water i s g e n e r a l l y a t a lower temperature (2lOC) t h a n the water
r e t u r n e d from the mine w o r k i n g s (28OC). The a d d i t i o n a l cost of o p e r a t i n g a
desalination plant underground where access is limited, is also not
p r o p e r l y e v a l u a t e d a t t h i s stage.
The most economical s i t i n g o f the d e s a l i n a t i o n p l a n t i s u n d e r g r o u n d as
this reduces p u m p i n g cost to the surface. It s h o u l d be a t the p o i n t of
highest TDS concentration as this results i n minimum p l a n t capacity per
u n i t removed.
The example illustrates the use of the program facilitated rapid
comparison of a l t e r n a t i v e water management systems on a cost a n d q u a l i t y
basis. I n general i t demonstrates t h a t , despite the f a c t t h a t desalination
costs quoted are often in excess of raw water costs, there are often
additional factors favouring desalination in industrial systems, namely:
cleaner water, less corrosion and blocking, less effluent, greater
conservation of natural resources, less pumping costs, and lower water
consumption.

TABLE 8.3 Optimum Flows to reduce TDS to 1603 mg/e

Best o b t a i n a b l e without d e s a l i n a t i o n , Cost, $/a = 6.544681E+6

Pipe Node 1 Node 2 Water Flow Increase TDS Cost Equilibrium TDS
number (116) ( W l ) (c/W (mg/l)
1 0 1 176 800 30 800
2 1 3 176 0 0 799
3 3 6 118 0 0 798
4 6 7 118 0 0 1239
6 0 11 -8 0 0 1303
6 0 7 -8 0 0 1233
7 2 4 36 0 0 1106
8 4 6 94 0 0 1216
9 0 6 18 1800 0 1216
10 6 8 112 400 0 1604
11 8 7 113 0 6 1233
12 8 9 212 0 0 1609
13 9 2 212 0 20 1603
14 7 10 78 0 0 1489
16 10 12 194 200 0 1697
16 I 11 146 0 0 1303
17 11 12 138 200 0 1697
18 0 10 116 1800 0 1489
19 12 8 213 0 10 1604
20 12 0 118 0 60 0
21 2 0 176 0 16 0
22 3 4 68 0 0 1106
127

REFERENCES

A b u l n o u r , A.M., S o r o u r , FA.H., Hammouda, F. and A b d a l Dayem, A.M.,


1983. S q u e e z i n g d e s a l t e d w a t e r c o s t s b y p r o p e r c h o i c e o f t h e d e s a l t i n g
t e c h n o l o g y a n d w a t e r management. D e s a l i n a t i o n , 44, 189-198.
B a k e r - D u l y , H.L.G., 1985. O p t i m i z a t i o n o f w a t e r r e t i c u l a t i o n s y s t e m s a t
L o r a i n e G o l d M i n e L i m i t e d , Proc. Mine V e n t i l a t i o n Society of South
A f r i c a Conf.
H o l t o n , M.C. and Stephenson, D., 1383. A c o m p u t e r model o f c i r c u l a t i n g
s e r v i c e w a t e r i n S o u t h A f r i c a n g o l d m i n e s . I n t l . J. M i n e Water, 2 ( 2 )
33-42.
L o u c k s , D.P., R e v e l l e , C.S. and Lynn, W . R . , 1967. L i n e a r p r o g r a m m i n g
m o d e l s f o r w a t e r p o l l u t i o n c o n t r o l . M a n a g e m e n t Science, 14, 8166-8181.
Rinaldi, S., Soncini-Sessa, R., Stehfest, H. and T a m u r a , H., 1979.
M o d e l l i n g and C o n t r o l o f R i v e r Q u a l i t y , McGraw H i l l , New Y o r k .
Smeers, Y . and T y t e c a , D., 1981. On t h e o p t i m a l l o c a t i o n o f w a s t e w a t e r
treatment p l a n t s , I n : J. T h i s s e and J. Z o l l e r ( E d s . ) , L o c a t i o n and
A n a l y s i s o f P u b l i c F a c i l i t i e s , N o r t h H o l l a n d , Amsterdam.
S m i t h , A.A., H i l t o n , E. and L e w i s , R.W., 1981. C i v i l E n g i n e e r i n g Systems
A n a l y s i s and D e s i g n .
Stephenson, D., 1983. D i s t r i b u t i o n o f w a t e r in g o l d m i n e s i n S.A., Intl.
M i n e Water J . , 2 ( 2 ) 21-30.
Stephenson, D., 1986. Computer analysis justifies desalination.
Desa I i n a t i o n , 58, 155-167.
Stephenson, D. a n d C o r b e t i s , S . , 1984. Economics o f d e s a l i n a t i o n o f w a s t e
w a t e r s f r o m t h e W i t w a t e r s r a n d , Proc. I n t l . Conf. o n Water Resources and
Desalination. Johannesburg, South A f r i c a , Water S u p p l y Improvement
Assn.
128

APPENDIX 8.1

MlNSlM PROGRAM FOR SIMULATING FLOW AND TDS IN CLOSED SYSTEMS

The program is for simulating flow and TDS changes in water


r e t i c u l a t i o n systems. It i s based on nodes a n d l i n k s , and calculates TDS
concentration at all nodes, and plots it at any selected node. Volume
v a r i a t i o n s a t nodes a r e p e r m i t t e d a l t h o u g h zero volume nodes can a l s o be
specified. The program will also draw a picture of the system at any
selected scale o r v i e w i n g angle. The program is in BASIC 3.0 for a HP
9816 micro computer.
To account for TDS build-up along a route, one may specify the
increase in TDS a l o n g the r o u t e in mg/l.
I f d e s a l i n a t i o n i s done, a n e g a t i v e increment i n TDS i s inserted.
When evaporation increases the TDS concentration at any node one
specifies a n e g a t i v e flow to t h a t node from another node such as node '0'
a t the node of o r i g i n a n d zero increase i n TDS a l o n g the l i n k route.
The volume a t each node i s set i n i t i a l l y a t a s p e c i f i e d value. I f this
v a l u e i s zero o r n e g a t i v e , a no-volume node i s assumed a n d inflow must
equal outflow. It is therefore not possible to specify flow variations
during the day from a zero volume node. From other positive volume
nodes, the flow can be s p e c i f i e d over so many h o u r s a d a y . Then in o r d e r
not to cause extreme volumes, t h e a v e r a g e flows i n t o a n d out of a l l nodes
should b a l a n c e d u r i n g each d a y , not the p e a k s w h i c h a r e s p e c i f i e d in the
data. i.e. Q.Tin should = 0 over 24 hours.
A maximum of 5 p i p e s o r l i n k s a r e p e r m i t t e d to each node.
When inputting data make sure the 'top' ( I ) node h a s been defined
before reading in data on the 'bottom' (J) node. If necessary use a
dummy p i p e from ' 0 ' to the node i n question to d e f i n e i t s co-ordinates.
Node 0 need not be so defined as no p i p e s from i t are plotted. Pipes to
node 0 a r e p l o t t e d b u t it i s not wise to have pipes to 0 a s they will
r e d e f i n e the TDS a t node 0, w h i c h c o u l d otherwise b e used as a s i n k f o r
e v a p o r a t i o n and a source f o r f i s s u r e s w h i c h would then increase i n TDS b y
a given figure.
Costs a r e c a l c u l a t e d i n D o l l a r s p e r annum if prices i n c/kl are input
in d a t a .

Tape or Disc Management

The programme MlNSlM can be copied onto new tapes. Data files
129

cannot. They have to be typed i n l i n e b y l i n e a f t e r the f o l l o w i n g i s set


u p on the tape.
CREATE "DATMIN", 100,88. T h i s creates a f i l e o f 100 records (for 100
p i p e s ) each 88 b y t e s long p e r m i t t i n g 11 x 8 b y t e numbers p e r record. To
erase a p r e v i o u s d a t a f i l e purge it a n d re-create it. Note a data file
m i g h t have to be closed manually after a bomb out. Alternatively while
executing the program i t pauses and asks whether graphics o r re-write of
d a t a i s r e q u i r e d . A new tape could be inserted a t t h i s stage.

MlNSlM L i s t of Symbols

A1 a n g l e of v i e w i n g p l a n e from X a x i s , degrees
A2 a n g l e of v i e w i n g p l a n e from 2 a x i s , degrees
B1-9 dummy i n p u t f o r a l t e r a t i o n s
C p r i c e c/kP
c2 t o t a l cost/rands p e r annum
E( i n i t i a l volume, m3. Use n e g a t i v e o r zero v a l u e to s i g n i f y
constant outflow over 24h. Must then b a l a n c e flows over 24h
not a t peaks.
F( ) new mg/P a t node
G( p o l l u t a n t concentration mg/P a t node
G1 G counter f o r p l o t s
GO used in c a l c u l a t i o n of TDS a t nodes
H( head, m not used
H1 size of device
I( top node of p i p e
J( bottom of p i p e
K ( I ,MI number of p i p e s connecting i n t o node ( u p to 5 p e r m i t t e d )
L( 1 length, m
L1 distance to device
MO p i p e counter
M Pipe
M1 number of nodes
M2 connecting p i p e counter
M3 counter f o r i n i t i a l flow c a l c u l a t i o n s
M5 p i p e no. of device
N node counter
NO device p e r p i p e
N1 node a t w h i c h TDS i s to b e p l o t t e d
N2 1 = TDS concentration p l o t r e q u i r e d
2 = volume a t node, m3
N3 0 = o l d data, 1 = new, 2 = r e v i s e d
N4 1 = g r a p h i c d i s p l a y , 0 = none, 2 = r e c o r d d a t a a n d stop
P( 1 i n p u t p o l l u t a n t concentration a t node 2, mg/e i n t o p i p e
PO in it i a I concentration
Q( flow P / s
91 ZQ
R( design flow i n p i p e P / s
S( 1 volume m3
s1 S coun'ter f o r p l o t s
T1 d u r a t i o n of simuln. d a y s
T2 s i m u l a t i o n i n t e r v a l , hours
T3 d r a w o f f hours/day (1st hours of d a y )
T4 time i n t e r v a l s f o r d r a w o f f p e r i o d
1 30

T5 device type; 1 = f l a n g e , 2 = v a l v e , 3 = tank,


4 = a r r o w , 5 = square
no. i t e r a t i o n s p e r d a y
i t e r a t i o n counter
d a y counter
d a y counter
screen X co-ordinate
screen Z co-ordinate
X m i n on screen
X max on screen
Z m i n on screen
Z max on screen
X co-ordinate
Y co-ordinate
Z co-ordinate

Data Input

The computer asks f o r the f o l l o w i n g d a t a to be typed i n i n t e r a c t i v e l y :

Ll System name
L2 Simulation p e r i o d , d a y s
L3 Time increment, d a y s
L4 Period in h o u r s p e r d a y d u r i n g which d r a w o f f occurs. The
b a l a n c e of the time water may flow to r e f i l l r e s e r v o i r s .
L5 I n i t i a l TDS of the e n t i r e systems, mg/P
L6 Node no. a t which a p l o t o f TDS versus time i s r e q u i r e d
L7 Type 0, 1 o r 2 depending on whether the o l d d a t a f i l e , a
new one o r a r e v i s i o n of the o l d one i s r e q u i r e d
L B et seq Type the f o l l o w i n g separated b y commas, w i t h one l i n e p e r
pipe o r conduit;

Top ( u p s t r e a m ) end node no.


Bottom (downstream) end node no.
x-co-ordinate of bottom node ( h o r i z o n t a l l y from a d a t u m )
y-co-ordinate of bottom node ( v e r t i c a l l y )
z-co-ordinate of bottom node ( i n t o screen)
Volume of storage a t bottom node, m’
Design flow in P / s
~ n p u tp o l l u t i o n , mg/e
Cost a l o n g route, cents p e r u n i t o f flow (ke)
Type a row of n i n e zeros to end t h i s d a t a
L a t e r the program may c a l l f o r a d d i t i o n a l g r a p h i c s d a t a p e r
pipe :
Pipe no. (counted from the top of the l i s t )
Position ( d i s t a n c e from top end o f p i p e ) to a device
Device t y p e ( 1 = f l a n g e , 2 = v a l v e , 3 = t a n k , 4 = a r r o w ,
5 = square)
Size, m to d r a w i t
Cost p e r u n i t size

The d a t a w i l l then be f i l e d f o r subsequent re-use b y the second


program MINOP. Examples of o u t p u t , g r a p h i c s a n d i n p u t a r e g i v e n in the
char, ter.
131

Program listing
1 I RE-STORE"MINS1M'
10 I MINE WATER RETIC SIMULATION I"M1NSIM"
1 1 GRAPHICS OFF
12 DUMP DEVICE I S 707,EXPANDED
1 4 PRINTER I S 707
2 0 ASSIGN O P a t h l TO "DATMIN" I CREATE "FILNAM",100,88 ag"DATM1N"
2 1 D I S P "SYSTEM NAME"i
22 INPUT NS
2 4 PRINT NE
2 5 INTEGER N ,NO ,N1 ,N2 ,N3 ,N4 ,M ,M0 ,M1 ,M2 ,M5 ,S I ,I( 99 ) , J ( 99 ) ,K( 99.5 )
2 7 D I M X ( 9 9 ) , Y ( 9 9 ) .Z( 9 9 ) ,U( 9 9 ) ,W(99) ,P( 9 9 ) , H ( 9 9 ) ,6( 9 9 ) , S ( 9 9 )
30 D I M E( 9 9 ) ,R( 9 9 ) ,P( 9 9 ) ,L( 9 9 ) .F( 9 9 ) ,C( 99 )
31 D I S P "DURN OF SIMULATION,DAYS"I
3 2 INPUT 1 1
33 D I S P "TIME 1NCREHENT.HOURS"t
3 4 INPUT 1 2
3 5 D I S P "FLO OVER HOURS/DAY"i
36 INPUT T3
37 D I S P " I N I T I A L TDS,mQ/l"(
3 8 INPUT P0
39 D I S P "PLOT TDS AT N0DE"i
4 0 INPUT N1
4 2 DEG
4 3 D I S P "OLD OR NEW OR REV D A T A ( 0 / 1 / Z ) " i
45 INPUT N3
61 NZ-1
62 T4-T3/T2 I I T S / d
66 1 7 - 2 4 / 1 2
70 G(O )=0
80 X ( 0 )-0
90 Y(0j-0
1 0 0 Z(0)-0
110 E ( 0 ) - 0
111 FOR N-1 TO 9 9
112 S ( N ) = 0
113 E ( N ) - 0
114 F ( N ) = 0
115 X ( N ) = 0
116 Y(N)=0
117 Z(N)=0
118 C(N)-0
1 1 9 NEXT N
120 c2-0
122 G( 1 )=P0
1 2 5 M1=0
130 FOR M=1 TO 9 9
1 4 0 I F N3C>1 THEN 190
1 4 5 I NEW PIPE DATA
150 DISP "N1 ,N2,X2,YZ,ZZ,U2,l/s,tmg/l,c/"i
1 6 0 INPUT I ( M ) , J ( M ) , X ( J ( M ) ) ,Y( J ( M ) ) ,Z( J ( M ) ) ,E( J ( M ) ) , R ( M ) ,P(M) ,C( M )
170 OUTPUT @ P a t h 1.Mi I ( M ) , J ( M ) , X ( J ( M ) ) .Y( J ( M ) ) ,Z( J ( M ) ) ,E( J ( M ) ) , R ( M ) , P ( M ) ,C(M)
1 8 0 GOT0 2 10
1 8 5 I OLD PIPE DATA
1 9 0 ENTER @ P a t h 1, M i I ( M ) , J ( M ) , X ( J ( M ) ) , Y ( J ( M ) ) , Z ( J ( M ) ) ,E( J ( M ) ) ,R(M),P(M),C(M)
2 1 0 I F I ( M ) + J ( M ) = 0 THEN 228
212 S(J(M))=E(J(M))
218 G(J(M))-P0
2 2 0 C2=CZtC( M )rR( M )*315
225 M l = M l + l
132

226 NEXT M
228 I F N3<2 THEN 256
230 FOR M0=l TO 99 I REV P I P E DATA
231 DISP "PIPE N o . " i
232 INPUT M
233 C2=CZ-C( M ) * R ( M ) * 3 1 5
234 DISP "N1 , N 2 , X 2 , Y 2 , ~ 2 , U 2 , l / s , t m g / l , c / " i
235 INPUT I ( M ) , J ( M ) , X ( J ( M ) ) , Y ( J ( M ) ) ,Z( J ( M ) ) ,E( J ( M ) ) , R ( M ) , P ( M ) , C ( M )
245 OUTPUT B P a t h l , M i I ( M ) , J ( M ) , X I J ( M ) 1 , Y ( J ( M ) ) ,Z( J ( M ) ) ,E( J ( M ) ) , R ( M ) , P ( M ) ,C(M)
246 I F I ( M ) t J ( M ) = 0 THEN 256
?48 I F MCMl THEN 251
249 Ml=MI+l
251 S ( J ( M ) )=E( J ( M ) )
253 G( J ( M ) )-P0
254 CZ=C2+C(M)*R(M)*315
255 NEXT M 0
256 FOR M - l TO M 1
257 L ( M ) = S Q R ( ( X ( J(M))-X(I(M)) ) ^ 2 + ( Y ( J ( M ) ) - Y ( I ( M ) ) ) * 2 + ( 2 ( J(M))-Z( I ( M ) ) ) ^ 2 1
258 NEXT M
262 PRINT "N1 N2 X2 Y 2 22 U2 Q t m g c / "
263 FOR M=1 TO M 1
264 PRINT USING 2651 I ( M ) , J ( M ) , X ( J ( M ) ) , V ( J ( M 1 ) ,Z( J ( M ) ) ,E( J ( M ) ) , R ( M ) . P ( M ) , C ( M )
265 IMAGE 20,2D ,5D, 4D,5D ,4D ,3D, 4D,3D
267 NEXT M
268 D I S P "LAYOUT GRAPHICS(0=NO,l=YES,2=RECORD DATA & STOP ) " I
269 INPUT N4
270 I F N4<2 THEN 2 8 0
271 ASSIGN O P a t h l TO "DATMIN"
272 FOR M=1 TO 99
274 OUTPUT @ P a t h 1, M i I ( M ) , J ( M ) ,XC J ( M ) ) , Y ( J ( M ) ) ,Z( J ( M ) ) ,E( J ( M ) ) , R ( M ) ,P(M) ,C(M 1
276 I F I ( M ) t J ( M ) = 0 THEN 2910
278 NEXT M
286 I F N4-1 THEN 1570
706 ALPHA OFF
705 GINIT
707 GRAPHICS ON
708 I F N2>1 THEN 8 5 0
710 WINDOW - . 5 , T I . - 2 0 0 , 2 * G ( I )t3000
71 1 C L I P 0 ,T1 ,0 ,2*Gc 1 )+3006
720 AXES 1 ,100
730 C L I P OFF
740 MOUE T1-1,10
750 LABEL " O A Y "
760 FOR T0=1 TO T I
770 MOVE TO-.5,-200
780 LC\BEL UALO(T0)
790 NEXT T0
800 FOR G I - 0 TO 2 * G ( l ) + 2 5 0 0 STEP 500
810 MOUE -.5,G1
820 LABEL UALS(G1 )
830 NEXT 61
835 MOUE .5,2*G( 1 )+600
840 LABEL "TDSmg/l NODE"LUALB(N1 )
845 60TO 890
850 WINDOW -.5,Tl ,-10,E(NI ) t l U
851 C L I P 0 , T l .-10,E(N1 ) + I 0
854 AXES 1 , l
855 C L I P OFF
133

857 MOUE T l - 1 ,0
868 LABEL "DAY"
860 FOR T0=l TO T1
862 MOUE T0-.5,-10
864 LABEL UALS(T0 )
866 NEXT 1 0
8 6 8 FOR S l - 0 TO E ( N 1 ) t l O STEP 10
870 MOUE -.5,S1
872 LABEL UCILL(SI )
874 NEXT S1
876 MOUE B,E(Nl ) + I 0
878 LABEL "UOLn3,NODE"bUALO(Nl)
890 I END OF LABELING
900 FOR N=1 TO M l t l I NODES CONS
905 M2=0
906 FOR M0=l TO 5
907 K ( N ,M0 )-0
908 I ( K ( N , M 0 ) ) = 0
909 J(K(N,M0))=0
910 NEXT MO
915 FOR M0=1 TO M 1 1 PIPES
920 I F I ( M 0 ) < > N THEN 940
925 MZ-M2tl
927 K ( N ,M2 )-M0
930 GOTO 960
940 I F J(M0)<>N THEN 968
945 M2=M2+1
950 K(N,MZ)=MB
960 I F M2-5 THEN 970
968 NEXT M0
970 NEXT N
975 I F NZ>l THEN 990
980 MOUE 0,G(N1 )
985 GOTO 1090
990 MOUE 0,S(Nl )
1090 FOR T=1 TO T l ! DAYS
1100 FOR T9=1 TO T7
1105 I F T9>T4 THEN 1160
1110 FOR M=1 TO M1
1145 Q ( M ) = R ( M )
1155 NEXT M
1157 GOTO 1200
1160 I F E ( I ( M ) ) < = 0 THEN 1200
1195 Q(M)-0
1200 TDS a STORAGE TINS
1210 FOR N-1 TO M 1 + 1
1220 I F E ( N ) > 0 THEN 1340
1225 Gl=Q
1228 Q1=.001
1230 FOR M0-1 TO 5
I 2 3 5 I F J ( K ( N , M B ) ) O N THEN 1280
1236 I F G( I ( K ( N , M 0 ) ) )+P(K(N ,M0) )<=0 THEN 1255
1240 G1=( G( I ( K ( N ,M0) ) ) t P ( K ( N , M I ) ) )*Q(K(N .M0) )+G1
1255 01 =Ol + Q ( K ( N ,M0 ) )
1280 NEXT MO
1282 G l = G l / Q l
1285 GOTO 1440
1340 G l m 0 1 S>0
1345 G0=G( N )*S( N )
134

1350 FOR MO-1 TO 5


1360 I F J(K(N,M0))<>N THEN 1390
1370 G1-61+(6( I ( K ( N,M0 ) ) )+P(K( N ,M0 ) ) )*Q( K ( N ,M0 ) )
1375 S(N)=S(N)tQ(K(N,MB) )*T2*3.6
1380 GOT0 1420
1390 I F I(K(N,MB))<>N THEN 1420
i 400 GI=GI-G(N')*Q(K(N ,MQ ) )
1412 S(N)=S(N)-Q(K(N,MQ) )*T2*3.6
1428 NEXT M 0
1430 61-( 61 *T2*3.6+60 ) / S ( N )
1440 F ( N ) = G l
1472 NEXT N
1474 FOR N-1 TO M l + l
1476 G(N)=F(N)
1488 NEXT N
1492 I F N2>1 THEN 1488
1485 DRAW T - l t T g * T 2 / 2 4 , 6 ( N l )
1486 60TO 1490
1488 DRAW T - l t T 9 * T 2 / 2 4 , S ( N l )
1490 NEXT T9
1492 NEXT T
1494 DUMP GRAPHICS 8707
1496 GOT0 2920
1570 DISP "XMIN ,XMAX ,ZMIN ,ZMAX ,XANGL ,ZANGL' I
1580 INPUT ~ 0 , ~ 9 , ~ 0 , ~,AZ 9 , ~ 1
1590 GINIT
1592 GRAPHICS ON
1595 DEG
1600 WINDOW UQ,U9,W0,W9
1601 FOR M=1 TO M 1 ! NODES
1602 U(M)=X(M)*COS(Al ) t Y ( M ) + S I N ( A l )
1604 W ( M )-Z( M )+COS( A2 ) t (Y ( M )tCOS( A1 )-X( M )*SIN( A 1 ) )*SIN( A2 )
1608 NEXT M
2170 FOR M=1 TO M l ! PIPES
2190 PEN 1
2195 I F I ( M ) - 0 THEN 2225
2196 I F J(M)-0 THEN 2230
2200 MOUE U ( I( M ) ) ,W( I( M ) )
2210 DRAW U ( J ( M ) ) , W ( J ( M ) )
2220 LABEL Uc\L$( J ( M ) )
2223 GOT0 2230
2225 MOUE U( J ( M ) ) ,W ( J ( M ) )
2228 LABEL UALO( J ( M ) )
2230 NEXT M
2235 FOR N0=l TO 3
2240 FOR N=1 TO 100
2241 I F N>Ml THEN 2415
2242 I F NO-1 THEN 2256
2243 I F N0=3 THEN 2269
2246 M5-N I ARROWS
2247 I F I ( N ) = 0 THEN 2410
2248 L l = L ( N ) / 2
2249 I F J ( N ) - 0 THEN 2410
2250 T5=4
2251 Hl-L1/10
2254 C0-0
2255 GOTO 2320
2256 M5=N ! TANKS
2258 L l = L ( N )
2260 1 5 1 3
135

2262 H l = E ( J ( N ) ) / 2 5
2264 C0=0
2266 GOT0 2320
2268 ALPHA ON
2 2 7 0 DISP "PIPEn ,X ,TYPE ,SIZE .COST/' 1
2280 INPUT M5,Ll ,T5,H1 ,C0
2 3 2 0 I F M5-0 THEN 2420
2340 X5=X( I ( M 5 ) ) + L l / L ( M 5 ) * ( X ( J(M5 1 )-X( I ( % ) ) )
2350 Y5=Y( I ( M 5 ) )+Ll /L(M5 ) * ( Y ( J( M5 ) )-Y( I(M5 ) ) )
2360 Z5=Z( I ( M 5 ) )+L1 /L(M5 )+(Z( J(M5) )-Z( I ( M 5 ) ) )
2 3 7 0 US=XS*COS(AI ) t Y 5 * S I N ( A 1 )
2 3 8 0 W5=Z5*COS(A2 )t(YS*CDS(Al )-XS*SSN(Al ) ) * S I N ( h Z )
2390 ON T5 60TO 2460,2490,2540,2590,2850
2400 I l=FLANGE,Z=UALUE,3=TANK,4=~RROW,S=S~UARE
2 4 1 0 NEXT N
2415 NEXT N0
2420 MOUE U0,WB
2430 C2=1NT(C2 )
2440 LABEL " R/s="LUALO(C2)
2445 DUMP GRAPHICS
2 4 5 0 60T0 7 0 0
2460 MOUE US ,W5tHI / 2
2 4 7 0 DRAW US ,W5-H1/2
2480 GOT0 2410
2 4 9 0 HOVE U5-H1/2,WStH1/2
2500 DRAW U 5 t H I 12 ,WS-HI / 2
2510 MOUE U5+Hl/2,WS+H1/2
2520 ORAW U5-H1/2,U5-H1/2
2530 GOTO 2410
2540 MOUE U5-HI / 2 ,WStWI
2550 DRAW U 5 - H l / 2 ,W5
2560 DRAW U S H 1 / 2 ,WE
2 5 7 0 DRAW U5+Hl/Z,W5+Hl
2580 60TO 2410
2590 I F U ( J ( M S ) ) < > U ( I ( M S ) ) THEN 2601
2591 IF W ( J ( f l S ) ) > W ( I ( M 5 ) ) THEN 2594
2592 U8=270
2 5 9 3 GOTO 2608
2594 U8=90
2595 GOT0 2608
2601 UE=ATN( (U( J( M 5 ) )-W( I( M5 ) ) ) / (U( J ( M 5 ) )-U( I( M5 ) ) ) )
2602 I F U8>=0 THEN 2606
2603 I F W ( J(M5) ) < W ( I(M5) 1 THEN 2608
2604 GOTO 2607
2606 I F U ( J ( M S ) ) > W ( I ( M S ) ) THEN 2608
2607 U8=U8t180
2608 UG=U5-Hl*COS(U8-45)
2610 W6=W5-Ht*SIN(U8-45)
2 6 2 0 U7-U5-Hl*COS(U8t45)
2630 W7-W5-Hl*SIN(UEt45)
2 8 1 0 MOUE U6,W6
2820 DRAW U5,W5
2830 DRAW U7,W7
2 8 4 0 6QTO 2410
2850 MOUE US-H1/2,WStHl
2860 DRAW U5-H1/2 ,W5
2a70 DRAW u 5 + ~ 12 1 ,w5
2880 DRAW U5tH1/2,W5tHI
2890 DRAW U5-H1/2,W5tHl
2900 60TO 2410
2920 END
136

APPENDIX 8.2 MINOP p r o g r a m f o r o p t i m i z i n g d i s t r i b u t i o n

MlNOP List of Symbols

p r i c e c/ke
cost
t o t a l cost
dummy
TDS
max. TDS desired, o r G of node w i t h m a x . T D S
increment i n TDS
max. increment i n TDS
t o t a l TDS - mg/s i n t o node
t o t a l flow i n t o node
TDS
G(I) - H(I)
max. TDS
top node
bottom node
p i p e no. connecting to node ( u p to 5 p e r m i t t e d )
number of p i p e connecting
number loops
best loop no.
branches i n loop
p o s i t i v e loop
p i p e s out node
loop counter
number loops a n d b e g i n number loop
number connecting p i p e s to node
p i p e number
p i p e counter
number of nodes
number of p i p e s in loop
number of connecting p i p e s from node
dummy
p i p e s out node
number of p i p e s in loop
r e d u c t i o n i n no. p i p e s in loop, o r , p i p e to node w i t h max.
TDS
p i p e number i n loop
b e g i n p i p e f o r loops
name
node counter
no. nodes
p i p e no.
i n p u t TDS, mg/e
flow e/s
dQ/dC
dQ
co-ord. not used i n MINOP
0,

I1
137

Notes on program

The program i s in BASIC for an HP 9816 series 200 micro computer. The
data f i l e i s obtained from the M l N S l M program in appendix 8.1.
138

Program MINOP listing


101 RE-STORE"MIN0P"
20 I "MINOP " OPTIMZS FLOS I N NETWORK SUBJECT TO TOS L I M I T S
3a PRINTER IS 707
40 A S S I G N @ P a t h 1 TO "DATMIN"
5 0 I D I S P "SYSTEM NAME"!
60!INPUT NO
7 0 I P R I N T NC
80 DIM Q( 9 9 ) ,G(99) ,P( 9 9 ) ,C( 9 9 ) , L ( 9 9 ) ,H( 9 9 )
90 INTEGER I ( 9 9 ) , J ( 9 9 ) , K ( 9 9 , 5 ) , K 1 ( 9 9 , 9 ) , L l ,L2 ,L3,M,M0,Ml , M 2 ( 5 0 , 5 0 ) , M 3 ( 9 9 ) .M4 ,M
S.M6(99),M7,M8,M9,N,Nl
1 0 0 M I 1 0 1 NO.PIPES
110 G ( 0 ) = . 1
120 N l = l
130 DISP "MAX TDS DES1RED"i
1 4 0 INPUT 6 0
1 5 0 FOR M=l TO 9 9
160 ENTER @ P a t h 1,M; I ( M ) , J ( M ) ,X ,Y ,Z ,E ,Q( M ) ,P( M ) ,C( M )
170 H( .I(M ) )=G0
180 I F I ( M ) t J ( M ) = 0 THEN 2 5 0
190 I F I(M);=Nl THEN 2 1 0
200 Nl=I(M)
216 If J ( M ) < = N l THEN 2 3 0
220 N1 =.Jc M )
230 MI-Mltl
2 4 0 NEXT M
250 H(0)-100000
2 6 0 FOR M0=l T O M 1
2 70 D I S P "ANY CHANGES? PIPENo,TOPn,BOTn,FLOI/s,POLmg/l , c / h 1 (O's=none ) " I
280 INPUT M , I ( M ) , J ( M ) ,QC M 1 ,P( M ) ,C(M )
290 I F I ( M ) + J ( M ) = 0 THEN 3 2 0
300 I F M > M l THEN M l = M l t l
310 NEXT MO
3 2 0 FOR N-0 TO N l I NODES
330 G( N )=G0
340 M3( N )=0
350 L( N )=0
360 FOR M=l T O M l I P I P E S FROM NODE
370 I F I ( M ) < > N THEN 4 0 0
380 M3( N )-M3( N ) t1
390 K l ( N ,M3( N ) )=M
400 NEXT M
410 FOR M0=1 TO M 1 I P I P E S TO NODE
420 I F J ( M Q ) < > N THEN 4 5 0
430 L( N )=L( N ) + 1
440 K ( N L ( N ) )=NO
~

450 NEXT M0
4 6 0 NEXT N
4 7 0 G( 0 1-0
4 8 0 L1=01LOOFS
4 9 0 FOR.M9=1 TO M1 IBEGINPIPE FOR LOOPS
500 L 0 - L l t l ITRY LOOP
51 0 L8=0
520 M G ( L 0 ) = l I N O . P I P E S I N LOOP
530 M2(L0,1 )=M91PIFES I N LOOPI
540 LE=LBIPOS LOOP
550 FOR L 3 = 1 TO MlIBRANCH ROUTINE
560 L8=0
570 L4=M3( J ( M 2 ( L0 ,M6( L0 ) 1 ) )
580 FOR M5=1 TO L 4 I P I P E S OUT NODE
139

5 90 I F M5-1 THEN 670


600 L6=LG+lIANOTHER POS LOOP FROM BRANCH
610 M6(L6 )=M6(L0)
620 FOR M7=1 TO M6(L6)-1
630 MZ(L6 ,M7)42(LB,M7 )ICOPIES PIPES I N PREU LOOP
640 NEXT M7
650 M 2 ( L6 ,M6(L6 ) )=K 1 ( J ( M 2 ( L 0 ,M6( LO ) - 1 ) ) ,M5 )
660 GOTO 690
670 M6(L0)=M6(L0)+1 IN0 PIPES I N LOOP
680 M2( L 0 ,M6( L 0 ) )=K 1 ( J ( M 9 ( L0 ,M6( L0 1- 1 ) ) ,M5 ) I NEXT PIPE
690 NEXT M51 CHEK LOOP CLOSURE
700 FOR M5-2 TO M6(L0)
71 0 FOR M7=1 TO M5
720 IF I( MZ(L0 ,M7) ) < > J M( Z ( L 0 ,M5) ) THEN 800
730 L1=L1 t 1
740 M6(L1 )4lS+I-M71SHUFFLE UP PIPES
750 FOR M8=l TO M 6 ( L l )
760 M Z ( L I ,M8 )-M2( L 0 ,M8+M7- 1 )
770 NEXT ME
780 L6=1
7 90 GOTO 840
800 NEXT M7
81 0 NEXT M5
820 GOTO 1000
830 I CHEK DUP LOOP
840 I F L l ; = l THEN 1000
850 FOR LZ=1 TO L1-1
860 M=O
870 FOR M7=1 TO M6(Ll )
880 M=Mt 1
890 ME-1
900 I F M Z i L l ,M)OM2(L2,M8) THEN 980
91 0 M8=M6+1
920 M=Mt 1
930 IF M:=MG(LI) THEN 950
940 M= 1
950 I F M8:=M6(L7) THEN 9G8
960 Ll=Ll-llREMOUE OUP L O O F
970 GOTO 1000
980 NEXT M7
990 NEXT L2
1000 I F L 8 < - 0 THEN 1030
1010 I F L 6 ‘ = L 0 THEN 1040
1020 L0=L0+ 1
1030 NEXl L3
1040 NEXT M9
I 0 5 0 FOR L:=l TO L I
1060 FOR M=I TO M 6 ( L 2 1
IQ701PRINT L:,MZ(L2,W)
1080 NEXT M
1090 NEXT L2
1100 GOSU8 1120
1110 GOTO 1300
1120 FOR L4=1 TO Mi
1130 G5-0
1140 FOR N=i T O N l
1150 Gl=.l
1160 G’L=.I
1170 FOR O=l TO L ( N )
1180 Gl=Gl t Q ( K ( N,O ) * ( G : I ( k‘: N , O ) ) ) t F ( K ( N ,O ) ) )
140

1190 62=62tQ(K(N,O))
1200 NEXT D
1210 63=6(N)
1220 6(N )=GI 162
1230 64=AES( 6( N )-63)/G(N )
1240 I F 64<65 THEN 1260
1250 6544
1260 NEXT N
1270 I F 65<.00I THEN 12901 MAX FC
1280 NEXT L 4
1290 RETURN
1300 FOR M=1 TO N1 IITNS
1310 RI=BIDQ/DC
1320 R3=0IDQ
1330 FOR L 2 = l TO LllEEST LOOP
1340 C1=0
1350 H1=0
1360 M0=6
1370 FOR M8=1 TO MSfL2 )
1380 C1 =C1tC(MZ(L2 ,M8 ) )
1390 I F Q ( M 2 ( L Z 3 M 8 ) ) < 0THEN 1540 INEXT LOOP
1400 M0=M8
14 10 Q( M2( L2 ,M8 ) )=Q( M2( L2 ,M8 ) ) t1
1420 I F J(M2(L2,M8))=0 THEN 1470 !NEXT PIPE
1430 I F G( J ( M 2 ( L2 ,M8 ) ) )-H( J( M 2 ( L Z ,M8 ) ) )<=HI THEN 1470
1440 HI =G( J( M 2 ( L2 ,M8 ) 1 )-H( J ( M 2 ( L2 ,M8 ) ) )
1450 M7=MZ(L2,MB)IPIPE TO NODEWITH MAX TDS
1460 60=G(J(MZ(LZ,M8)))
1470 NEXT ME
1480 I F H l i = l THEN 1540
1490 GOSUE 1120
1500 I F ( G Q - G ( J ( M 7 ) ) ) / C I ~ ' = R 1THEN 1540
1510 Rl=(G0-G( J(M7) ) ) / C l
1520 R3=(G0-H( J(M7)) ) / t G Q - G ( J ( M 7 ) ) )
1536 L3=L2
1540 FOR M8=1 TO MQ
1550 Q( MZ(L2 ,ME) )=Q(MZ(L? , M a ) ) - 1
1560 NEXT M8
1570 NEXT L t
158@ FOR M7=1 TO M6(L3i
1590 Q ( M ~ ~ L ~ , M ~ ) ) I Q ( M ~ ( L )~t ,RM3 ~ )
1600 NEXT M7
1610 NEXT M
1620 C 2 = 0
1630 PRINT " P n N l N2 1 / s tTDSng1 c / L 1 T D S 2 '
1640 FOR M-1 TO M1
1650 PRINT USING 1660;M,I(M) , J ( M ) , Q ( M ) ,P(M) , C ( M ) ,G( J(M))
1660 IMAGE 2D ,4D ,4D ,4D ,6D ,50 ,SO
1670 C2=C2tC(fl)*Q(M)*315
1680 NEXT M
1690 C2=INT(C2 )
1706 PRINT "COST ,R/a="i C Z
1710 ASSIGN 0Pathl TO
1720 EN0
141

CHAPTER 9

INTEGER PROGRAMMING PLANN ING OF TREATED WASTEWATER CONVEYANCE FOR


ARTIFICIAL RECHARGE OF AN AQUIFER

INTRODUCTION

The international growth i n water demand over the last few decades
has been p e r s i s t e n t l y h i g h . T h i s r a t e of growth i s l i k e l y to c o n t i n u e a s a
l a r g e p r o p o r t i o n of the p o p u l a t i o n is increasing r a p i d l y in standard of
living. The a v a i l a b i l i t y of new sources of water h a s tended to l a g b e h i n d
demand. Even i f new sources were a v a i l a b l e the cost of p r o v i d i n g to meet
any possible drought extreme in developing areas could be high, on
account of the u n r e l i a b i l i t y of r i v e r flow. Surface water can o n l y be used
t o i t s f u l l e s t extent i f a l t e r n a t i v e sources a r e a v a i l a b l e to meet essential
demands i n times of d r o u g h t . For t h i s reason the w o r l d i s now l o o k i n g to
groundwater a n d wastewater to meet s h o r t f a l IS.
A scheme i s i n v e s t i g a t e d here to s u p p l y from groundwater o n l y a t the
r a t e a t which i t can be n a t u r a l l y replenished. Separate studies a r e b e i n g
conducted on t e r t i a r y treatment of wastewater but the idea of using the
wastewater to a r t i f i c i a l l y recharge groundwater with wastewater is only
now r e c e i v i n g consideration. Such research is a long term project and
cannot be expected to re1 ieve c u r r e n t droughts, w h i c h however p r e c i p i t a t e d
research i n t o a l t e r n a t i v e sources of water.
It i s proposed to use groundwater i n conjunction w i t h surface water
resources in such a way that deficiencies in surface water can be
supplemented b y groundwater, resulting in a higher overall availability.
Surface water resources can then be u t i l i z e d to a g r e a t e r degree since
groundwater reserves can be d r a w n on in times of shortfalls in surface
r i v e r s (Paling, 1984). The r a t e of recharge w i l l also be r e l a t i v e l y slow
owing to limited suitable wastewater being available, the possibility of
natural purification and the limited permeability of the soil. The
h y d r a u l i c s of the recharge process should be i n v e s t i g a t e d w i t h s i t e tests.
The case study a n a l y z e d i s the Witwatersrand area, a h i g h growth r a t e
conurbation based o r i g i n a l l y on m i n i n g . Groundwater constitutes a t present
o n l y one percent of the average d a i l y supply to t h e Rand Water Board of
2400 M l / d to the Witwatersrand area. Privately owned boreholes for
farming and gardening purposes are however in common use as a
I

Fig. 9.1 Dolomite deposits i n the Witwatersrand a r e a


143

supplement to the formal supply. F o l l o w i n g a series of low r a i n f a l l years


i n the e a r l y 1980's water r e s t r i c t i o n s were introduced, several emergency
programmes initiated, and the study of available groundwater sources
intensified .
The main source of groundwater is in weathered zones, cavities and
fissures of dolomite deposits, outcrops of which occur in a wide circle
around Johannesburg (Figure 9.1 1. These deposits, approximately one
kilometer thick, d i p gently away from the centre a t a slope of some ten
degrees. To the South dolomite o v e r l a y s the quartzites and small pebble
conglomerates of Black Reef Series, the Ventersdorp and W i twatersrand
geological System, the l a t t e r renowned f o r i t s a u r i f e r o u s reefs, w h i c h were
at one time extensively mined on the Witwatersrand. The shales and
quartzites of the Pretoria Series overly these series, and vertical
intersections b y syenite dykes create a number of virtually independent
groundwater compartments. Since some of the r e s u l t i n g compartments have
been sanctioned for dewatering in order to allow safer mining, the
groundwater levels i n these compartments have dropped considerably. Other
compartments have not been dewatered because of the danger of sinkholes
forming on the surface. The outcrops of dolomite a r e g e n e r a l l y covered b y
a l l u v i a l deposits of v a r y i n g depth.
Published studies on a r t i f i c i a l recharge by infiltration with partially
treated wastewater a r e almost entirely conducted in p r i m a r y aquifers of
considerable depth, u n d e r l a i n b y an impermeable layer. Important aspects
of the performance of these schemes a r e the permanence of the i n f i l t r a t i o n
r a t e and the reduction i n contaminants b y b a c t e r i o l o g i c a l processes a n d b y
adhesion to soi I p a r t i c l e s . Successful p r o j e c t s a r e r e p o r t e d from the U.S.A.
(Bouwer et a l . , 1980), I s r a e l ( l d e l o v i t c h and M i c h a i l , 1984) and A u s t r a l i a
(Mathew et al., 1982). Careful selection of the infiltration site and
extensive testing of the p u r i f y i n g c a p a c i t i e s of the combined p r i m a r y and
secondary aquifer in the Witwatersrand area will have to precede a
decision on the q u a l i t y of the i n f i l t r a t i o n water.
An actively operated groundwater reservoir wil I create a fluctuating
groundwater t a b l e and increased groundwater flow velocities. Therefore a n
increased a c t i v i t y of s o l u t i o n a n d erosion processes i n the dolomite may be
expected, in some cases followed by the occurence of sinkholes and
subsidence. Built-up areas located over some of the dolomitic areas leave
certain compartments u n s u i t a b l e for the envisaged scheme. Simulation of
groundwater movements a n d m o n i t o r i n g of the n a t u r a l freshwater/wastewater
interfaces w i l l have to be done.
Artificial recharge not only increases the available amount of
Fig. 9.2 S e w a g e w o r k s and r e c h a r g e s i t e s
145

groundwater, it also enables, by its more stable supply, to develop


schemes for groundwater use in conjunction with surface waters. A
s i m p l i s t i c model ( P a l i n g , 1985) indicates t h a t b y c o n j u n c t i v e use o f surface
a n d groundwater the minimum guaranteed d r a f t can b e increased b y 10% of
the total s u p p l y compared w i t h the s i t u a t i o n in w h i c h each source s u p p l i e s
individually.
Sewage treatment in the Witwatersrand a r e i s s t i l l m a i n l y o r g a n i z e d on
a municipal basis. Johannebsurg operates three works South of the
watershed which divides the town in two distinct areas as far as
wastewater collection is concerned, with a combined effluent output of
approximately 500 M l / d (megalitres p e r day). Sewage from surrounding
m u n i c i p a l i t i e s i s processed i n works w i t h c a p a c i t i e s a s small as 6 Ml/d.
East of Johannesburg a p o t e n t i a l a r t i f i c i a l recharge scheme c o u l d involve
the works located in the m u n i c i p a l i t i e s of Germiston, Boksburg, Benoni,
B r a k p a n and S p r i n g s ( F i g u r e 9.2 and Table 9.1).

Table 9.1 Sewage works a n d Recharge site.

No. Name Discharge ( M l / d ) E l e v a t i o n (m)


a v e r a g e 1984

1 Davey ton 8 1600


2 Mc Comb 6 1580
3 Jan Smuts 11 1610
4 Rynfield 11 1620
5 Ancor 30 1580
6 Benon i 15 1640
7 Mapleton -- 1580
8 Dekema 56 1530
9 Rondebul t 39 1560
10 Vlakp laas 44 1520

It has been historical p r a c t i c e to discharge the effluent in streams


that form t r i b u t a r i e s of the' Vaal River upstream o f the intake of the
local Rand Water Board. I n d i r e c t reuse in t h i s form h a s been p r a c t i s e d i n
the Witwatersrand area since 1923 and takes place after alleged self
p u r i f i c a t i o n in streams and reedbeds, and after d i l u t i o n with fresh river
water. I n t e n s i v e m o n i t o r i n g of the streams by the Rand Water Board in
146

recent years h a s i n d i c a t e d a considerable contamination o f these streams,


mainly by industrial discharge and leaching of the numerous goldmine
dumps.
Dolomite compartments suitable for artificial recharge and abstraction
are limited. Dewatered compartments in the gold mining area and
compartments with built-up a r e a s on top have been mentioned. Another
impediment c o u l d be t h e p e r c o l a t i o n from contaminated streams.
Effluent from established sewage water works would have to be
t r a n s p o r t e d to the dolomite compartment by pipeline. Canals might have
been a cheaper alternative, b u t closed pipes are prefered for sanitary
reasons. For sewage works i n close p r o x i m i t y to each o t h e r in comparison
w i t h the distance to a n y seepage a r e a combined c o n d u i t s would probably
be most economic whereas o t h e r sources may j u s t i f y seperate conduits. The
problem of minimizing total conveyance cost could be considered as a
network to be optimized. I n the W i t w a t e r s r a n d s i t u a t i o n clearly distinct
clusters are discernable, which enabled the complexity of the network
model to be reduced to a minimum. Although r e l o c a t i o n of e x i s t i n g sewage
treatment works cannot be justified, future new works may be more
economically s i t e d over the seepage areas. T h i s would use the a d v a n t a g e
of economy of scale, b u t may increase conveyance cost due to the g r e a t e r
peak to average flow r a t i o f o r r a w sewage.
Integer network programming has found several applications in the
f i e l d of sewage conveyance. Wanielista a n d Bauer (1972) a p p l i e d i t to the
c e n t r a l i z a t i o n of wastewater treatment facilities in the Econ R i v e r basin
Florida, Leighton a n d Shoemaker (1984) used it in a similar fashion for
the r e g i o n a l i z a t i o n of wastewater c o l l e c t i o n and treatment systems in Long
I s l a n d , New York.

COST ANALYSIS

P i p e diameters were based on maximum p e r m i s s i b l e flow v e l o c i t i e s decided


for p r a c t i c a l reasons. The next larger commercially available pipe size
was selected in each case. Based on a cost p e r metre in f i g u r e 9.3 and
w i t h the length of the p i p e l i n e section known, the t o t a l costs f o r supply
a n d construction may be c a l c u l a t e d .
For purposes of estimating pumping heads, the friction head was
calculated using the Darcy equation. The costs f o r pumps was estimated
w i t h the r e l a t i o n s h i p s i n Table 9.2. The equations were based on d a t a f o r
the range Q = 10 - 50 Ml/d. For intermediate heads the costs are
interpolated.
147
D
0
N
0
0
E
0
D
E
0
z
D
N
D
z
D
m
0
0
W
0
0
0
0
0
N
0
w
c)
V
. -
0 0 0 0 0 0 0
- 0 N
0 0 OD
0 0
L n w
0 0
N
m u
0 E- c
.r
148

TABLE 9.2 Pump costs

Head Cost (Q i n Ml/d)


(m 1 ($1

20 207.5 * Q + 4625
40 217.5 * Q + 4725
60 230.0 * a + 5800

The costs for the pump motors were based on the shaft power,
increased b y a 20% safety m a r g i n (see Table 9.3).

TABLE 9.3 Motor costs

Motor Power Voltage costs


(kW) (V) ($/kW

0 - 250 400 60
250 - 1900 1000 - 3300 70
1900 - 6570 6600 - 11000 80

The number of i n s t a l l e d pump sets depends on the flow to be h a n d l e d


a n d the c o n t i n u i t y of t h i s flow volume. For the present calculations one
b a s i c a n d one back-up set of pumps was h e l d to be s u f f i c i e n t for flows
under 50 M l / d a n d two b a s i c a n d one back-up sets f o r b i g g e r flows.
F i n a l l y the costs f o r p i p e l i n e , pumps a n d motors are added up. This
procedure i s repeated f o r each of the 113 v a r i a b l e s p e r c l u s t e r .
149

MATHEMAT I CAL FORMULAT ION

In the papers by Wanielista and Bauer (1972) and Leighton and


Shoemaker (1984) r e g i o n a l wastewater conveyance to c e n t r a l i z e d treatment
plants is dealt with in one comprehensive computer model. For the
W i twatersrand a r e a e f f l u e n t conveyance to recharge sites can successfully
be broken down in a number of subsystems. A network of the
Witwatersrand was developed that could incorporate a l l the subsystems. In
p a r t i c u l a r the most complex subsystem comprises the s i x treatment works in
the Benoni, Brakpan and Springs municipalities (node 1 to 6) and a
proposed infiltration area at Mapleton (node 7 ) . The network with its
possible flow d i r e c t i o n s i s presented in f i g u r e 9.4.
I t i s the subsystems which could p o s s i b l y s u p p l y the Mapleton a q u i f e r
which a r e considered here. At design stage i t i s u n c l e a r whether e f f l u e n t
should be conveyed from node 3 to node 5 o r the o t h e r way r o u n d in o r d e r
to minimize the costs. Both options a r e therefore l e f t f o r possible selection
in the optimization procedure. The same considerations apply to the link
between nodes 3 and 6. The i n t r o d u c t i o n of two p o s s i b l e flow directions
excludes the a p p l i c a t i o n of a dynamic programming approach as used b y
Smith et a l . , (1983). On the other h a n d integer programming can be

-
a p p l i e d to solve t h i s problem i n the f o l l o w i n g way.

Rynfield 0---------
--------ODaveyton
,'; \, -0' i

MC Comb

Maple1:on

F i g . 9.4 Network w i t h possible flow d i r e c t i o n s


150

Each plant has a certain design discharge. Every pipeline section


o r i g i n a t i n g from a p l a n t must a t least be a b l e to convey this discharge.
As most l i n k s a r e supposed to convey e f f l u e n t from o t h e r p l a n t s a s w e l l , a
series of flow r a t e s can be defined as a r e s u l t of d i f f e r e n t combinations o f
d i s c h a r g e volumes. By r e p r e s e n t i n g each flow r a t e in each p i p e l i n e section
b y a seperate i n t e g e r decision v a r i a b l e the network o p t i m i z a t i o n c a n then
be formulated a s a n i n t e g e r programming problem. The number of v a r i a b l e s
depends on the number of network nodes as well as on the number of links
w i t h t h e i r specific d i r e c t i o n . In the present example the network can be
described b y 113 v a r i a b l e s .
The flow rate dictates under certain conditions mentioned below the
required pipe diameter and pump capacities. Subsequently the costs
connected with these requirements can be determined. The objective
function can be expressed as the sum of the products of the integer
v a r i a b l e s and the r e l a t e d cost factors, or
Objective f u n c t i o n = CCi * Xi
The cost factors Ci are calculated for each flow volume and each
p i p e l i n e section as i n d i c a t e d i n the cost a n a l y s i s . A separate p r o g r a m was
developed to c a l c u l a t e thse cost factors, based on the input of (1) the
maximum flow volume from each node, ( 2 ) the e l e v a t i o n of each node, (3)
the l e n g t h of each l i n k , ( 4 ) the maximum p e r m i s s i b l e flow v e l o c i t y a n d ( 5 )
the e f f e c t i v e p i p e roughness. P r o v i s i o n was made to e l i m i n a t e c e r t a i n links
o r s u p p l y nodes i n o r d e r to m a i n t a i n the f l e x i b i i t y r e q u i r e d to a d j u s t the
network for other clusters of treatment plants. The cost calculation
program was extended in such a way that the format of the output
program makes it immediately suitable for submission for optimization,
u s i n g the I .B.M. mixed i n t e g e r programming p a c k a g e MPSX-370/MIP.
The c o n s t r a i n t m a t r i x i s based on two simple p r i n c i p l e s which exploit
the essential f e a t u r e of decision v a r i a b l e s :
a) Y - C Xi 5 zero
T h i s expression i s used to describe the relationship between different
variables. I f Y = 1 at least one X. must b e present. I n t e r a c t i o n s of this
t y p e a r e m a i n l y of a p r o g r e s s i v e n a t u r e , but some r e g r e s s i v e steps h a d to
be i n c l u d e d e.g. i f l i n k 4-6 conveys Q1 + Q4, the l i n k s 2-3, 2-5 and 2-7
may convey o n l y Q2.
b) L X . = zero o r . u n i t y .
For the r i g h t h a n d side t o equal one, only one variable assumes the
v a l u e one. I f the r i g h t h a n d side i s set a t zero, all v a r i a b l e s must be
zero. I f a l l p o s s i b l e flows from a particular node a r e grouped together
a n d then set a t one, o n l y one flow w i l l be selected.
151

The combination of these two expressions guarantee unique paths


through the network a n d together w i t h the o b j e c t i v e f u n c t i o n the p a t h w i t h
the lowest costs w i l l be found.

RESULTS

Optimal effluent network configurations were computed for a potential


artificial recharge scheme c a r r i e d around Mapleton. A visual inspection
appeared to i n d i c a t e the minimum total pipe length was a s estimated in
f i g u r e 9.5a. lising the flow volumes in table 9.1 and a maximum flow
velocity of 2.2 m/s the total construction costs for this selected
configuration would be $13.4 million. Under the same conditions the
Optimized c o n f i g u r a t i o n shown in f i g u r e 9.5b would cost $11.2 m i l l i o n .
The r e l a t i o n between flow r a t e and p i p e diameter is Q = v ( ~ / 4 D
) 2.
Since the p i p e diameter i s thus i n v e r s e l y p r o p o r t i o n a l to the s q u a r e root
of the flow velocity, a reduction of the maximum velocity to 1.2 m/s
r e s u l t e d i n a minimum construction cost of $17.5 million (figure 9 . 5 ~ ) .
The o r i g i n a l problem c o u l d be expanded slightly by the inclusion of
sources 8 , 9 a n d 10. F i g u r e 9.6a shows the s i t u a t i o n in which two c l u s t e r s
supply individually to one recharge site. Each cluster is optimized
separately, r e s u l t i n g in a cost of $11.2 m i l l i o n f o r the n o r t h e r n c l u s t e r as
indicated previously, and $5.1 million for the western cluster, bringing
the t o t a l to $16.3 m i l l i o n .

F i g . 9.5 Network o p t i m i z a t i o n . Node 1 to 6 represent sewage works


and node 7 the i n f i l t r a t i o n site. See also T a b l e 9.1 a n d
f i g u r e 9.1
152

F i g . 9.6 L i n k a g e of networks. Node 1 to 6 a n d node 8 to 10 represent


sewage works a n d node 7 the i n f i l t r a t i o n site. See a l s o
Table 9.1 a n d f i g u r e 9.1

An a l t e r n a t i v e arrangement which may save cost i s to supply effluent


from the northern cluster via node 9 (see f i g u r e 9.6b). The change in
i n p u t d a t a f o r the n o r t h e r n c l u s t e r i n v o l v e s f i v e new v a l u e s f o r the l e n g t h
of the l i n k s between node 2, 3, 4, 5, 6 on the one h a n d a n d node 9 on
the other, resulting in a cost of $11.1 million. The fact that the
c o n f i g u r a t i o n of the optimal network remains the same as i n f i g u r e 9.6a is
coincidental.
The cost optimization of the western cluster is performed after
increasing the flow via node 9, resulting i n an optimized cost of $6.1
million for this cluster. The total cost thus amounts to $17.2 million.
Hence the lay-out of the sewage works i n t h i s case i s such that a linkage
of the two c l u s t e r s does not r e s u l t i n any f u r t h e r cost reduction.

+' nodes 1 , 2

nodes 1 , 2 , 4 , 5 , 6 - supply points


node 7 - demand p o i n t
node 3 - l o c a t i o n of booster station
node 3' - alternative s i t e for
booster stati a,

F i g . 9.7 Optimum location f o r a booster s t a t i o n


153

SUMMARY AND CONCLUS IONS

Integer programming p r o v e d to be a useful tool f o r the e v a l u a t i o n o f a


least-cost configuration of a p i p e l i n e network. The support p r o g r a m was
designed to suit a network with seven nodes and rather complex
interrelations. It takes approximately half a second on an I.B.M. 3083
computer to c a l c u l a t e the cost factors and to compile a p r o g r a m s u i t a b l e
f o r integer optimization. For a c l u s t e r of s i x s u p p l y nodes a n d one demand
node optimization r e s u l t s were obtained i n less t h a n 30 seconds u s i n g the
I .B.M. mixed integer programming package MPSX-37O/MIP. Complicated
cases can be handled by subdividing a system into subsystems and
o p t i m i z i n g each c l u s t e r i n d i v i d u a l l y .
The f l e x i b i l i t y of the program enables one to i n t r o d u c e changes w i t h
minimal e f f o r t a n d to compare the r e s u l t i n g a l t e r n a t i v e s .
This feature may be illustrated by the follohing example of an
economic o p t i m i z a t i o n of the location of a booster s t a t i o n ( f i g u r e 9.7).
By v a r y i n g the d i s t a n c e between node 3 a n d the other f i x e d nodes a
sequential search w i l l r e s u l t i n the optimum solution. The "range" facility
o f MPSX-370/MIP can f o r each run g i v e a sensitivity analysis and thus
prevent the search from becoming a random process. The general
applicability i s accompanied b y the d i s a d v a n t a g e that no p r o v i s i o n s a r e
made f o r i n t r o d u c i n g the costs f o r obstacles l i k e r o a d s a n d r i v e r s .
Computer analysis of the most economic pipeline configuration for
sewage effluent conveyance to an artificial groundwater recharge site
resulted in a cost r e d u c t i o n of 16% over an optimum solution based on
visual inspection. A lower flow velocity increased the total construction
costs as the influence of a n increased p i p e diameter s t r o n g l y outweighs the
reduced cost in pumping equipment. Depending on the lay-out of the
sewage works a f u r t h e r r e d u c t i o n in the combined costs c o u l d p o s s i b l y be
a t t a i n e d b y i n t e g r a t i o n of i n d i v i d u a l clusters.

REFERENCES

Bouwer, H., Rice, R.C., Lance., J.C., and Gilbert, R.G., 1980.
R a p i d - i n f i l t r a t i o n Research a t F l u s h i n g Meadows Project, Arizona. J.
Water Polut. Control Fed., Vol. 52, No. 10, p 2457.
I d e l o v i t c h , E. and M i c h a i l , M., 1984. S o i l - a q u i f e r Treatment - A New
Approach to an O l d Method of Wastewater Reuse. J. Water P o l l u t .
Control Fed., V o l . 56, No. 8, p 936.
Leighton, J.P. and Shoemaker, C.A., 1984. An I n t e g e r Programming
Analysis of the Regionalization of L a r g e Wastewater Treatment a n d
Collection Systems. Water Resources Research, Vol. 20, No. 6, p 671.
154

Mathew, K., Newman, P.W.G. and Ho, G.E., 1982. G r o u n d w a t e r R e c h a r g e


w i t h Secondary Sewage E f f l u e n t . A u s t r a l i a n Water Resources C o u n c i l ,
Technical P a p e r No. 71, C a n b e r r a .
P a l i n g , W.A.J., 1984. O p t i m i z a t i o n of C o n j u n c t i v e Use of G r o u n d w a t e r and
S u r f a c e Water Resources in t h e Vaal B a s i n . Proceedings o f t h e H a r a r e
Symposium, IAHS P u b l . No. 144.
P a l i n g , W.A.J., 1985. Economic O p t i m i z a t i o n of A l t e r n a t e Water Resources
f o r the W i t w a t e r s r a n d . Water Systems Research Programme, U n i v e r s i t y o f
the W i t w a t e r s r a n d , Report No. 4/1985.
P a l i n g , W.A.J. and Stephenson, D., 1985. I n t e g e r P r o g r a m m i n g of T r e a t e d
Wastewater Conveyance f o r A r t i f i c i a l Recharge o f a n A q u i f e r . J. Int.
SOC. Ecol. M o d e l l i n g ( 7 ) .
Smith, A.A., H i n t o n , E. a n d Lewis, R.W., 1983. C i v i l E n g i n e e r i n g Systems,
A n a l y s i s and Design, John Wiley G. Sons.
W a n i e l i s t a , M.P. and Bauer, C.S., 1972. C e n t r a l i z a t i o n of Waste Treatment
F a c i l i t i e s , J. Water P o l l u t . Control Fed., V o l . 44, No. 12, p 2229.
155

CHAPTER 10

OPTIMAL PLANNING OF REGIONAL WASTEWATER TREATMENT

I NTRODUCT ION

The Witwatersrand o r 'White Waters Ridge' i s a n elevated r i d g e in the


h e a r t of South A f r i c a formed by the gold-bearing quartzite rock strata
emerging above the surface of the s u r r o u n d i n g c o u n t r y . It i s where gold
was originally discovered in 1886 and this sparked the growth of
secondary industry and tertiary commercial development. There are now
about f o u r m i l l i o n i n h a b i t a n t s of the area. The present water consumption
which i s s u p p l i e d b y the Rand Water Board, averages 1700 Ml/day and i s
i n c r e a s i n g a t a r a t e of 6 p e r cent a year.
The Witwatersrand r u n s i n a n east-west d i r e c t i o n a n d forms a n a t u r a l
watershed (Fig. 10.1). It is the source of the number of streams and
o r i g i n a l l y there were many s p r i n g s a l o n g the r i d g e , hence the name 'White
Waters R i d g e ' . The water supply to Johannesburg was o r i g i n a l l y pumped
from the ground as the water resources of the a r e a were not p l e n t i f u l (on
account of the fact that the elevated ridge was an origin of streams
r a t h e r than a r i v e r b a s i n ) . The streams to the n o r t h form tributaries of
the Limpopo R i v e r , a n ephemeral r i v e r w h i c h flows eastwards to the I n d i a n
Ocean. The streams to the south flow i n t o the Vaal River, a tributary of
the Orange R i v e r , which flows westwards to the A t l a n t i c Ocean.
It i s from the Vaal River that the W i t w a t e r s r a n d d r a w s most of its
water. The Vaal Barrage, 50 km south of the Witwatersrand, was
constructed in 1923, followed by the Vaal Dam in 1933. The combined
r e l i a b l e sustained y i e l d of these sources is 3300 Ml/day and the Rand
Water Board h a s r i g h t s to 2400 Ml/day. Some water must also be passed on
to downstream users. The Vaal, or 'murky' r i v e r as i t i s translated is a
flashy r i v e r a n d c a r r i e s much s i l t (170 mg/l on the a v e r a g e ) beyond the
Vaal Dam. On the other hand it has relatively little dissolved salts
( a v e r a g e 100 m g / l ) o r i g i n a t i n g m a i n l y from s a l t s leached from farmlands.
The Vaal River is now being supplemented by water diverted from the
r e l a t i v e l y untapped Tugela R i v e r , 300 km away. P l a n s were considered f o r
t a p p i n g the Orange R i v e r a t i t s source a n d d i v e r t i n g these waters to the
Vaal b a s i n . These d i v e r s i o n schemes a r e expensive.
On the other h a n d wastewater treatment technology i s now advancing
r a p i d l y and the cost of treatment i s a p p e a r i n g more a t t r a c t i v e . Since o n l y
about 50 p e r cent of water i s used consumptively on the Witwatersrand,
f
157

there i s scope f o r water reclamation and recycling. In f a c t this i s now


happening indirectly. Most water returned to the sewers on the
Witwatersrand finds i t s way, a f t e r treatment, to streams which discharge
to the Vaal Barrage. The Rand Water Board (1977) pumps 1000 M l / d a y from
the Barrage. Most of the b a l a n c e of i t s supply i s taken d i r e c t l y from the
Vaal Dam through a p i p e l i n e , which may be supplemented i n the f u t u r e b y
a canal leading from the dam wall to the Zuikerbosch pumping station.
Some of the e f f l u e n t from the Witwatersrand which finds i t s way to the
Vaal Barrage a n d i s not r e t u r n e d to the Witwatersrand b y the Rand Water
Board, i s consumed b y communities f u r t h e r downstream. The q u a l i t y of the
effluents entering the Vaal Barrage thus affects the cost of treatment
before f u r t h e r use of the water i s possible.
A number of possible schemes to re-use the wastewaters of the
Witwatersrand a r e possible:

1. P a r t i a l treatment of waste water treatment p l a n t s on the Witwatersrand,


and return of the effluent to the Vaal Barrage via streams, where
f u r t h e r p u r i f i c a t i o n occurs. The e f f l u e n t s a r e d i l u t e d b y r e l a t i v e l y p u r e
r i v e r water then, a f t e r re-treatment, pumped back to the W i t w a t e r s r a n d
and/or passed dOWnStream of the Barrage.
2. Convey wastewaters from the Witwatersrand to the banks of the
Barrage, a n d p u r i f y them a t a combined works before r e - c y c l i n g .
3. Reclaim e f f l u e n t on the Witwatersrand to a sub-standard and re-cycle it
i n a separate d i s t r i b u t i o n system f o r non-hygienic purposes.
4. Reclaim to a high standard and re-cycle together with the water
pumped from the Vaal R i v e r .
5. Install a low-capital, h i g h operating-cost, reclamation f a c i l i t y on the
Witwatersrand and m a i n t a i n t h i s as a standby in case of n a t u r a l r i v e r
droughts. Draw from the Vaal River a much h i g h e r d r a f t than could
reliably be drawn, and use the reclamation plant when shortfalls
occur.
6. Install low-capacity reclamation facilities on the Witwatersrand and
discharge the p u r i f i e d e f f l u e n t i n t o storage dams; e i t h e r on the surface
o r underground. Draw on the Vaal R i v e r to a h i g h degree as for (5)
and use the stored e f f l u e n t when s h o r t f a l l s occur in the Vaal R i v e r .
7. Pass partly treated effluents downstream of the Vaal Barrage in
constructed condu i t s.

Past p l a n n i n g and construction of waste water treatment facilities has


been on a local municipal basis, whereas bulk water supply was the
158

r e s p o n s i b i l i t y of the r e g i o n a l Rand Water Board. The establishment of a


r e g i o n a l waste water a u t h o r i t y i s now b e i n g contemplated. This w i l l enable
comprehensive planning at least cost to be achieved. Regional treatment
facilities and combined sewage o u t f a l l s can be planned with consequent
s a v i n g s i n cost due to scale. The location of treatment facilities can be
selected to result in least overall cost i.e. of sewers, wastewater
treatment, and water s u p p l y .
I t i s w i t h these ideas i n m i n d t h a t a p r o j e c t to s t u d y the water s u p p l y
a n d waste water system of the Witwatersrand was embarked on.

THE MATHEMAT I CAL MODEL

The system of dams, streams, wastewater treatment plants, potable


water works and c o n d u i t s between the W i t w a t e r s r a n d a n d the Vaal River is
g r o w i n g more complex over the years. As p o i n t e d out, a regional planning
authority responsible for overall planning is desireable. Such a body
should have at its disposal data assimilation facilities and systems
a n a l y s i s models to f a c i l i t a t e p l a n n i n g . A s u i t e of computer programs for
optimizing planning of wastewater treatment works, outfall sewers and
water works should be a t hand.
The system i s described l a t e r b y equations a n d constraints which are
in some cases not linear, a n d n o n l i n e a r programming methods a r e needed
to a r r i v e a t a least-cost plan. A s i m p l i s t i c mathematical model of p a r t o f
the system i s assembled below, and methods o f s o l u t i o n a r e o u t l i n e d l a t e r .
The chapter goes on to describe a method of linking neighbouring
basins by a master program, which could also consider various time
horizons. The set of c o n s t r a i n t s developed below i s f o r s t a t i c c o n d i t i o n s in
a p a r t i c u l a r b a s i n . S t a t i s t i c a l l y averaged v a l u e s of flows, water q u a l i t i e s
and consumptions are taken. To allow for variations by probability
distributions would be theoretically possible but would increase
computat iona I time man yfo Id.
Consider the simplistic system in F i g u r e 10.2. The diagram embodies
the f o l l o w i n g concepts: the water requirements of a m a j o r consumer such a s
the Witwatersrand ( 3 ) c o u l d be met from s u r f a c e resources (11, partially
t r e a t e d wastewaters r e t u r n e d to the r i v e r a t ( 2 ) o r reclaimed waste water
from (5). The wastewaters from (3) could be treated a t ( 4 ) followed b y
t e r t i a r y treatment a t ( 5 ) o r d i s c h a r g e d i n t o the r i v e r a t (6) after limited
treatment. Reclaimed water is assumed to be circulated in the same
d i s t r i b u t i o n systems as r i v e r water. The problem i s to determine what each
f l o w should be a n d what s t a n d a r d of treatment i s d e s i r a b l e .
159

SUE-SKTEM I

'. \
\
Fig. 10.2 Water c i r c u l a t i o n d i a g r a m

Each source, consumer, treatment p l a n t o r j u n c t i o n i s r e f e r r e d to as a


node. The v a r i a b l e s a r e the flow r a t e s between the nodes numbered i n the
d i a g r a m and the respective p o l l u t a n t concentrations, designated Q. . and
I-J
Pi-j respectively for flow from node i to node j. The flows are all
expressed i n Ml/day.
The p o l l u t i o n load may be conservative such as total dissolved solids
(TDS) o r a non-conservative v a r i a b l e such as biochemical oxygen demand
(BOD). I n the former case there must be a mass b a l a n c e of pollutant in
the system and i n the l a t t e r case the effect of the r i v e r a n d b a r r a g e s i n
d i l u t i n g the p o l l u t a n t must be assessed. BOD i s considered i n the present
model. Values of BOD a r e expressed i n mg/l and the product of flow in
M l / d a y and p o l l u t i o n in mg/l i s p r o p o r t i o n a l to the t o t a l load in tons/day
discharged b y a stream o r conduit.
The object of the study i s to minimize the cost of the system. It is
convenient to convert a l l costs to a common basis, say a n n u a l i n t e r e s t a n d
redemption on c a p i t a l cost p l u s r u n n i n g costs, a l l expressed in cents p e r
kilolitre. Cost coefficients, or rates, are therefore required for each
variable. Some of the cost coefficients, for instance for conveyance of
water i n closed conduits, a r e n o n l i n e a r a n d must be approximated b y the
anticipated incremental costs. Although nonlinear objective functions are
theoretically feasible, it is necessary to simplify the model as far as
possible since there a r e f u r t h e r n o n l i n e a r c o n s t r a i n t s as w i l l be r e v e a l e d
f u r t h e r on. I f there a r e e x i s t i n g c o n d u i t s these w i l l e f f e c t i v e l y h a v e zero
capital cost. But if the f u t u r e flow along a route is likely to exceed
existing capacity, i t i s o n l y the incremental cost of the new c o n d u i t s a n d
works which need to be considered. Allowance f o r peak f a c t o r s i s made in
s i z i n g the v a r i o u s works.
160

Associated with each flow rate Q. . is a conveyance cost coefficient,


I-J
which i s designated Ci. The v a r i a b l e p o r t i o n of the cost component which
i s a f u n c t i o n of flow o n l y i s t h u s

‘1 ‘1-2
‘2’2-3
+ ’ ‘3‘3-6 + ‘4‘4-3 (10.1)
Note that Q6-7 is fixed so its cost is not variable i.e. need not be
considered. The cost of conveyance i n n a t u r a l channels i s zero.
P u r i f i c a t i o n costs comprise a component p r o p o r t i o n a l to flow Q . . and a
I-J
component proportional to pollutant load Q. .P.
I-] I-J
. or proportional to
pollutant load removed, Q . . ( P . .-P.
I-J I-J
. )
l-J/2
where subscript 2 refers to
conditions after treatment. The l a t t e r component, i.e. cost p r o p o r t i o n a l to
load removed, i s d i f f i c u l t to e s t a b l i s h a n d i n fact i t i s often assumed that
works a r e designed to produce a n effluent of reasonable standard with
treatment costs p r o p o r t i o n a l to flow r a t e . Vie w i l l consider the general case
a n d designate the coefficients of Q
3-6 P3-4 a n d Q2-6 P2-3 as C5 and C6
respectively. Since Q3-6 + Q3-4 is a constant, the flow-proportional cost
can b e omitted.
The equations o r c o n s t r a i n t s d e s c r i b i n g the system a r e formulated next.
Hall (1977) formulated the system w i t h similar constraints, but at that
time was unaware of a simple method of solution.
For flow b a l a n c e a t the v a r i o u s nodes:
At some source nodes such as (1 1, the yield may be limited but in our
case we consider u n l i m i t e d augmentation possible, a t a cost. At consumer
nodes the s u p p l y must be s u f f i c i e n t :

42-3 + Q4-3 = a1 (10.2)


‘6-7 = a2 (10.3)
At consumer nodes the wastewater o u t p u t i s known:

‘3-4 + ‘3-6 = a3 (10.4)


At treatment p l a n t s a n d o t h e r nodes the flows must balance:

Q3-4 - 44-2 - Q4-3 = 0 (10.5)

‘1-2 + ‘4-2 - ‘2-3 - ‘2-6 = (10.6)


‘2-6 + ‘3-6 - ‘6-7 - ‘6-8 = (10.7)
Note t h a t a l t h o u g h the v a r i a b l e Q6-7 could be e l i m i n a t e d using equation
(10.3) i t complicates the cost function and any later changes to the
constants which may b e desired. The number of variables i s nevertheless
minimized i n simple cases b y s u b s t i t u t i n g Q f o r Q4-5 a n d a n d Q3-6
4-3
for Q3-9 and Qg-6. It will be observed that the so called constraints
(10.2)-(10.7) are in fact all equations. They could equally well be
c o n s t r a i n t s of the less-than or greater-than type, in which case slack
v a r i a b l e s would be i n t r o d u c e d to form equations. The waste, Q6-8, i s i n
fact a slack variable, but w i t h more meaning than a purely algebraic
161

slack variable. I n equation (4) the waste water output is given as a


constant. This constant i s in fact a f u n c t i o n of the consumption a2, but i t
i s easier to i n s e r t a constant.
There a r e other forms of c o n s t r a i n t s on the flow v a r i a b l e s which c o u l d
be incorporated. For instance b a s i n c h a r a c t e r i s t i c s may l i m i t the amount of
waste water C13-6 which could (economically) be discharged beyond the
barrage (10.2). O r i f the s u p p l y C12-3 was considered as two components;
one through e x i s t i n g c o n d u i t s a n d one t h r o u g h new conduits, there would
be a l i m i t on the c a p a c i t y of the e x i s t i n g conduits.
The next set of c o n s t r a i n t s a p p l i e s to the p o l l u t a n t s . Certain l e v e l s of
p o l l u t i o n may be known:
P I - 2 = a4 (10.8)
I n fact i t i s assumed f o r the BOD study t h a t P1-2 = 0.
There may be t o l e r a b l e l i m i t s on c e r t a i n l e v e l s o f p o l l u t i o n :
‘4-2 5 a5 (10.9)
‘2-3 5 a6 (10.10
‘6-7 5 a7 (10.11
There i s an extent of n a t u r a l p u r i f i c a t i o n i n r i v e r s :
(10.12
’4-2 - ’4-2/2 = a8
P2-3 - P2-3/2 = ag ( f o r f l o w 2 - 6) (10.13)
I n the case of TDS as the p o l l u t a n t there would b e n e g l i g i b l e r e d u c t i o n
of P a t waste water treatment p l a n t s a n d the TDS a f t e r reclamation p l a n t s
could be taken as zero. i n the case of BOD, i t can be assumed P1-2 = 0
a n d P5-3 = 0, and there i s some reduction i n P a t 4 a n d 9:
‘3-4 - ‘4-2 = (10.14)
(10.15)
‘3-4 - ‘9-6 = all
A mass balance of p o l l u t a n t s must be m a i n t a i n e d a t nodes:

(10.16)

(10.17)

Q3-4P3-4+Q3-6P3-4-Q2-3P2-3-Q4-3p5-3 = al
2 (10.18)
Note P2-6 e q u a l s P2-3, P4-5 equals P4-2 a n d P3-9 e q u a l s P3-4 so these
s u b s t i t u t i o n s a r e made f o r s i m p l i f i c a t i o n .
It is implicit in the optimization program that all variables are
non-negative a n d r e a l so these c o n s t r a i n t s a r e not s t a t e d e x p l i c i t l y .
The general model then is to minimize (10.1) subject to constraints
(10.2)-(10.17). A method of solution i s o u t l i n e d in the n e x t section.
162

OPTIMIZATION METHOD

It will be observed that the objective function and constraints are


linear except for constraints (16)-(18). These constraints involve
two-dimensional p r o d u c t s of v a r i a b l e s . There a r e techniques f o r c o n v e r t i n g
these products to separate functions, f o l l o w i n g w h i c h a technique known a s
separable programming may be employed to optimize the system.
Hadley (1964) proposed a simple method o f transforming a p r o d u c t QP
b y i n t r o d u c i n g two new v a r i a b l e s M a n d N, such t h a t

M = (Q + P)/2, N = (Q - P)/2 (10.19)


then

QP = M‘ - NZ (10.20)
and

Q = M + N , P = M - N (10.21)
M and N a r e u n r e s t r i c t e d i n s i g n b u t t h i s i s p e r m i t t e d in the d e l t a method
of separable programming (IBM, 1976).
The separable programming algorithm is based on the fact that a
separable f u n c t i o n can be approximated b y piecewise I i n e a r functions. The
p o l y g o n a l a p p r o x i m a t i o n i s represented b y a set of special variables, so
a n y v a l u e of M can be represented a s follows:
M = Mo + GIDl + G2D2 + ...Gk D k +...GRDR (10.22)
where the Ds represent i n t e r v a l s of M a n d the special v a r i a b l e s GI, ...,GR
a r e defined as follows:
f o r M in i n t e r v a l k ,

G , = G 2 = G k-1 = 1 (10.23)
O(Gk( 1 ( 10.24)
Gk+l = Gk+2,..GR = 0 (10.25)
i.e. M comprises a set o f i n t e g r a l i n t e r v a l s D u p to k - 1 p l u s a f r a c t i o n
of i n t e r v a l k .
Note that
M’ = Ma
0
G E+
1 1
+...
G E +... GRER
k k (10.26)
where each i n t e r v a l Ek corresponds to an interval Dk. Thus for the
approximations i n Fig. 10.3, Mo = 0, k = 4 a n d Gk = 0.3. The v a l u e of M
can be confined to a known range.
Although there i s a p o s s i b i l i t y of a t t a i n i n g a local optimum t h i s chance
i s reduced i f the problem i s solved w i t h the special v a r i a b l e s set initially
at their upper bound, a n d then w i t h them set at their lower bounds, to
verify the r e s u l t s . It will be observed that 2R variables are introduced
i n t o the model f o r each p r o d u c t in the o r i g i n a l c o n s t r a i n t s . It i s therefore
d e s i r a b l e to s i m p l i f y the o r i g i n a l system a s much a s p o s s i b l e to minimize
163

the number o f products.


The f o l l o w i n g s i m p l i f i c a t i o n s a r e introduced. The system i s subdivided
i n t o two subsystems (see F i g . 10.21, which will b e l i n k e d v i a a master
p r o g r a m a c c o r d i n g to D a n t z i g ' s decomposition p r i n c i p l e (Dantzig, 1963) a n d
a p p l i e d b y Stephenson (1970) to link r i v e r b a s i n s . O n l y sub-system 1 is
considered here and there will be shadow values imposed o n the cost
coefficients of Q P a n d Q2-6P2-3 b y the master p r o g r a m (C7 a n d C8).
3-6 9-6
(10.31, (10.71, (10.11 1, (10.13) a n d (10.17) a r e thereby e l i m i n a t e d .
E l i m i n a t i n g Pg-6 using (10.151, cost c o e f f i c i e n t s become C o g = C3 - al 1
C7,Cf5 = C5 + C7 a n d C I 6 = C6 + C8.
Certain variables , namely Plm2 a n d P5-3 a r e zero so c o n s t r a i n t (10.8)
i s eliminated, and by substituting from (10.12) and (10.4) into (10.16)
and (10.18) respectively, the number of products i s reduced. Now the
problem i s :

subject to:

0.28)
0.29)
0.30)
0.31)
0.32)
(10.33)
(10.34)
( 10.35)
( 10.36
(10.37)
(10.38)

Put

'4-2~4-2 = M - ~ ' t 1 (10.39)


Q2-3P2-3 = M i - N: (10.40)
Q2-6P2-3 = M - N: (10.41)

Q3-4p3-4
= M Z - Nt (10.42)
Then (10.37) a n d (10.38) c a n b e r e p l a c e d b y e q u a t i o n s (10.43)-(10.52):
164

(10.43
(10.44
(10.45
(10.46
(10.47)
(10.48)
(10.49)
(10.50)
(10.51)
(10.52)

Each M2 a n d NZ comprises composite p o l y g o n a l functions according to


equation (10.22) and it is necessary to define the intervals in the
computer d a t a i n p u t . The o b j e c t i v e f u n c t i o n must a l s o be r e - w r i t t e n as
min c1a1-2 + c2a2-3 + c * ~ Q +~ c4a4-3
- ~ + c; (a3P3-& - M: + N):
(10.53)

t /

Fig. 10.3 Polygonal a p p r o x i m a t i o n of a s e p a r a b l e f u n c t i o n

T h e problem may thus be solved by straightforward techniques, and


once the procedures are adequately programmed, the process may be
expanded, using the decomposition principle, to consider various time
horizons. The problem may also be considered i n more d e t a i l than outlined
here. Using decomposition principles it would be p o s s i b l e to incorporate
sub-programs for individual waste water treatment works (e.g. CIRIA,
1975). This would ensure optimization of each component. Dynamic
programming methods c o u l d be employed to study pollution along stream
reaches a n d optimize the s o a c i n g a n d s t a n d a r d of waste water p u r i f i c a t i o n
165

works d i s c h a r g i n g i n t o the stream. Where m u l t i p l e decisions a r e r e q u i r e d ,


f o r instance f o r a l t e r n a t i v e p u r i f i c a t i o n p l a n t locations, o r sewer o u t f a l l s ,
m i x e d integer p r o g r a m m i n g c o u l d be employed. Computer s i m u l a t i o n s o f the
system a s a l s o r e q u i r e d t o study extreme c o n d i t i o n s a n d cost s e n s i t i v i t i e s
t o supplement the shadow v a l u e s produced b y the o p t i m i z a t i o n . There are
o t h e r s o p h i s t i c a t e d techniques f o r o p t i m i z i n g n o n l i n e a r waste water systems
(Chiang and Lauria, 1977; Pratishthananda and Bishop, 1977) but the
above-formulated problem s i m p l i f i e d to a neat solution.
To incorporate a l l the concepts in a large number-crunching program
would not be realistic though, and interactive programming is
recommended, i.e. human i n t e r v e n t i o n a t each step. The p l a n n i n g process
c o u l d be continued as data are updated by successive iteration of the
master p r o g r a m a n d sub-programs.

REFERENCES

Chiang, C.H. a n d L a u r i a , D.T., 1977. H e u r i s t i c a l g o r i t h m f o r waste water


p l a n n i n g . Proc. Amer. SOC. Civ. E n g r s 103 No. EE5, 863-876.
CIRIA, 1975. Cost e f f e c t i v e sewage treatment - the c r e a t i o n of a n
o p t i m i z i n g model. C l R l A Report 46, London.
D a n t z i g , G.B., 1963. L i n e a r Programming a n d Extensions: P r i n c e t o n
U n i v e r s i t y Press, Princeton, New Jersey, USA.
H a d l e y , G., 1964. N o n l i n e a r a n d Dynamic Programming, pp. 448-465:
Addi son-Wesley , Reading.
H a l l , G.C., 1977. A method f o r o p t i m i z i n g the c i r c u l a t i o n o f water in
u r b a n regions. N a t l . I n s t . Water Res. Council f o r Scient. a n d I n d u s t .
Res. P r e t o r i a , R.S.A.
IBM, 1976 Mathematical Programming System Extended/370, P r o g r a m
Reference Manual, 2nd e d i t i o n , pp. 230-251.
P r a t i s h t h a n a n d a , S . a n d Bishop, A.B., 1977. A n o n l i n e a r m u l t i l e v e l model
f o r r e g i o n a l water resources p l a n n i n g . Waf. Resour. B u l l . 13 No. 3,
61 1-625.
Rand Water Board, 1977. Annual Report.
Stephenson, D., 1970. Optimum d e s i g n o f complex w a t e r resource projects.
Proc. Amer. SOC. Civ. E n g r s 96, no. HY6, 1229-1246.
Stephenson, D. 1978. Optimal p l a n n i n g of r e g i o n a l wastewater treatment.
Proc. IAHS Symposium. Model I ing the Water Qua1i t y o f the H y d r o l o g i c a l
Cycle. Baden, 125. 351-360.
166

CHAPTER 1 1

SIMULATION OF SEWER FLOW

I NTRODUCT ION

With the development of s u b u r b a n a r e a s w i t h i n c i t i e s to t h e i r limits i t


i s becoming necessary to consider s u b d i v i s i o n a n d more intense r e s i d e n t i a l
densities i n suburbs which were previously only sparsely populated. The
effect of more intense development on the services, such a s sewerage, for
a n area must be considered before such increase i n d e n s i t y i s permitted. A
study of the consequences of increased l o a d i n g i s d i f f i c u l t , p a r t i c u l a r l y as
the effect may cause a chain reaction down the length of the sewer
system. A method of identifying possible problem areas and methods of
p a s s i n g the increased flow was sought. A model for rapidly selecting a
p a r t i c u l a r area o r f o l l o w i n g the flows t h r o u g h a system was developed.
There i s frequently a n a p p r e c i a b l e time-lag as h y d r o g r a p h s flow down
the system as well as attenuation due to lateral dispersion of the
hydrograph, i.e. routing. I n o r d e r to a l l o w for these effects a computer
s i m u l a t i o n program seemed to b e the most logical approach. The program
can draw d a t a from e x i s t i n g l a n d use i n v e n t o r i e s wherein d a t a concerning
a l l stands w i t h i n the m u n i c i p a l area (i.e. floor areas, number o f rooms
and land usage type) are retained. In addition data files containing
engineering data (i.e. sewer lengths, slopes, connections, diameters, drops
and condition) are compiled. The latter data may be used for other
p r o j e c t s such as d a t a r e t r i e v a l for drawing sections, establishing depths
of connections a n d management o f the sewerage system a t a l a t e r stage. As
there a r e about 135 000 stands w i t h i n the Johannesburg a r e a a systematic
a n d e f f i c i e n t way of s t o r i n g the d a t a was r e q u i r e d f o r t h i s a p p l i c a t i o n .
The process of computerization also enables engineers to estimate
sewage design flows effectively. Whereas it was previously necessary to
design sewerage systems f o r the estimated peak flows based on averages
over ten hours of the d a y , it i s now possible to divide the flow into
different components. Stormwater ingress, infiltration through joints in
manholes, leaks from sanitary fittings and sewage discharges may be
accounted f o r separately. The a c t u a l time d i s t r i b u t i o n of sewage d i s c h a r g e s
may also be considered. This will affect the hydrograph lags where
successive hydrographs contribute to an outfal I. Considerable data
therefore h a d to be gathered in o r d e r to p r o v i d e the subdivision for the
a n a l y s i s a n d the development of c o n t r i b u t o r y h y d r o g r a p h s .
167

Considerable work has been done in South Africa, in particular by


Shaw (1963) a n d Crabtree (1976), on e s t a b l i s h i n g c o n t r i b u t o r h y d r o g r a p h s .
The a c t u a l l a g g i n g of h y d r o g r a p h s a n d consideration of r o u t i n g effects a n d
p r o b a b i l i t i e s of d i f f e r e n t connections discharging simultaneously h a v e not
been tackled on a r e a l scale. The a n a l y s i s of flow in storm d r a i n s has
probably received more attention (Stephenson 1981) due to the ease of
synthesizing inflow hydrographs and the larger scale of storm sewer
systems (Stephenson a n d Hine, 1986).

HYDRAULIC ANALYSIS

In employing a computer simulation method allowance for time lag,


backwater effects, routing and probability effects could be included.
However, if al I these components were considered in one program,
computational time would be increased considerably. The hydraulic
equations were therefore s i m p l i f i e d b y o m i t t i n g some of these effects and
time l a g r o u t i n g was employed. Peaks due to individual lavatory flushes
and so on can also be shown using probability theory to be rapidly
attenuated (Chan a n d Wang, 1980).
One computer model simulates the flow down sewers a n d accounts for
time l a g as well as d i f f e r e n t types a n d times of inflow. However as the
individual lengths of sewer number 135 000, i t i s often not convenient o r
in fact a p p r o p r i a t e to a n a l y s e the flow in each sewer. Another program
therefore e x i s t s f o r abstracting data from relevant areas which require
a n a l y s i s and even f o r condensing the d a t a so t h a t a number o f lengths of
sewer could be considered together to speed a n a l y s i s (Constantinides, 1982).
I t i s possible to i d e n t i f y the type of inflow i n each case i n o r d e r to
build the correct contributor hydrograph. One of the objects of the
program i s to study the l a g and a t t e n u a t i n g effect on h y d r o g r a p h s . The
times of day at which the flows start and increase and subsequently
decrease a r e therefore important a n d f i e l d measurements were r e q u i r e d .

FLOW MEASUREMENTS

Measurements were made in manholes as n e a r a s possible to the source


of sewage in o r d e r to minimize the time l a g a n d t o a v o i d a t t e n u a t i o n due
to r o u t i n g down the sewers. The flow depths in t h e sewers wer gauged a t
manholes over a p e r i o d of weeks a n d the r e s u l t i n g h y d r o g r a p h s plotted.
The v a r i a t i o n s from week to week were s l i g h t and s i m i l a r weeks were
averaged f o r compil i n g the hydrographs. The o b s e r v a t i o n s t a k e n a t night
168

in dry weather were assumed to indicate leakage plus infiltration. By


comparing these r e a d i n g s a t the end of summer a n d the end of winter in
dry periods it was possible to estimate the relative proportions of
infiltration and leakage from the plumbing systems. Observations made
during and after summer storms indicated storm inflow to the system.
R a i n f a l l i n Johannesburg i s n o r m a l l y r e s t r i c t e d to the summer season when
h i g h i n t e n s i t y storms of short duration occur during the afternoons. The
sewers were assumed to flow unsurcharged during storm flow and
surcharged conditions were d i s c a r d e d a s they would h a v e been d i f f i c u l t to
use for e s t i m a t i n g a c t u a l c o n t r i b u t i n g flows.
The peak flow r a t e d u r i n g the d a y f o r h i g h e r income housing averages
1.17 I/min per house excluding leakage, infiltration and stormwater
ingress. The amount of leakage from the p l u m b i n g system i n t o the sewerage
system i s estimated to be 0.06 I/min per house, over 24 hours a day
throughout the year. The i n f i l t r a t i o n in l e a k i n g sewers i s estimated to be
0.05 I/min p e r metre of sewer p e r metre diameter. This i s greater for older
sewers in poor soil. The additional flow during and after storms is
estimated to a v e r a g e 1 % of the p r e c i p i t a t i o n over the catchment. This w i l l
v a r y w i d e l y depending on the methods o f c o n t r o l l i n g stormwater inflow into
gullies. T h i s flow i s also associated w i t h the one-year recurrence interval
storm over n h o u r s which i s estimated to be 2 mm/h f o r general analysis.
The flow suggested to peak flow design i s 0.05 mm/h which is 1% of the
two-year storm of 5 mm/h. The e f f e c t i v e c o n t r i b u t i n g area i s about 50 m
w i d t h of catchment p e r metre of sewer.
Up to 5%, and in i s o l a t e d cases even more, of the rainfall over an
assumed 50 m wide strip over all sewers was found to e n t e r sewers in
some cases. The a c t u a l sewer flows often increased b y over 50% - even to
69% d u r i n g a n d a f t e r a storm.
I n the case of the f l a t a r e a s a special g l a s s f i b r e flume was used in
the invert of a manhole. This has a curved bottom and a hump which
made it possible to measure relatively small flows from a block of 273
flats. T h i s e l i m i n a t e s the problem of a s c e r t a i n i n g e x i s t i n g sewer g r a d i e n t s .
Another advantage is the fact that the results are not affected by the
existence o f s i l t a n d so on. Where possible, a conventional flume was used
to gauge flows. The normal method of r e c o r d i n g a t these flumes was on a
c i r c u l a r chart w i t h an integrated total flow r e a d on a meter; the charts
a r e changed a t weekly intervals. To g i v e continuous flow rates a Fisher
P o r t e r meter was installed for several weeks t o g i v e a c c u r a t e d a t a at 15
minute intervals.
169

The input hydrographs were reproduced by the computer using a


F o u r i e r series t y p e of c u r v e f i t .

H i g h e r income r e s i d e n t i a l

An area o f approximately 1220 h a w i t h well e s t a b l i s h e d medium-sized


houses and newer houses including town houses was selected for the
development of the h y d r o g r a p h taken in a h i g h e r income r e s i d e n t i a l area
(Fig. 11.1). Equivalent house u n i t s were based on a 150 m2 f l o o r area
normalized b y r a i s i n g t h a t to the power of 0.8 to a l l o w f o r the r e d u c t i o n
in flow p e r u n i t of f l o o r a r e a as houses a n d f l a t s increase in size. Hotel
and servant accommodation is allowed for at the rate of one house
e q u i v a l e n t u n i t f o r every three rooms. Shops, offices, schools a n d churches
a r e allowed f o r a t the r a t e of one house e q u i v a l e n t per 300 m2 of floor
area.
Metered Calculated
(November 1981)
Daily Iota1 kl 58 5 58 7
Average Its 67 8 68 0
Peak 11s 1164 1156
Minimum 11s 24 0 24 0

'"t - Calculated flow


.-. 0. Melered llow

I I
0 3 6 9 12 15 18 21 24
Time of day h

Fig. 11.1 Comparison of c a l c u l a t e d a n d metered f l o w s in h i g h e r


income r e s i d e n t i a l areas

Minimum o r base flow was assumed to be due to i n f i l t r a t i o n i n t o sewers


a n d manholes o r to leakage w i t h i n the b u i l d i n g s . This was shown to be
equal to 41% of the average, o r 20% of the maximum flow in d r y weather.
At 5.15 a.m. the flow rises r a p i d l y to within 80% of maximum flow,
which occurs at 8 a.m. This flow pattern suggests that higher income
residents in Johannesburg get up a t 5 a.m. onwards. Generally, production
workers start work at 7 a.m. and office workers at 8 a.m. Traffic
congestion demands an e a r l y s t a r t for those t r a v e l l i n g the 12 km i n t o the
city.
170

The evening peak occurs at about 7 p.m., indicating preparation of


meals, ablutions a n d so on, and activity ceases at midnight. Maximum
flow p e r u n i t was found to average 1.17 I/min.

Low income residential

Detailed land use d a t a were not available for the study of the low
income residential area (Fig. 11.2). Details of houses and flats were
a b s t r a c t e d from c o n s t r u c t i o n d r a w i n g s a n d checked on s i t e before the sewer
d a t a were used. The a r e a chosen embraced most of the newer sections of
Lenasia - a town 25 km south o f Johannesburg - a n d the sewage flow was
monitored with a flume. A time l a g of one hour was allowed for when
comparing the h y d r o g r a p h measured with the hydrograph at the p o i n t of
origin.

Monday ,Tuesday Wednesday Average Calculaled


Daily Iota1 kl 461 462 460 461 463
Average. 11s 531 5 32 5 30 531 54
Peak. Us 14 77 1532 1450 1486 14 7
Minimum. U s 0.15 0.28 0.28 0.23 0.3

15
-
-- - Calculaled l b w
Monday 30 January 1984
... Tuesday 31 January 1984
........... Wednesday 1 February 1984

5
a
E

l 1

0 3 6 9 12
Time of day h
15 18
. 21 24

F i g . 11.2. Comparison o f c a l c u l a t e d a n d metered f l o w s i n low


income r e s i d e n t i a l a r e a s

Rapid increases i n flow occurred a t 5.30 a.m., p e a k i n g a t 7.15 a.m. A


smaller peak a t 10 a.m. i n d i c a t e d g r e a t e r a c t i v i t y i n the home a f t e r the
departure of the working population than in higher income areas,
particularly on Mondays - the traditional washing day. A greater
p r o p o r t i o n of the f a m i l y c o u l d remain a t home, which i s also indicated b y
the smaller evening peak a t 8 p.m.
Maximum flow p e r house u n i t amounted to 0.46 I/min. The low minimum
flows metered are due to the recent construction of all sewers and
b u i Idings.
171

Apartment buiIdi n g s

A large complex known as Helderberg in Berea, Johannesburg, was


selected for the study of an apartment (flats) area (Fig. 11.3). It
comprises 273 f l a t s with a total of 585 rooms. Metering was c a r r i e d out
close to t h e b u i l d i n g so the time l a g was minimal.
Points of interest of the hydrograph include the low level of
infiltration and leakage, the sharp rise and subsequent fall in the
m o r n i n g peak, the secondary morning peak a t 10 a.m. each d a y except on
the Thursday a n d the d i s t i n c t i v e peaks a t 7.30 p.m. a n d 9.30 p.m. which
may have something to do with the television viewing of flat-dwellers.
Inflow starts later than i n other residential areas, possibly due to the
g r e a t e r p r o x i m i t y of these f l a t s to places of work. Maximum flows appear
t o be h i g h e r t h a n i n other types of development, i.e. 2.05 I/min p e r u n i t .

Monday Tuesday Wednesday Thursday Average Calculated


Daily lolal hI 238 2295 241 9 2374 2367 240
Average 11s 2 74 2 66 2 82 276 275 28
Peak 11s 8 66 7 45 8 66 8 31 027 86
Minimum lls 0 27 041 0 46 052 042 03

10 - - Calculated (low
---
-.-
Monday 21 November 1983
Tuesday 22 N o v e m k 1983

Time 01 day h

F i g . 11.3 Comparison of c a l c u l a t e d a n d metered f l o w s i n a f l a t a r e a

Commercial a r e a s

For the study of a commerical a r e a ( F i g . 11.4) a p o r t i o n of t h e c e n t r a l


business d i s t r i c t of Johannesburg was selected. House e q u i v a l e n t u n i t s were
obtained on the b a s i s of 300 mz of f l o o r a r e a a n d amounted to 3026 u n i t s ,
i n d i c a t i n g a t o t a l f l o o r a r e a of 90.8 ha. The development i s exclusively
commerc i a I.
A r a p i d increase in flows occurred a t 6 a.m. The peak m o r n i n g flow
occurred a t 1 1 a.m. Flows then d e c l i n e d u n t i l a r a p i d increase o c c u r r e d a t
around 3 p.m., resulting in the peak daily flow at 4 p.m. These
phenomena a r e i n d i c a t i v e of normal o f f i c e hours. The afternoon peak must
b e due to the use of l a v a t o r i e s a n d washing j u s t before s t a f f l e f t work.
172

The base f l o w s are high relative to higher income r e s i d e n t i a l areas,


w h i c h can be a t t r i b u t e d to a h i g h leakage r a t e .

20-23 April 1982 5-8 July 1982 Average Calculalecj


Daily average total kl 4955 6 47185 4837 0 4849 8
Average 11s 57 4 54 6 56 0 56 1
Peak 11s 90 5 94 7 92 6 93 2
Minimum 11s 31 9 30 0 30 9 30 3

150 I - Calculated llnw


......Weekday tlow 20-23 Aprll 1982
-..... . Weekday l l w 5-8 July 1981

I I

0 3 6 9 12 15 18 21 24
Time 01 day h
F i g . 11.4 Comparison of c a l c u l a t e d a n d metered flows in a commercial area

Industrial

Several industrial areas were investigated in detail to establish a


reasonably r e a l i a b l e method of simulating flows. Types of industry were
generally mixed b u t d i d not include any heavy industry. Initially units
were established based on 100 mn of floor area which showed large
discrepancies in flows due to t h e predominance of h i g h o r low water usage
b y i n d i v i d u a l f i r m s w h i c h bore no r e l a t i o n to f l o o r a r e a . Variation of the
f l o o r area p e r u n i t d i d not therefore g i v e consistent r e s u l t s from one a r e a
to another. I t was found t h a t a c t u a l water s u p p l y g a v e the best i n d i c a t i o n
of e f f l u e n t discharge. Water meter r e a d i n g s a r e n o r m a l l y taken e v e r y three
months i n Johannesburg a n d stored in computer files. It was possible to
e x t r a c t meter flows over a three-month p e r i o d a n d then base u n i t s on an
average d a i l y flow of 800 I / d a y . An i n d u s t r i a l a r e a o f 160 h a was selected
f o r t h i s study ( F i g . 5 ) . M a j o r i n d u s t r i e s i n c l u d e d yeast m a n u f a c t u r e w h i c h
has a v e r y h i g h water usage.

CONCLUSIONS

The composite h y d r o g r a p h s p r e p a r e d from sewer flow measurements (Fig.


1 1 .l-11.5) indicate varying peak times and the importance of assessing
i n d i v i d u a l h y d r o g r a p h s i s thereby emphasized. Out-of-phase peaks w i I I not
be c u m u l a t i v e a n d as a result sewer capacities need not be the sum of
peaks. The time l a g of individual contributor hydrographs also adds to
173

the attenuation effect.


The computer s i m u l a t i o n program i s used f o r the p l a n n i n g a n d design
of extensions to the sanitary sewer collection network. The program is
used to i d e n t i f y bottlenecks, study the effects of re-zoning or subdivision,
plan new facilities, size temporary diversion works, plan alternative
routes, size t r u n k sewers a n d estimate loads a t o u t f a l l works.
The program is linked to a land use inventory for assessing
contributions and to plotting program for drawing sewer longitudinal
sections. It is proposed to estimate future flows using a land use
c l a s s i f i c a t i o n established in the computer d a t a b a n k .

Monday Tuesday Wednesday Thursdav Average Calculated


Dailylolal k l 10317 11157 10565 10423 10615 10647
Average I/s 1194 129 1 122 3 1206 1228 1232
Peak lls 216 5 208 5 208 5 2165 2125 2132
Minimum 11s 60 2 69 9 75 0 67 4 68 1 74 5

300 -
--
-..-.-.-
Calculaled llow
Monday 5 July 1982
Tuesday 6 J U I ~1982
Wednesday 7 July 1982
....... Thursaay 8 July 1982

,7 E FE R EN C ES

Chan, W.Y.W. a n d Wang, L . K . , 1980. Re-evaluating H u n t e r ' s model f o r


r e s i d e n t i a l water demand, J. Am. VJat. Wks Ass.
Constantinides, C.A., 1982. Comparison o f time l a g and k i n e m a t i c flow i n
conduits. Water Systems Research Programme, University of the
Witwatersrand.
Crabtree, P.R., 1976. Flow a n d i n f i l t r a t i o n g a u g i n g in sewers. N a t i o n a l
B u i l d i n g Research Institute, Concil for Scientific and Industrial
Research. P r e t o r i a .
Shaw, V.A., 1963. The development of c o n t r i b u t o r h y d r o g r a p h s f o r s a n i t a r y
sewers and t h e i r use in sewer designs. Civ. Engr. S. A f r . 5, No. 9,
246-252.
Stephenson, D., 1981. Stormwater h y d r o l o g y a n d d r a i n a g e , E l s e v i e r , p 276.
Stephenson, D. and Hine, A.E., 1986. Simulation of sewer flow. M u n i c i p a l
Engineer, 3. 107-112.
174

APPENDIX 11.1

PROGRAM SEWS IM

T h i s i s a micro-computer (HP9816) o r i e n t a t e d version of a program to


store d a t a and s i m u l a t e flows down s a n i t a r y sewage networks.
Contributor hydrograph characteristics are programmed for various
types of development e.g. residential upper class (type 11, lower class
(type 2), industrial (type 3) a n d commercial ( t y p e 4). The actual peak
flows per unit (P/min/100m2) must b e . inserted in the data. Equivalent
number o f house u n i t s (HE) or 100m’ in t h e case of non-residential, is
r e q u i r e d f o r each p i p e , a n d peak in P/min/HE f o r each section.
Hydrographs a r e accumulated w i t h time l a g proceeding down a l l sewers.
Time l a g s a r e based on f u l l p i p e v e l o c i t y a s h y d r o d y n a m i c a n a l y s i s would
be too time consuming a n d not worth the e f f o r t .
Hydrographs over a n y p e r i o d of time e.g. 24h (starting at midnight)
f o r any time i n t e r v a l (e.g. l h ) f o r a n y number of selected p i p e s ( s a y 10
maximum) a r e t a b u l a t e d , and plotted i f required.
Also indicated are maximum flow for every pipe, its capacity and
overflow volume i f not adequate. A summary of pipe lengths and inflow
areas i s t a b u l a t e d a t the end.
Each sewer i s i d e n t i f i e d b y a number. The n u m b e r i n g system f o r sewers
can be selected such t h a t the f i r s t two d i g i t s of a 6 - d i g i t number i n d i c a t e
the o u t f a l l region, the second two the s u b u r b a n d the l a s t two the a c t u a l
p i p e which i s assumed t h e same as the top end manhole.

Effect o f Local Peaks ( P r o b a b i l i t y a n d R o u t i n g )

The design flow from a house connection f o r sewer design is typically


1,5P/min. The a c t u a l peak d i s c h a r g e i s c o n s i d e r a b l y h i g h e r b u t when the
effect of a number of houses is accumulated the above figure is
reasonable. A proof t h a t the instantaneous peaks c a n be neglected follows.
A typical toilet f l u s h occurs a t a rate of 204 in 7 sec i.e. 3P/s.
Frequency of f l u s h i n g a t peak p e r i o d s i s once e v e r y 2,5 minutes = 1/150s
p e r house. Therefore a f t e r 7 houses mean flow/house assuming o n l y 1 house
flushes a t a time, i s 1,2P/min - w h i c h i s a normal design flow. i.e. after
10 houses o r so the flushes a v e r a g e to g i v e a normal flow design f i g u r e .
The p r o b a b i l i t y of any two houses f l u s h i n g simultaneously is (7/150)’ =
1/400, a n d of 3 houses 1/8000 etc. i.e. remote so the coincidence o f a peak
from each house together i s remote. I n any case those peaks a r e rapidly
175

attenuated b y the r o u t i n g effect described below.

Routing effect (Graphs from Stormwater Hydrology a n d D r a i n a g e Stephenson,


1981 ) .
In order to shorten running time, the program does not include
h y d r a u l i c r o u t i n g effects. The f o l l o w i n g section i s proof of the fact that
r o u t i n g has a n e g l i g i b l e effect on peak flows. Routing i s the s p r e a d i n g of
a wave and corresponding reduction in peak due to hydrodynamic forces.
I t i s superimposed on the time l a g effect.
Consider the depth corresponding to a flow of 3P/s f o r 7s in a 150mm
dia. d r a i n a t a slope of 1/100:
From a c h a r t f u l l flow Qf = 2OP/s.
Therefore Q/Qf = 0.15.
Therefore r e l a t i v e depth a t 3 t / s from the c h a r t i s
y/D = 0.25
Now f o r a reduction in depth from 0.25D to 0.125D (i.e. half original
d e p t h ) from the c h a r t

Ey
0 15*x
~0.013a100 - = 15m
0.003'7

.'. x = t2m

i.e. depth h a l v e s i n 12m of sewer pipe.

Therefore the r o u t i n g effect i s v e r y r a p i d to s t a r t w i t h i f the flow is


v e r y low.
On the other h a n d the r o u t i n g effect on a h y d r o g r a p h from 100 houses
i s c a l c u l a t e d below:
Flow r a t e q = 1.5P/min x 100/60s = 3e/s as well, b u t Q(volume) i s now
0.003 x 8h x 3600. i.e. depth w i l l h a v e over 12rn x 3600 x 0/7 = 50 km
i.e. n e g l i g i b l e r o u t i n g over the f i r s t km o r so and b y then the number of
c o n t r i b u t i n g houses w i l l f a r exceed 100.

Non-Circular Conduits

Sewers are sometimes non-circular e.g. rectangular culverts or egg


176

shaped. Then the 'diameter' in the d a t a may be r e p l a c e d b y 4R where R


i s the h y d r a u l i c r a d i u s , equal to A/P and A i s t h e cross sectional area,
a n d P the wetted perimeter, at full flow. This procedure results in the
correct time l a g i n the computations, w h i c h i s taken f o r the f u l l flow. The
flow c a p a c i t y of the c o n d u i t w i l l however b e i n c o r r e c t l y indicated in the
results. The a c t u a l d i s c h a r g e c a p a c i t y is:

A
Q = A v =
- De2 Qe
4

i.e. the i n d i c a t e d flow Q i n the computer p r i n t o u t should be m u l t i p l i e d b y


the true cross sectional area divided by ( nDe'/4) where De is the
e q u i v a l e n t diameter 4A/P i.e. true capacity Q = (P2/4nA)Qe.

I n f l o w Components

I n f l o w to each sewer i s assumed to comprise four components; sewage flow


from connections, stormwater ingress, a function of sewer length, steady
groundwater i n f i l t r a t i o n which i s a f u n c t i o n of s e w e r length, and leakage.
Each parameter can be s u p p l i e d i n the d a t a a n d u n t i l more a c c u r a t e d a t a
i s available, the f o l l o w i n g f i g u r e s a r e suggested:

Sewage I n f l o w : Peak net i n f l o w r a t e of 1.0 litres per minute per house


equivalent i s average i n m i d d l e c l a s s r e s i d e n t i a l areas.

Stormwater : From gulleys, manholes and leaks, mm/h. About 1% of


p r e c i p i t a t i o n r a t e . e.g. 1% x 10mm/h = O.lmm/h.

I n f i l t r a t i o n : 0.15 l i t r e s p e r m i n u t e p e r metre of sewer p e r metre diameter.


Increase for o l d sewers.

L e a k s : from c i s t e r n s , d r i p p i n g taps etc. 0.15 l i t r e s p e r m i n u t e p e r house


equivalent. Increase f o r o l d e r a n d l a r g e r p r o p e r t i e s .

Inflow distribution assumed is a series of sin waves. The relative


peaks of each of the 3 s i n waves f o r any type hydrograph a r e designated
Q1, Q2 a n d 43 in the program. The time in hours when each of those
peaks occur a r e T11, T21 a n d T31 a n d the time at which the sin waves
(positive half only) start are T10, T20 and T30 respectively (all in
hours). The f i r s t wave should s t a r t at T10. These values are built into
177

the program a n d the p r o g r a m must be e d i t e d to change them.

DATA

Data can be stored a n d edited in a separate f i l e , named a t the time of


establishment and identified when SEWSIM i s run. Data requirements are
identified in the d a t a form. As the d a t a file is appended to the main
program a t the time of r u n n i n g u s i n g the "GET" statement, l i n e numbers
should be equal to and greater than 2000 to a v o i d o b l i t e r a t i n g program
lines. The d a t a f i l e should end w i t h END. Data can b e i n free format w i t h
commas s e p a r a t i n g numbers.
Three items of d a t a a r e not used i n SEVJSIM. These a r e sewer depth,
drop a n d g r o u n d level. They a r e intended f o r a d a t a l o g g i n g a n d p l o t t i n g
program l a t e r . Zeros may be inserted a t t h i s stage.

Program Output

The program sorts the p i p e d a t a i n t o order and identifies the lowest


m a n h o l e ( s ) . I f there i s more than one unconnected (downstream) manhole
the d a t a should be checked. A p i c t u r e i s d r a w n of the system w h i c h can
be copied u s i n g DUMP GRAPHICS. Then press CONTINUE f o r the p r o g r a m to
r o u t e the flows through the system and t a b u l a t e maximum flow etc. i n each
pipe.
Hydrographs are also tabulated for nominated pipes. Note that the
h y d r o g r a p h s a r e t a b u l a t e d a t the times corresponding to when they would
reach the e x i t of the system ( t h e lowest manhole) and to get the actual
time at which the tabulated flow occurs subtract the lag time of the
correspond i ng p i pe from the t a b u I a t ed t ime.
The h y d r o g r a p h s a r e a l s o p l o t t e d on the screen a t the correct times. To
p l o t a h y d r o g r a p h DUMP GRAPHICS then/or continue.
F i n a l l y a t a b l e summarizing t o t a l p i p e l e n g t h a n d house u n i t s f o r each
type of development i s g i v e n .
'LO! Fi'E-Sl'OHE"SEWSI~1"
12 ! D. STEPHENSON
15 N d = T 0 2
. W 1'1'8 7 16-2560 - 1b 03.87 I

16 DlJMP O E V I C E I S Nd
18 PRINTER I S Nd
20 D I M M ( 4 0 0 ) . M d ( 4 0 0 ) . D ( 4 0 0 ) .S(408). X ( 4 0 0 ) .Hq(400) . Y ( 4 0 0 ) .H(40GI). I t (400) . T x (400)
. T I (400) . Q ( 4 0 0 ) . Q c ( 4 0 0 ) . Q w ( 4 0 0 ) , G ! s ( 4 0 0 ) ,Qi (400) .Ql (400) . Q m ( 4 0 0 ) .Qv(4P)0) . J d ( 4 0 B )
3v1 DJM I21 (99) . Q 2 ( 9 9 ) , Q 3 ( 9 9 ) . T 1 8 ( 9 9 ) . T 1 1 (99) . T 2 0 ( 9 9 ) . T 2 1 (99) . T 3 0 ( 9 9 ) , T 3 1 (99) . M h ( 9
9 ) , Jh (99) ,at (29.99) ,He(99) .G1 (400)
31 COM NSCZ03
40 I N P U T "NAME '?".NS
60 READ N s . T s . T i . N h , O l l !NO.SECNS. S I M L N h . T I N C h .Nhydqphs.GLmBOTMH
70 FOR Jn=l TO Nh ! HYDROGHAPHS
80 HEAD M h ( J n ) !PIPE NO OF HYGPHS
90 NEXT Jn
100 J2=B
110 FDR K = l TO Ns ' SECT DATA
I20 READ Mb,Itk,Np,Am,Fw,Fs,Fi,Fl ! HOTM MH, ZONE TYPE,Npipes,MANNINljn.I-/MIN/HEQ
,mm/h STORM,INFIl/MIN/M/M.LK/MIN/HEQ
130 J2=J2+Np
140 M d ( J 2 ) = M b
150 FOR J=J2-Np+l TO J2 ! P I P E DATA
160 READ M ( J ) , D ( J ) . X ( J ) , S ( J ) , H q ( J ) ,Y(J) , H ( J ) ,G1 (;I) !NO. ,DIClmm,I..m,Sm/m.H~USE EQU
VS(l00m2),DETTHm,DROP BOTMENDmm,GLmMH
180 I F J < = J Z - N p + l THEN 200
190 M d ( J - l ) = M ( J )
200 I t . ( J ) = I t k
205 Q v ( J ) = K
260 T x ( J ) =X ( J) +Am*4". ( 2 / 3 /) ( D ( J )/ 1000) ." ( 2 / 3 )/S( J)". 5 ! L..AG, 5
*
270 Qc ( J) =. 7 8 5 / A m / 4 . " ( 2 1 3 ) (D( J) / 1000) ( 8 / 3 +S
,*% ) ( J) 5+ 1000 !CAPAC I TY ,L./s
280 Qw (J)=Fw+Hq ( J ) /h0/ 1000 !PEAK M3/S INFLOW
290 Q s ( J ) = F s * X ( J ) * l 0 0 / 1 0 0 0 / ~ 6 0 0 ! DO. STORM
300 Qi ( J ) = F i + X (,I) +D ( J )/ 1 0 0 0 / 6 0 / 1000 ! DO. I N F I L T N
310 Q1 ( J ) = F l + H q ( J ) /60/1000 ! DO. L E A K S
320 NEXT J
330 0 (K)=J2-Np+l ! TEMP. TOPMH FOR PLAN P1.OT
340 9rn (F:) =Np
350 NEXT K
400 T40=12 !STORM START h
410 T41-17 !PEAK h
420 91(1)=.7 ! RESID4UPCLAS. 1S T P E A K = l
430 92(1)=.5 ! 2 N D PEAK
440 93(1)=.6 ! 3 R D PEW
450 T10(1)=5 !START h 1ST PEW MUST BE 1st HG TO START
460 Tll(i)=9 ' P E A K h 1ST PEAK
470 T20(1)=7 ! S T A R T h ZND PEAK
480 T21(1)=13 !PEAK h 2ND PEAK
490 T30(1)=14 !START h 3RD PEAK
500 T 3 1 ( 1 ) =20 !PEAK h 3RD PEAK
510 91(2)=.8
520 92(2)=.7 !RESIDLPOOR
530 c-13(2)=.6
540 T10(2)=6
550 T11 (2)=8
560 T20(2)=7
570 T Z 1 ( 2 )=13
580 T30(2)=16
590 T 3 1 ( 2 ) = 2 0
4500 01(3)=.4 ! INDUST
6llb 02(3)=.S
620 93 ( 3 ) =. 45
630 T10(3)=6
640 T 1 I ( 3 )= I 0
650 T 2 0 ( 3 ) = 4
660 T 2 1 (3)=13
670 T 3 0 ( 3 ) = 1 3
675 T 3 1 ( 3 ) = 1 6
680 01 (4)=.3 !COMMERCIAL
690 02(4)=.7
700 03(4)=.45
710 T 1 @ ( 4 ) = 6
--.
720 1 1 1 ( 4 ) = 1 0
/d T 2 0 ( 4 ) = 6
740 T 2 1 ( 4 ) = 1 3 d

W
w
t);
i
B 61
t.4 4
m IXI
2 z
W iii
I I
!- L
w
a
+ r.l
a -
1030 NEXT E
1032 GINIT
1033 GRAPHICS ON ! DRAW LAYOUT
1034 GCLEAR
1035 WINDOW - 1 0 0 0 . T l ( N l ) + 1 0 0 0 , 0 , N s + l
1039 FOR E = l TO Ns
1040 FOR J = B ( E ) TO O m ( K ) + O ( # ) - l
1041 MOVE T 1 ( N l ) - T l ( J ) . K
1042 DRAW T 1 ( N l ) - T l ( J d ( J ) ) - 2 0 . K
1043 L l = T l ( N l ) - T l (J)-20
1044 C S I Z E 3..4
1045 MOVE L1 .K
1046 ! D I R l+PT/b
1047 LABEL M ( J )
1048 LDIR 0
104.9 NEXT J
1050 Jmm=Om ( K ) +O ( K ) - 1
1051 I F J d ( J m m ) = 0 THEN 1055
1052 MOVE T 1 ( N l ) - T l ( J d ( J m m ) ) - 1 0 . K
1053 DRAW 1 1 (N1 ) - T 1 ( J d ( J m m ) ) , G ! v ( J d ( J m m ) )
1055 NEXT K
1060 PAUSE
1070 FOR J=1 TO J2
1071 (;7m(J)=0 ! MAX. FLOW
1072 Qv(J)=0 !SPILL VOL.
1073 NEXT J
1074 Ti=0 !COUNTER FOR T H
1075 FOR T h = T i TO T s STEP T i ! T I M E h AT BOTTM
1076 FOR J=1 TO 52 I N I T I C I L I Z E FLOWS
1077 R(J)=0
1078 NEXT J
1080 FOR J=1 TO 52
J 090 T S T h - T I ( J ) /3h00 !PIPE T I M E FOR REACHNG E X I T A T T h
1092 I F T : : . T 1 0 ( I t ( J ) ) THEN 1100
1093 T=T+24
1100 IF T > = T l B ( I t ( J ) ) THEN 1110 !HYOROGRAPH ORDINATE PER PIPE
1290 NEXT J 5
1300 NEXT J
1 3 1 0 FOR J = 1 TO J Z
1320 I F R(J)<=Qm(J) THEN 1340
1 3 3 0 Qm(J)=Q(J)
1 3 4 0 I F Q ( J ) < P c ( J ) / l 0 0 0 THEN 1360
1350 R V ( J )=Qv (J)+ ( Q ( J )-Qc ( J )/ 1000)+3600+Ti
1360 NEXT J
1370 T i = T J + l
1380 FOR Jn=l TO Nh !HYDRAPH POTNTS
1 3 Y 0 R t (Jn,T j )=a ( J h ( J n )) + l B 0 0
1400 NEXT J n
1.410 NEXT T h
1420 PRINT "SEWER NETWORK ANALYSIS , "
1430 PRINT N8
1440 PRINT 'I P I P E D I A SLOPE HUNITS LAGh M A X L s QCAP XFIJL.1. OFL.Om3"
1450 IMAGE DDDDDD.DDDDD.D. DDDD.DDDDD.DDD. DD.DDDI)D.T)I)DDD.onnI>. D.DDDD. D
146R FnR J=l TO 52
1470 T 1 ( J ) = T l (J)/360cI
1400 Qm (J ) =Qm (J)+1800
1490 P f = a m (J) /Qc (J ) + 1 0 0
1500 PRINT USIND 1450:M(J) ,D(J) . S ( J ) .Hq(J) .T1 ( . 7 ) .Qm(,l) . R c ( , l ) . P f . R v ( J )
1 5 1 0 NEXT 3
1520 IMAGE DDDDDDDD.#
1525 IMAGE DDDDD.DD.#
1526 IMAGE DDDDD.DD
1 5 2 7 IMAGE DDDDDDDD
1530 PRINT
1540 PRINT "SELECTED HYDROORAPHS AT":Ti ;"h 1NTVL.S STCIHTTNG A T - T l FKIR PTPF
1 5 4 5 PRINT USING 1520;Ti
1550 FOR Jn=l. 7'0 Nh-1
1551 PRINT USING 1520;M(Jh(Jn))
1552 NEXT J n
1 5 5 3 PRINT USING 1 5 2 7 : M ( J h ( N h ) )
1 5 6 0 FOR T k = l TO T i
1565 PRINT USING 15'20:Tk
1570 FOR Jn=l TO Nh-1
1571 PRINT I.JSING 1525:Qt (-7n.Tk)
1 5 7 2 NEXT J n
1573 PRINT IJSING 1526:Qt ( N h . T k )
1575 NEXT TIC
1380 GCLECIR
1600 FOR Jn=l TO Nh
1602 QCLEAR ! P L O T HYDROORCIPHS ON SCRN- DUMP GRAPHICS & / O R CONT
1603 CSIZE 4
1605 Omi=Qm(Jh(Jn))
1610 WINDOW -3,24.5,-0.Qmj*l. 1
1620 CIXES 1,1,0,0,12,10
1630 MOVE l , Q m ( J h ( J n ) )
1700 *
Qm (Jh (Jn ) ) = I N T (Qm ( Jh (Jn ) ) 1000)/ 1000
1701 L A B E L Qm (Jh (Jn) ) ;" L / s PIPE" :Mh (Jn)
1710 MOVE 6,0
1711 LCIBEL. N S
1720 MOVE 22.0
1721 L A B E L "h 24"
1724 MOVE -2.10
1725 LhBEL 10
1730 FOR T=2 TO T i
1740 Th=(T)*Ti-Tl (Jh(Jn))
1750 MOVE T h - T i .Qt ( J n , T - l )
1755 DRAW Th,Qt ( J n . T )
1760 NEXT T
1765 MOVE -T1 ( J h ( J n ) ). Q t ( J n , T - 1 )
1766 DRAW T i - T l ( J h ( J n ) ) , Q t ! J n , l )
1770 PAUSE ! T Y P E CONT (El-) TO DO NEXT t i Y B 13R DUMPFH TO DRAW EX
1780 NEXT Jn
1790 FOR 1 = 1 TO 6
1800 He(I)=0
1810 NEXT I
1820 FOR J = 1 TO 52
1830 H e ( 1 t ( J ) ) = H e ( I t (J)) + H q ( J )
1840 XleX1+X(J)
1850 NEXT J
1855 PRINT
SEWSIM DATA FORM

--
NAME

2 W ! DATA ..........
NO. OF SECTIONS, SIMULATION DURATION h, T I M E INCREMENT.h, NO. HYDROGRAPHS REQUIRED, G.L.(M) BOTM M i .

2010 DATA ........................ .................................... ..................


P I P E NO. OF HYDROGRAPHS

2020 DATA ..........................................................


SECTION DATA: ( 1 L I N E PRECEEDING EACH SECN.)

BOTTOM MH NO., ZONE TYPE (1-41, NO. PIPES I N SECN, MANNINGN, SW/min/HE, STORM Chnm/h, I N F I L T N L/mm/m/m, EAKG L/mi /HE.
2030 DATA ...........................................................................................................................
P I P E DATA: ( I L PER P I P E )
P I P E NO.(=TOP MH NO.), m
-, LENGTH m , SLOPE m/m, HE=HDWE E a U I V S ( l W m ’ ) , (DEPTH TOP MHm, DROP BOTM mm, GLm

2040 DATA ............................................................................................................................

G i v e e a c h d a t a f i l e a name - c a l l e d up when SEWSIM i s RUN.


Store in ASCII form e.g. SAVE “FILENAME”
187

SAMPLE DATA F I L E

SEWDAT (ASC I I FORMAT)

7000 DATA 4 , 2 4 , 1 . 4 , 5 0
2001 DATA lil,1'21,211,.311., 1 1 . 3 , 1 , 2 ,. 0 1 J , 1 1 , 0 ,. , 0
ZQCI? WTA 111,100,100 ,.0100,100,1.30,49
2061.3 DATA 1 1 2 , 3 0 0 , 3 0 0 , . 6 1 1 0 0 , l ~ 0 , 1 , 0 , 4 8
2804 DATA 112,2,9, . 0 2 , I , 0 , . 1 , . 0
288'3 DATA 1 2 1 , 1 5 0 , 1 B 0 , . 0 1 5 , 1 0 0 , 1 , 0 , 4 7
2806 DATA 1 2 2 , 1 5 0 , 3 0 0 ,.004,75,1,C!Il46
2807 D(?ITA 123,200,i00,.802,95,1.5,30,45
2 0 0 8 DATA 122,3,1,.02,1,0,a..1
2089 DATA 211,150,100,.81,100,1,0,48
2 0 1 0 DATA 123,4,1,.02,1,0,0,0
2 0 1 1 DATA 311,100,iQ0,.002,100,1,0,47
100pI0 END

:YF'B2'mX. 78B
VOLUME LABEL: B982.5
F I L E NAME PRO TYPE FiEC/FILE HYTE/HEC ADDPESS

SEWDAT ASCII 3 256 50


BEWSIM PHDG 7 3
_IL 256 53

Note _-__
SAME "SEWDAT", don't STORE ,I...
188

END P I P E WITH NO D . S . P I P E 2 112

SEWER NETWORK hNALYSIS FOR


PIPE DIA SLOPE HUNITS LAGh MAXLs PCAP %FULL OFLOm3
111 100 .0100 100 .I2 2 4 37.3 0.0
112 308 .0100 100 .07 10 84 12.1 0.0
121 150 .0150 100 .44 2 12 14.0 0.0
122 150 .0840 75 .40 4 6 69.3 0.0
I23 200 .0020 95 .I6 7 10 73.3 0.0
211 150 .a100 100 .45 2 10 16.7 0.0
311 100 .00?0 100 .31 2 2 112.5 0.0

SELECTED HYDROGRAPHS AT I h INTULS STfiRTING AT -TL FOR P I P E ...


I 111 121 21 1 31 1
1 * 29 .03 .I7 0.00
L.
1 .03 .03 .I7 0.00
3 0.00 .03 .I7 0.00
4 0.0Q .03 .I7 0.00
5 0.00 -03 .17 0.00
6 .
a0 .03 .17 0.0Q
7 .79 .68 .67 .32
8 1.25 1.45 1.03 .7b
9 1.56 1.69 1.32 1.10
10 I .67 1 .?O I .51 1.36
11 1.57 .96 1.58 1.50
1: 1.29 1.11 1.52 1 .5l
13 .96 1.18 1.36 1 .41
13 .97 1.18 1.33 1 .21
15 !.18 1.10 1.51 I .47
16 1.34 .94 1.65 1 .69
17 1.40 .94 I .56 I .43
16 1.36 1.03 1 .:7 * 76
19 1.21 1.00 .bl .34
20 1.17 1.01 .51 .08
21 1.07 1 .OO .37 0.Q0
22 .89 .64 .23 0.08
23 .73 .56 .I7 0.08
24 .53 .20 .I7 0.00
189

I 1.684 L / s PIPE 121

1.688 L/s PIPE 31 1

- h 24
190

CHAPTER 12

SEWERAGE SYSTEMS MANAGEMENT

The interest in sewerage systems has increased in recent years as


management of existing systems is improved in order to cope with
i n c r e a s i n g flows and to improve catchment water balances. The design,
s i m u l a t i o n a n d management o f such systems i s t h e s u b j e c t of much r e s e a r c h
(Yen, 1987). Dual systems pose particular problems in older ,areas as
p o l l u t i o n of waterways i s becoming more o f a problem. Rehabilitation of
old systems, including re-lining to increase throughput i s also taking
p l a c e (Adams a n d Zukovs, 1987).
The o p e r a t i o n o f a l a r g e u r b a n sewer system was o p t i m i z e d b y S c h i l l i n g
a n d Petersen (1987) using I i n e a r programming. The storm/waste combined
sewer system in Brenner, West Germany, comprises sewer pipes,
pumpstations, storage ponds and a wastewater treatment plant. Unless
adequately controlled, t h e system i s l i a b l e to f l o o d l o w l y i n g s u b u r b s w i t h
severe economic consequences. The optimization model was run in
conjunction with a catchment s i m u l a t i o n model. The reason f o r this was
t h a t the o p t i m i z a t i o n model was, o f necessity, a s i m p l i f i e d model assuming
linear constraints. Conduit s t o r a g e therefore, a complex function of flow
rate as given by the St. Venant flow equations, could not easily be
i n c l u d e d in a l i n e a r model.
A r a i n d a t a c o l l e c t i o n n e t w o r k was coupled t o a catchment model o n a
r e a l time b a s i s t o p r e d i c t flow r a t e s ( F u c h s e t al., 1987).

LEARNING SIMULATION PROGRAM

The p r o g r a m used a n i t e r a t i v e l e a r n i n g process t o o p t i m i z e o p e r a t i o n o f


t h e system. That is, successive r u n s used p r e v i o u s r e s u l t s to improve o n
the operating r u l e using a r t i f i c i a l intelligence.
The sewer system studied was designed to treat lower flows whereas
o v e r f l o w i n storms r a n t o r i v e r s a n d lakes. I n c r e a s i n g p o l l u t i o n awareness
forced t h e system t o b e improved. At t h e same time a s r e d u c i n g overflows,
p a r a l l e l o b j e c t i v e s were t o reduce p u m p i n g e n e r g y costs a n d avoid street
f lood ing .
The problem was set up t o m i n i m i z e a cost function without violating
constraints. A formal system (referred to as a production system) is
e s t a b l i s h e d w i t h t h r e e components:
191

A w o r k i n g memory w i t h a l l d a t a
A r u l e base
A n i n t e r p r e t e r to choose and a p p l y p r o d u c t i o n s
Improved control i s achieved b y a l t e r i n g the r u l e base o r a d d i n g new
ones. For each u n s a t i s f a c t o r y production a l i s t i s created.
A meta production systems was f u r t h e r added. Meta p r o d u c t i o n s do not
a f f e c t the w o r k i n g memory b u t can change the content of the r u l e base.
The meta system i s e v a l u a t e d b y the control interpreter. A simple example
demonstates the technique:
Stormflow could be stored in a detention basin freely, while street
f l o o d i n g would be an u n s a t i s f a c t o r y state. T h i s r u l e could be described b y
a meta f u n c t i o n as flows:
( W E > 1.0) + (pump too L O W ) (Value = - 1 )
Whenever the water level in the sewers i s h i g h e r t h a n manhole level which
may cause street f l o o d i n g t h i s r u l e i s a p p l i e d . At a n y selected time i f the
meta production rule is applicable the decision PUMP = OFF is counter
r u l e d i.e. the corresponding productions a r e decreased b y 1 .
The facts i n the w o r k i n g memory a t the time may h a v e been
W E = 0.4 where W E = water e l e v a t i o n
R I = 10 where R I = r a i n f a l l i n t e n s i t y .
Another s i t u a t i o n may also have been stored in the experience memory,
e.g.
WE = low
R I = 10
I t i s possible b e t t e r productions c o u l d h a v e been a p p l i e d .
The total l i s t i n memory may now be

Va Iue WE
- -
RI
-3 LOW 9
-1 LOW 10
-6 LOW a
-10 LOW 11

A new p r o d u c t i o n i s created t a k i n g the c o n d i t i o n p a r t of the o l d one. A


second condition i s added of the form
N I op x
where op i s e i t h e r < o r > a n d x i s the median v a l u e of R I in the l i s t of
experience memory, weighted w i t h the level of punishment.
Thus s t a r t i n g w i t h the o l d p r o d u c t i o n
(WE = LOW) + (pump = OFF)
192

the new p r o d u c t i o n w i l l s w i t c h pump on because the systems knows t h i s is


connected w i t h h i g h r a i n f a l l i n t e n s i t y .
Hence the new p r o d u c t i o n i s
( W E = L O W ) ( N I > 9.75) + (pump = O N )
The new p r o d u c t i o n i s assigned a v a l u a t i o n of 0 a n d stored in the rule
base.
I f the s i t u a t i o n WE = 0.4
R I = 10
occurs a g a i n the last r u l e s w i l l b o t h a p p l y , but the l a t t e r r u l e i s chosen
as h a v i n g lowest e v a l u a t i o n l e v e l . Street f l o o d i n g i s t h u s avoided.

OPT I M I ZAT ION

The same problem was simplified into a linear system for direct
optimization at discrete times. Sewers were lumped into three
subcatchments.

Wser I1

Hbsssliisc

I2
-1

Fig. 12.1 Process v a r i a b l e s f o r the s i m p l i f i e d systems


193

A rainfall/runoff model was used to compute inflow hydrographs. The


r e m a i n i n g system consists o f two o f f - l i n e ponds a n d two t r u n k sewers w i t h
backwater effects from the pumps. The system can be described by 18
v a r i a b l e s ( F i g . 12.11, namely:
- i n f l o w I 1 i n t o the pump sump of the downstream p u m p i n g s t a t i o n , inflow
12 h a l f w a y up the upstream station, and 13 into the sump of the
upstream s t a t i o n .
- the pumping r a t e s PR3 i n t o the upstream pond, P2 from the upstream
i n t o the downstream system, PRl i n t o the downstream pond, a n d PKA to
the treatment p l a n t .
- recycled flow from the ponds to the system ( R R l a n d RR3, r e s p e c t i v e l y ) ,
- the stored sewage i n the t r u n k s (V12 a n d V3, r e s p e c t i v e l y ) and i n the
ponds (R1 a n d R3, r e s p e c t i v e l y ) ,
- overflow PO1 into the Weser estuary, 01 into the downstream creek
Wasserlose, and 03 i n t o the upstream creek Krimpelfleet,
- flood volumes which cannot be handled by the system (F12 and F3,
respectively 1.
The s i m p l i f i e d model was verified. This was done through a detailed
a n d p h y s i c a l l y precise model.

Optimal Control a s a L i n e a r Programming Problem

The task in the o p e r a t i o n of the Bremen combined sewer system were


drainage (i.e. m i n i m i z a t i o n of f l o o d i n g ) a n d environmental p r o t e c t i o n (i.e.
m i n i m i z a t i o n of combined sewer f l o w ) w h i l e keeping the cost of operations
as low as possible. Since it is impossible to achieve perfect flood
protection a n d no overflow simultaneously p r i o r i t i e s h a v e to be specified.
They include:
1 . minimum f l o o d i n g (F12, F3)
2. minimum overflow i n t o the creeks (01, 0 3 ) ,
3. minimum overlfow i n t o the e s t u a r y ( P o l ) ,
4. minimum pumping i n t o the ponds ( P R l , PR3),
5. minimum use of the ponds ( R l , R3)
U n i t costs c a r e specified for every c u b i c metre flooded, c u b i c metre
overflow. etc. Using the technique of l i n e a r programming the o p e r a t i o n a l
o p t i m i z a t i o n problem was formulated a s
n
z
min t = l cv3tV3t + cR3tR3t + cv12tV12t + c r l t R l t + cRR3tRR3t + cP2tP2t
+ cPR3tPR3t + cF3tF3t + co3t03t + cRRltRRlt + cPKAtPKAt
+ cPRltPRlt + cF12tF12t + cPOltPOlt + c o l t O l t
1 94

TABLE 12.1 O p t i m a l Control S t r a t e g y f o r M a j o r Storm 0708

t PM PRl Po1 93
_--- ,---------------- -----
----
C 0.0 1.0
.---------------- 10
1280
0.26
-----
1730
1 1630 0 0
2 14268 1277 3800 1277 0 1280 1730
3 'I4268 10000 3600 9414 691 1260 1730
4 14268 10000 3600 Be68 9000 0 1730
6 ill88 10000 3600 6346 6346 1260 1730
6 14268 10000 3600 2946 2946 1260 1730
7 14268 10000 3600 1373 1373 1260 1730
8 14268 10000 3800 603 603 1260 1730
9 14264 10000 $600 a 0 1260 1730
10 13814 10000 3600 0 0 1260 1730
11 13049 10000 3600 0 0 1260 1262
12 13088 10000 3600 0 0 1260 193
13 11037 10000 3600 0 0 1260 0
14 10108 10000 3800 0 0 1260 0
16 9125 10000 3600 0 0 0
16 a i r 4 10000 3800 0 0 0 0 0 1260 0
17 7163 10000 3600 0 0 0 0 0 1260 0
18 6182 10000 3800 0 0 0 0 0 1260 0
19 5201 10000 3600 0 0 0 0 0 1260 0
20 4220 10000 3800 0 0 0 0 0 1280 0
21 3231 10000 3600 0 0 0 0 0 1260 0
22 1612
-~~~ 9549 3600 0 0 0 0 451 171 0
23 531 6560 3600 0 0 0 0 989 171 0
24 531 6490' 3600 0 0 0 0 2070 171 0
25 531 4420 3800 0 0 0 0 2070 171 0
28 531 2350 3600 0 0 0 0 2070 0
27 531 260 3600 0 0 0 0 2070 171 0
28 531 0 1610 0 0 0 0 280 171 0
29 531 0 1530 0 0 0 0 0 171 0

t R3 PR3 F3
.---__----- ------------
c 0.3 1.0 1001
.---------- --__----_
1 4944 4944 0
2 9600 6588 5710 1932 12283 13558 3483
3 9600 6588 270 0 6588 9972 6118 1762
4 9600 6032 0 0 6032 12715 6032 752
._
5 9600 3656 0 0 3656 8410 4916 275
6 9600 2047 0 0 2047 5080 3307 205
7 9600 2011 0 0 2011 3519 3271 194
8 9600 1577 0 0 1577 2752 2837 191
9 9600 997 0 0 997 2147 2257 189
10 9600 396 0 0 396 1701 1656 189
11 9600 0 0 0 0 1386 812 189
12 9600 0 0 0 0 1170 171 i 89
13 6704 0 0 896 0 1170 171 189
14 7615 0 0 1069 0 1170 171 189
15 6626 0 0 1089 0 1170 171 ias
16 5437 0 0 1089 0 1170 171 189
17 4848 0 0 1089 0 1170 171 189
~~~

18 3269 0 0 1089 0 1170 171 189


19 2170 0 0 1089 0 1170 171 189
20 1081 0 0 1089 0 1170 171 189
21 0 0 0 1081 0 1170 171 189
22 0 0 0 0 0 171 189
195

subject to the c a p a c i t y c o n s t r a i n t s

V3 5 1730 m3
R3 5 9600 rn’
V12 5 14268 m’
R1 5 10000 m3
P2 5 0.70 m’/s
PR3 5 3.66 m’/s
PKA 5 2.00 m’/s
PRI 5 8.20 m’/s
PO1 5 5.00 m’/s

a n d the dynamic c o n s t r a i n t s f o r each of the f o u r storage u n i t s

V3t+l - V3t - RR3t + P2t + PR3t + F3t = 13t


R3t+l - R3t - PR3t + RR3t + 03t = o
V12t+l - V12t - P2t-2 - RRlt + PKAt + P R l t + F12t = Ilt + 12t-1
R l t + l - R l t - P R l t + RRlt + POlt + Olt = o

The flow time from the inflow site 12 to the downstream pump was
taken as one time step (i.e. 30 m i n ) and the flow time between the two
pumping stations as two time steps.
The problem was solved w i t h s t a n d a r d software. A typical result is
presented i n Table 12.1 f o r a 210 m i n storm and inflow forecasts of one
time step o n l y (0 to 30 m i n from a c t u a l time). The table includes u n i t
costs c of the o b j e c t i v e function. Sensitivity a n a l y s e s showed that these
c o u l d be specified q u i t e a r b i t r a r i l y , provided that u n i t costs o f different
o r d e r s of magnitude a r e a l l o c a t e d to objectives of d i f f e r e n t p r i o r i t y .

SEWER MAINTENANCE DATA PROCESSING IN JOHANNESBUG

Johannesburg has nearly four thousand kilometres of sewerage to


operate and m a i n t a i n on a continuous basis. Many of the areas a r e prone
to abuse and blockage a n d the n a t u r e o f the topography a n d c l i m a t e make
maintenance a h i g h cost i n the system. That is, intense storms o f t e n r e s u l t
in ingress into sewers and this may bring surface debris and other
f o r e i g n matter which block the sewers. There i s a l s o u n a u t h o r i s e d access
in many places it is suspected, as articles obviously not from the
sanitation system are often found in sewers. Despite the high rate of
1%

g r o w t h i n Johannesburg many o f the sewers a r e o l d a n d some a r e of poor


q u a l i t y r e q u i r i n g r e g u l a r maintenance a n d replacement a n d r e p a i r s .
While a t f i r s t i t may a p p e a r t h a t r e a d y a v a i l a b i l i t y of l a b o u r i n South
Africa should f a c i l i t a t e c l e a n i n g a n d at the same time the maintenance
should provide labour opportunities, the management of such a system
obviously imposes severe problems a t higher levels. Maintenance of logs
for i d e n t i f y i n g trouble spots i n the system would b e of great value to
managers and sewerage engineers. This type of data is useful for
b u d g e t i n g f o r r e p a i r work such a s to manholes a n d even pipes requiring
replacement as well a s m i n o r items such a s manhole l i d s a n d s t e p i r o n s a n d
benching i n manholes. There i s also much to be g a i n e d from a n a l y s i s of
maintenance d a t a in the way o f types of blockage. For instance l o c a l i t i e s
where f o r e i g n objects a r e f r e q u e n t l y encountered can be narrowed down
a n d the i n h a b i t a n t s o f t h a t township made a w a r e of the t r o u b l e s caused b y
such p o l l u t i o n . Where sand i s f r e q u e n t l y f o u n d in sewers i t may p o i n t to
r o a d s r e q u i r i n g s u r f a c i n g as stormwater can r e a c h sewers by unexpected
ways.
Overflows a n d l i f t i n g of manhole lids in c e r t a i n a r e a s may p o i n t to
inadequate sewer capacities. Alternatively they may indicate corroded
sewer l i n i n g s o r roots w h i c h b l o c k the sewers. Here a g a i n i d e n t i f i c a t i o n of
frequency and l o c a l i t y o f such inadequacies i n d i c a t e s where maintenance i s
most u r g e n t l y r e q u i r e d .
The human management side i s also v e r y complex. The o u t l y i n g depots
where such maintenance takes p l a c e employ some s i x h u n d r e d people, w h i c h
a r e g e n e r a l l y o r g a n i z e d i n t o gangs a t each depot. The s u p e r v i s o r s r e p o r t
to managers who take messages a n d t r a n s m i t the teams to problem points.
Even managers a n d f r e q u e n t l y s u p e r v i s o r s a r e not h i g h l y trained a n d the
type of logs they keep are often difficult to process. However the
computerization of the l o g keeping on a n e x p e r i m e n t a l b a s i s a t one of the
depots has proved s a t i s f a c t o r y a n d w i t h i n the c a p a b i l i t i e s of the e x i s t i n g
t y p e of s t a f f . T e r m i n a l s connected to the m u n i c i p a l i t y ' s m a i n computer at
head o f f i c e a r e used a n d once b a s i c k e y b o a r d s k i l l s h a v e been p i c k e d u p
then spread sheet type data l o g g i n g h a s p r o v e d p o s s i b l e and in fact of
g r e a t advantage to the engineers a t head o f f i c e concerned w i t h planning
a n d the engineers concerned w i t h b u d g e t i n g a n d maintenance a n d design.
Although Johannesburg's obvious solution is through its mainframe
computer w i t h o u t l y i n g t e r m i n a l s , in f a c t many s m a l l e r m u n i c i p a l i t i e s may
r e s o r t to m i n i o r even m i c r o computers to h a n d l e t h e i r system. The l a t t e r
would be p o p u l a r w i t h the smaller m u n i c i p a l i t i e s where one s t a t i o n o n l y is
maintained.
197

The use of micro computers also enables micro g r a p h i c s to be used to


identify t r o u b l e areas. A screen map can highlight zones with frequent
blockages. With the advent of the computers many fields in the Civil
Engineering f i e l d have been opened up to the b e n e f i t s which can accrue in
both the design and constructional areas and the management and
administration areas. Due to initial costs and a natural reluctance to
adopt new methods progress i s sometimes slow b u t i t can u s u a l l y be s a i d
that w h i l e computers do not necessarily save money they can definitely
g i v e b e t t e r r e s u l t s a t the end of the day.

AppI i c a t i o n to Johannesburg's system

Thus it was with this i n t e n t i o n of giving an improved service that


Johannesburg h a s persevered w i t h computerization o f many of i t s functions.
T h i s chapter o u t l i n e s the progress made in the p r o v i s i o n a n d maintenance
of sewerage r e t i c u l a t i o n .
The analysis of sewer systems to identify potential overloading by
sewers has already been established and has been used to analyse
townships for existing and future flows. In some cases the effect of
s u b d i v i s i o n of stands has been assessed a n d a c c u r a t e estimates of costs
g i v e n f o r a d d i t i o n a l sewerage work (Stephenson a n d Hine, 1982 a n d 1985).
Sewer r e t i c u l a t i o n s need r e g u l a r p l a n n e d c l e a n s i n g i f serious f l o o d i n g
a n d subsequent danger to h e a l t h i s to be avoided. I f r e g u l a r c l e a n s i n g of
p u b l i c sewers i s well o r g a n i z e d many of the blockages which occur can be
avoided. Maintainance of p r i v a t e l y owned sewers i s not the r e s p o n s i b i l i t y
o f the sewerage a u t h o r i t y b u t i n Johannesburg i t i s the o f f i c i a l p o l i c y to
unblock these sewers i f asked to do so b y the owner. In many cases the
owner i s the local a u t h o r i t y so t h a t there i s a vested interest to ensure
t h a t these a r e well m a i n t a i n e d so as to reduce the number of blockages.
Conventional systems h a v e been used to r e c o r d the work c a r r i e d out
u s i n g c a r d s etc. which h a s been successful but time consuming. I t was
considered t h a t records of c l e a n s i n g work and the c l e a r i n g of blockages
c o u l d be more e f f e c t i v e l y done b y computer a n d that r e t r i e v a l of records
a n d p l a n n i n g o f work would be made easier.
Consequently the M a i n t a i n a n c e Data System has been e s t a b l i s h e d a n d i s
being applied where Sewer Data has been established giving sizes and
lengths of sewers together w i t h a u n i q u e manhole numbering system.
198

Data i s compiled b y depot a d m i n i s t r a t i v e s t a f f on Forms wich have a


numerical format s u i t a b l e f o r i n p u t to the computer. Details are abstracted
from work r e p o r t s d a i l y .
Forms used in the f i e l d g i v e township a n d street names which become
numerical township codes a n d manhole numbers before b e i n g entered into
the computer. I n c o r p o r a t e d i n the c l e a n s i n g r e p o r t i s a n inspection of each
manhole a n d sewer l e n g t h i n c l u d i n g the measurement of the d e p t h o f flow.

Processing of Sewer Maintenance Data

The processing of sewer maintenance data has reached an advanced


stage u s i n g the programs and techniques d e s c r i b e d below.
The workforce i s d i v i d e d i n t o gangs w h i c h work on e i t h e r c l e a n i n g of
sewers o r c l e a r i n g of blockages.
The c l e a n i n g of sewers i s recorded b y the gang leaders in the f i e l d
each day a n d manhole numbers a r e o b t a i n e d from keyplans showing the
sewer network.
On the following working day information is abstracted by depot
a d m i n i s t r a t i v e s t a f f a n d i n s e r t e d i n a numerical format. ( T a b l e 12.1)

TABLE 12.1 Cleansing of Sewers

nEcono OF SYSTEMATIC CLEANINO L SEWERMANHOLE CONDITION


199

Program UPDATE i s then used to provide details of sewer diameter,


length and slope which is added to the data file. These details are
o b t a i n e d from the Sewer Data F i l e ( T a b l e 12 .2 )
A MERGE program i s used to add new data to a master maintenance
f i l e (Table 12.3).
Program MAINTENANCE produces a r e p o r t of the sewers cleaned between
g i v e n dates as r e q u i r e d a n d p r i n t e d out a c c o r d i n g to each township.
Program GANGS produces a r e p o r t of work c a r r i e d out by each g a n g
between g i v e n dates.
T h i s r e p o r t could form t h e b a s i s f o r a bonus scheme ( T a b l e 1 2 . 4 ) .
Blockages a r e recorded as reported with the time a n d date recorded.
The details of actual clearance giving the time started and completed
appear on work r e p o r t sheets a n d enable d a t a to be completed. "Private"
blockages which are within stand boundaries are denoted by a stand
number a n d "Main" blockages which occur in p u b l i c sewers a r e denoted b y
a manhole reference number.
Reports which a r e found to b e problems i n the water reticulation or
storm water system a r e g i v e n a code which enables the computer to i g n o r e
t h a t report a p a r t from showing how many of the r e p o r t s h a v e been r e f e r r e d
elsewhere f o r a c t i o n ( T a b l e 1 2 . 5 ) .
Program BLOCKMACRO produces a report of blockages in townships
between g i v e n dates. The l e n g t h of time taken to c l e a r the blockage and
possible cause i s shown. The time which elapsed between the report and
completion of clearance i s also c a l c u l a t e d to help identify administration
problems, lack of s t a f f etc. The s e v e r i t y of a blockage i s also shown b y
i n d i c a t i n g the number of houses flooded as the r e s u l t of a "main" blockage
and for a "private" blockage i f the house o r y a r d i s flooded ( T a b l e 1 2 . 6 ) .
A macro program produces a r e p o r t of all the work each gang has
done in u n b l o c k i n g sewers between g i v e n dates. Numbers of blockages a n d
t o t a l time spent i s shown ( T a b l e 1 2 . 7 ) .
A program produces a r e p o r t of a l l the stands a n d sewer lengths where
there has been more than one blockage in a given time period. This
information can be very useful in identifying possible defects and
overloading of public sewers and also when answering queries about
repeated blockages on p r i v a t e stands.
TABLE 12.2 E x a m p l e of U p d a t e d Sewer D a t a F i l e h)
0
0
35032913086120202 51.01.5 1 1 3 1 152 15.54 100.0
35032913186120203 41.01.0 2 152 70.26 100.0
35032913286120203 ti 1.0 1.0 2 2 152 57.79 100.0
35032913386120203 &l. 01.5 2 1 1 152 60.01 00.0
350329 13986120202 5 1.02.0 3 152 49.22 80.0
35032914086120202 51.02.5 3 5 152 75.68 00.0
35032914186120202 51.03.0 5 1 6 152 49.10 110.0
35132915086120103 01.01.0 2 2 152 76.90 70.0
35132915106120103 61.01.0 1 4 152 70.73 70.0
35132915286120103 41.02.0 6 3 152 64.00 60.0
35132915306120103 41.01.5 1 1 6 1 152 62.97 60.0
35132915486120102 51.02.0 3 152 65.40 66.0
35132915586120102 51.02.0 5 2 152 62.97 46.0
35132915686120102 51.01.5 4 2 1 152 63.03 79.2
35132315786120102 SO. 50.5 1 152 63.00 79.2

TABLE 12.3 Sewer M a i n t e n a n c e Records

I S E U E R PlCIINlEWNCE R M R D S

REOUESTED CTLIRT WITEX- 861201


REOUESTED END DRTE I- 861204
TOUNSHIP sso - uwsm
DEBRIS MRNMOLE CONDITION SEUER CONDITION
l4RN GRNG tcIhlti SEWER !saNn DEPTH DF
NO No SIZE
~~~-W R S
~ LENGTH BKTS RUB R f f i MET YO00 GLRSS ROOTS FRT BEN 1NV U W SLBB COV FRR STEP COP FLOu RS X JNTS PIPE OIL
329130 2 S 1.0 IS.% 1.S 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 19. 0 1 0
3-2131 S 0 1.0 70--% 1.0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1z. 0 0 0
3--91;2 3 A 1.0 57.79 1.0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 12. 0 6 0
z2913 1 A 1.0 60.01 1-s 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 6. 0 0 0
229139 2 S 1.0 A9.22 2.0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 19. 0 6 0
‘329100 2 5 1.0 75-68 2. S 0 0 1 0 0 0 0 0 0 0 0 0 0 0 6 22. 0 6 0
3,3101 2 S 1.0 A9.10 J. 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 29. 0 0 0

TOTnLSI 7.0 S77.60 12.S 1 2 1 0 1 0 0 1 0 0 0 0 0 1 1 6 1 6


ISEUER PlCIINlENRNCE RECORDS
c
0
N
86 T V 9 oo ' 0 00 ' 0 00 ' 0 86 'SVV s 'L :
- s7u101
vv ' V S t 00 ' 0 00 ' 0 00 ' 0 vv ' f 6 Z S T I 1x3 UISUN31 -f SS
vs ' 6 8 I 00 '0 00 ' 0 00 ' 0 9s '68 f 0 '* UI S U N 3 1 'OSZ
H18N31 +OOOI 666/00S 662/002 66f/001 StlllOH 3WUN 3a03NMOI
1UlOl :tj313wu1a do U 3 N U 3 1 3Stj3M3S do S H U N 3 1 NU313 d IHSNMOl
N
0
N
TABLE 12.6 Output f i l e

1 S E U L R OLOCKAGC R C C O R D S

R L P U E S I E D S T A C T DATE:- 87ClUS
*L:UC:TED CND D A T i :- 1175107

TOYNSnXP 52 - BOSlOWT 40. OF S T A N D S - 1418 AREA(WECTARES> - 106.1

BLOCKAGL PRIVATE: nA1M: I L O C K A G E CLEARED M O U I S FRO@! RAIN: CIIVATL:


DCPORTFD STAND SLYER START FINISH JOB I E P O R T l N G TO GANG GANG MOUSES 11110 MOUSE
MOUll DATE NO no DATF TlRf DATE T I R E TIRC CORPLETIOM SUMDAT NO S I Z E CAUSE L F F E C T E D FLOODED FLOODED

I ~ 7 ~ 1 ~ 0s z z ~ i m 8 7 r i . r ~ iz:40 87itios i 4 : o o 1.20 6 0 5 1 4 2 I! n

EC’PLAIhTS
--
Z O n P L A I h T : R E F i R R E O TO b A T E R S R A N C M
R L F E R R L O 1 0 R O A D S AND YORKS
0
D
I S E Y i l l BLOCKAGE R C C O R D S

R L C U E S l r O S l 8 R T DATE:- 117OlOf
D L i U C S T L D E k J C A T i :- 6 7 0 1 0 7

lOYNSHIP 157 - ELDORAEO PAPK 30. OF S T A W S - 19n9 AREACMECTLRES) - 128.6

@LOCKAGE PRIVATE: RAIN: BLOCKAGE CLELRED MOURS cnon MA1N: PRIVATE:


iEPOPTlL STAGO SEYER START Fltiisn JOB R E P O R T I N G TO GANG GAYG nousEs T A ~ D HOUSE
niui DATL ID L’J DATL TIBE DATE TINE TIME COfiPLETIOH . SUMDAT N O S l Z E CAUSE A F F E C T E D FLOODED FLOFDED
? C7ulub 4Jl o I I C I ? ~ I:DD 87~106 a:so 0.30 2 0 2 3 2 0 1 1
20 3
REFERENCES

Adams, B.J. and Zukovs, G., 1987. P r o b a b i l i s t i c models f o r combined sewer


systems r e h a b i l i t a t i o n s a n a l y s i s . In Beck (1987).
Beck, M.B. (Ed.) 1987. Systems A n a l y s i s in Water Q u a l i t y Management.
IAWPRC Conf. London, Pergamon
Fuchs, L., M u l l e r , D. a n d Neumann, A., 1987. L e a r n i n g p r o d u c t i o n systems
f o r the c o n t r o l o f u r b a n sewer systems. In Beck (1987).
S c h i l l i n g , W. a n d Petersen, S.O., 1987. Real time o p e r a t i o n o f urban
d r a i n a g e systems, v a l i d i t y a n d s e n s i t i v i t y o f o p t i m i z a t i o n techniques.
In Beck (1987).
Stephenson, D. a n d Hine, A.E., 1982. Computer a n a l y s i s o f Johannesburg
Sewers. Proc. I n s t n . Munic. Engrs. S.A. IMESAF, 7 ( 4 ) A p r i l . p13-23
Stephenson, D. a n d Hine, A.E. 1985. Sewer Flow Modules f o r V a r i o u s t y p e s
o f development in Johannesburg. Proc.. I n s t . Munic. Engrs. S.A. (10)
Oct. p31-41.
Stephenson, D. and Hine, A.E., 1987. Maintenance program for
Johannesburg Sewerage Systems.
Yen, B.C. ( E d . ) 1987. Proc. 4 t h I n t l . Conf. U r b a n Storm Water H y d r o l o g y
a n d D r a i n a g e , Lausanne.

TABLE 12.7 Outout f i l e

SLYER BLOCKAGE RECORDS

l E O U E S T E D S T l R l DATE:- 070105
SEOUESTED END DATE :- 870107

€AM: 1 SIZE: 4

NO. OF NO. OF TOTAL


ToYNsnIP PRIYITE nAIN NO. O f PRIVATE- MAIN- TOThL-
lOYNCODE lAME BLOCKbCLS BLOCKAGES BLOCKAGES JOB T I M E JOB T I M E JOE T I M E
155 E L D O R l D O PARK2 2 0 2 b.25 0.00 4.25
177 ELDORADO PARK4 4 0 4 7.25 0.00 7.25
342 K L I P S P R U I T YES 1 0 1 0.25 O.OG 0.25
TOTALS:- 7 0 7 12.15 0.00 12.15
204

CHAPTER 13

WATER QUALITY MON I TOR I NG NETWORKS


b y Thomas G. Sanders, Colorado State U ive i Y

NECESS I T Y FOR NETWORKS

Environmental legislation and general water quality awareness have


been responsible f o r recent increased m o n i t o r i n g a n d s a m p l i n g of water in
streams. Such m o n i t o r i n g and testing can be expensive and a scientific
approach to m i n i m i z i n g costs w h i l s t m a x i m i z i n g b e n e f i t s i s d e s i r a b l e .
The assumption that a water quality monitoring network can detect
trends in water quality, check compliance with stream standards, and
measure ambient water quality, etc., is incorporated into legislation for
water q u a l i t y management in the U n i t e d States. The legal view of water
qua1 i t y monitoring envisages conclusive information being generated to
actively guide government's water quality management efforts. When
implemented, however, water quality monitoring is viewed more from a
technical feasibi I i t y stand-point. That is, the problems involved in
obtaining conclusive information w i t h the available resources force many
compromises a n d half measures, the consequences of which a r e often not
f u l l y understood.
Monitoring performed by government agencies is, in many cases,
conducted over large geographic areas (defined by political and not
necessarily hydrologic boundaries) covering many k i lometres of streams.
Simply collecting samples in such a situation often becomes a major
problem; so major, i n fact, that i t becomes a n end in itself. In many
cases, little thought is given to the representativeness of the water
samples or types of data analysis techniques to be used or even the
ultimate use of the data. Consequently, the majority of resources are
devoted to c o l l e c t i n g d a t a as i t i s the most immediate problem.
By using most resources to physically collect water samples, little
resources a r e l e f t to consider the representativeness of the sample i n time
a n d space, d a t a a n a l y s i s o r d a t a use. A b a l a n c e d ( c o l l e c t i o n versus use)
m o n i t o r i n g system should therefore be developed so the e n t i r e m o n i t o r i n g
system should be examined and designed simultaneously (a systems
approach).
The purpose of t h i s chapter i s to r e v i e w the monitoring system and
then d e l i n e a t e the impacts t h a t such a systems a p p r o a c h of m o n i t o r i n g w i l l
h a v e on network design b y c o n s i d e r i n g the w a t e r q u a l i t y v a r i a b l e s to b e
205

monitored, the sampling location a n d s a m p l i n g frequency.

MONITORING SYSTEM FRAMEWORK

Before a monitoring network can be designed the goals of the


m o n i t o r i n g program should be delineated, and specific objectives applied.
I n addition, the decisions to be made based upon information from the
network and the subsequent actions should also be well developed p r i o r to
the collection o r a s i n g l e b i t of data.
The a c t u a l operation of a m o n i t o r i n g system can be categorized into
f i v e major functions:

1. Sample Collection
2. Laboratory Analysis
3. Data H a n d l i n g
4. Data A n a l y s i s
5. I nformation U t i I i z a t i o n

These f i v e functions serve as the feedback loop from in-stream water


quality conditions of water quality management decision making. A
management agency i s c o n s t a n t l y m a k i n g decisions (e.g. r e l a t i v e to site
approvals, regulations, pollution abatement, etc.) that affect water
qua1 i t y . Without a monitoring feedback loop accurately documenting the
effects of those decisions, the management's past success and future
direction are uncertain.
M o n i t o r i n g network design i s an o v e r r i d i n g a c t i v i t y ( c o v e r i n g the f i v e
operational f u n c t i o n s l i s t e d above) that should c a r e f u l l y i n t e g r a t e sample
collection (e.g. location a n d frequency) with the type of data analysis
used to o b t a i n the i n f o r m a t i o n r e q u i r e d a n d actually utilized in decision
making. Thus, the design of water q u a l i t y m o n i t o r i n g networks must take
i n t o account the decision m a k i n g process, the t y p e and level o f s t a t i s t i c a l
a n a l y s i s a p p l i e d to the d a t a , a n d u l t i m a t e use of the d a t a collected.

FACTORS I N NETWORK DESIGN

M o n i t o r i n g network design, as a p l a n n i n g / d e s i g n type function which


guides m o n i t o r i n g operations, can i t s e l f be broken down i n t o three m a j o r
componen ts:
1. Selection of Water Q u a l i t y V a r i a b l e s to Monitor
2. Sampling Station Location
206

3. Sampling Frequency

The term water quality variable is used instead of water quality


parameter because water q u a l i t y i s a random v a r i a b l e a n d c a n be d e f i n e d
by statistical parameters such a s the mean a n d standard deviation. In
addition, the term parameter is most often used to define constants of
deterministic equations or models and it can lead to confusion by
i d e n t i f y i n g i t a s a random v a r i a b l e .
Each of these factors in network design effects all the monitoring
system's o p e r a t i o n a l f u n c t i o n s I i s t e d p r e v i o u s l y a n d v i c e versa. The degree
of impact, however, depends upon the purpose a n d g o a l s of the m o n i t o r i n g
system.

SELECT ION OF WATER QUALITY VARIABLES TO MEASURE

The selection of the water q u a l i t y v a r i a b l e to be sampled w i l l depend


to a l a r g e extent on the objectives of the sampling network and the
background or frame of reference of the individuals responsible for
developing the o b j e c t i v e s of the monitoring network. When a sampl i n g
network has its primary objective to monitor compliance with stream
stndards, the v a r i a b l e s sampled a r e the ones s p e c i f i e d i n the l e g i s l a t i o n ,
for example, dissolved oxygen (DO). DO is sampled because stream
s t a n d a r d s specify a minimum l e v e l which should not be v i o l a t e d . Dissolved
oxygen a n d o t h e r v a r i a b l e s deemed most i m p o r t a n t a n d i n c l u d e d in stream
s t a n d a r d l e g i s l a t i o n were those r e l a t e d to water s u p p l y , col iform b a c t e r i a ,
biochemical oxygen demand (BOD), temperature, turbidity, and suspended
a n d dissolved solids, because most i n d i v i d u a l s e n t e r i n g the f i e l d o f water
quality management d u r i n g the last few decades have a background in
s a n i t a r y engineering.
Since i n d i v i d u a l s o t h e r t h a n besides s a n i t a r y ( e n v i r o n m e n t a l ) engineers
became interested i n water q u a l i t y , the number of water quality variables
which should be sampled routinely has increased. This compounding
syndrome cannot a n d should not be the major v a r i a b l e selection mode f o r a
permanent, routine sampling program, but instead can be easily
accommodated in the much discussed synoptic surveys. The increasing
p o p u l a r i t y of synoptic s u r v e y s w i t h s a m p l i n g agencies i s p r o b a b l y due to
the f a c t t h a t the s u r v e y s a r e in fact a n a p p l i c a t i o n of a systems a p p r o a c h
to water quality monitoring. Unlike the permanent, routine sampling
programs, the objectives, the use of the d a t a , the s a m p l i n g locations, the
sampling frequency, the variables to be sampled as well as the data
207

analysis procedures and decisions to be made should be developed


completely before the survey i s undertaken.
Both sampling location and sampling frequency can be developed
independently of the water quality variable to be analyzed, as both
location and frequency a r e specified f o r the c o l l e c t i o n of the water sample
( t h e analyses a r e made l a t e r ) . However, both c r i t e r i a a r e affected b y the
water quality variable being monitored. For example, sampling once a
week at a single point in a river may be more than adequate for
monitoring the relatively stable river temperature, but may be hardly
adequate for m o n i t o r i n g r a p i d l y v a r y i n g coliform bacteria concentrations.
Therefore, before a water q u a l i t y m o n i t o r i n g network can be designed in a
systematic fashion, the variables to be monitored should be s p e c i f i e d so
that t h e i r n a t u r a l and/or man-made variation in time a n d space can be
considered when designing the monitoring network. In addition to
considering water qua1 i t y variables, their respective units should be
delineated. Network design differs if a daily mean (flow weighted)
concentration is needed as opposed to an instantaneous grab sample
concentration, the former being a result of several samples with flow
measurements spaced d u r i n g a 24-hour period, while the latter comprises
only a single sample (generally in the daytime, between 8.00 a.m. and
4.30 p.m.1.
In reality, the specification of the water quality variable to be
monitored p r i o r to i n i t i a t i n g network design would be ideal. In p r a c t i c e ,
however, network design i s specified a n d one must know o r determine what
water quality variables can be accurately monitored with the existing
network.

SAMPL I NG STAT ION LOCAT ION

The location of a permanent sampling station in a water’ quality


m o n i t o r i n g network i s probably the most critical aspect of the network
design, b u t a l l too often never properly addressed. Expediency a n d cost
comprises lead i n many cases to s a m p l i n g from b r i d g e s o r near existing
r i v e r g a u g i n g stations. Whether the s i n g l e g r a b sample from the b r i d g e o r
the gauging station is truly representative of the water mass being
sampled is not known, but generally is assumed to be by both the
collectors and users of the water quality data. Using river stage for
e s t i m a t i n g discharge, measurement anywhere i n the lateral transect would
indicate exactly the river discharge. However, t h i s does not necessarily
follow when measuring water quality variable concentrations. In fact
208

F i g , 13.1 Macrolocation of Sampling Stations W i t h i n a R i v e r Basin


Using the Percent Areal Coverage a s the C r i t e r i a S p e c i f y i n g
Locat ion
209
research indicates the opposite, that rarely will a single sample be
i n d i c a t i v e of the average water q u a l i t y i n a r i v e r cross section.
Sampling locations for a permanent water quality network can be
c l a s s i f i e d i n t o two levels of design: macrolocation a n d microlocation, the
former b e i n g a f u n c t i o n o f the specific objectives o f the network and the
latter being independent of the objectives but a function of the
representativeness of the water sample to be collected.
The macrolocation within a river basin usually is determined by
political boundaries, a r e a s of major p o l l u t i o n loads, population centres,
etc. Macrolocation can be specified, a s well, a c c o r d i n g to percent areal
coverage u s i n g b a s i n c e n t r o i d s (Sanders et al, 1986). This methodology
locates sampling p o i n t s in a systematic f a s h i o n m a x i m i z i n g information of
the e n t i r e b a s i n w i t h a few s t r a t e g i c a l l y located stations. F i g u r e 13.1 is
an example of locating sampling stations using basin centroids and
sub-basin centroids w i t h percent a r e a l coverage a s the c r i t e r i a .
The procedure f o r l o c a t i n g sampling s t a t i o n s i s d e r i v e d b y d e t e r m i n i n g
the c e n t r o i d o f a r i v e r system. Each c o n t r i b u t i n g e x t e r i o r t r i b u t a r y (this
i s a stream without defined tributaries) i s given the v a l u e o f one; an
i n t e r i o r stream r e s u l t i n g from the intersection of two e x t e r i o r tributaries
would have a v a l u e of two. Continuing downstream i n the same manner, as
streams intersect, the r e s u l t a n t downstream s t r e t c h of r i v e r would h a v e a
value equal to the sum of the values of the preceeding intersecting
stream. At the mouth of the r i v e r , the v a l u e o f the f i n a l r i v e r section w i l l
be equal to the number o f c o n t r i b u t i n g e x t e r i o r t r i b u t a r i e s , 22 in F i g u r e
13.1. D i v i d i n g the v a l u e of the f i n a l s t r e t c h of the river by two, the
v a l u e of the c e n t r o i d of the b a s i n , 1 1 i s calculated. The section of r i v e r
h a v i n g a v a l u e equal to t h a t of the c e n t r o i d d i v i d e s the b a s i n into two
sections and i s the location of the sampling station with highest order
( t h e assumption i s made t h a t there e x i s t s a s a m p l i n g s t a t i o n a t the mouth
of the r i v e r b a s i n ) . I n many cases, when applying t h i s procedure to a
r i v e r basin, there i s u s u a l l y not a stream h a v i n g a v a l u e e q u a l to that
of the centroid. When this occurs, the stream segment having a value
closest to the c e n t r o i d i s chosen. The n e x t o r d e r o f sampling locations is
determined b y f i n d i n g the c e n t r o i d v a l u e of the two equal sections above
a n d below the initial river basin centroid. The procedure is continued
u n t i l a percentage of a r e a l coverage i s a t t a i n e d .
The percentage of area coverage specified b y the m o n i t o r i n g agency is
defined as the number of sampling stations d i v i d e d b y the m a g n i t u d e of
the basin. Intrinsic in this objective procedure is the concept of a
sampling station hierarchy that o r d e r s the importance of each sampling
210

station in the b a s i n (Sharp, 1973). T h i s p r o v i d e s a r e a l i s t i c methodology


i n which a r a t i o n a l implementation progam c a n proceed: the most important
stations (highest order) are built first and as the resources become
available, additional s t a t i o n s can be b u i l t . As each succeeding h i e r a r c h y
of s t a t i o n s a r e e s t a b l i s h e d the percentage of a r e a l coverage i s increased.
Having established the macrolocations within a river basin, the
microlocation specifies the river reach to be sampled while the
microlocation specifies the p o i n t i n the r e a c h to be sampled. This point is
t h e location of a zone in the r i v e r r e a c h where complete m i x i n g e x i s t s a n d
o n l y one sample i s r e q u i r e d from the l a t e r a l transect in o r d e r to o b t a i n a
representative (in space) sample. Being a function of the distance
downstream from the nearest o u t f a l l , t h e zone of complete m i x i n g can be
estimated u s i n g v a r i o u s methodologies.
Given the assumptions t h a t a p o i n t source pollutant distribution in a
stream approximates a Gaussian d i s t r i b u t i o n , a n d t h a t b o u n d a r i e s can be
modelled using image theory, the following equation can predict the
d i s t a n c e downstream in a straight, u n i f o r m channel from a point source
p o l l u t a n t to a zone of complete m i x i n g (Sanders et a l . , 1977).

(J 2u
- Y
LY - 2oy (13.1)

where L is the mixing distance for complete lateral mixing, a y is


Y
distance from source to farthest lateral boundary, u is mean stream
velocity and D i s the l a t e r a l t u r b u l e n t d i f f u s i o n coefficient.
Y
Estimates of D can be made u s i n g e q u a t i o n 13.2
Y

D = 0.23 du' (13.2)


Y

where d i s depth of flow u* i s shear v e l o c i t y g i s acceleration flow


due to g r a v i t y R i s h y d r a u l i c r a d i u s S i s slope o r t h e h y d r a u l i c g r a d i e n t
(Sanders e t al., 1977).
Unfortunately, there may not e x i s t in a g i v e n r i v e r reach any points
of complete m i x i n g due i n p a r t to the random n a t u r e of the aforementioned
mixing distance, inapplicability of the assumptions used in the
determination of the m i x i n g distance, o r more often t h a n not, not enough
river l e n g t h o r t u r b u l e n c e to assure complete m i x i n g within the s p e c i f i e d
river reach. On the o t h e r hand field v e r i f i c a t i o n of a completely mixed
zone p r i o r to l o c a t i n g a permanent s a m p l i n g s t a t i o n c a n be e a s i l y done b y
collecting m u l t i p l e samples in the cross section and analyzing the data
u s i n g a we1 I-known one- o r two-way a n a l y s i s of v a r i a n c e techniques.
21 1

If there is not a completely mixed zone in the river reach to be


sampled, there a r e three a l t e r n a t i v e s :
( 1 ) Sample anyway a t a s i n g l e p o i n t a n d assume i t i s r e p r e s e n t a t i v e ( t h i s
i s a general approach adopted t o d a y ) ;
( 2 ) Don't sample the r i v e r reach a t a l l , because t h e d a t a w h i c h would be
obtained does not represent the e x i s t i n g r i v e r quality, b u t only the
q u a l i t y o f the sample volume collected. In o t h e r words, the data is
useless;
( 3 ) Sample a t several p o i n t s in the l a t e r a l transect c o l l e c t i n g a composite
mean, which would be r e p r e s e n t a t i v e of the water q u a l i t y in the r i v e r
a t that p o i n t i n time a n d space.
I f the sample i s not r e p r e s e n t a t i v e of the water mass, the frequency of
sampling as well as the mode of data analysis, interpretation and
presentation and the realistic use of the data for objective decision
m a k i n g becomes inconsequential. I n s p i t e of t h i s f a c t , c r i t e r i a to e s t a b l i s h
s t a t i o n locations f o r r e p r e s e n t a t i v e s a m p l i n g h a v e received r e l a t i v e l y little
a t t e n t i o n from many i n s t i t u t i o n s a n d agencies responsible f o r water q u a l i t y
monitoring.

SAMPLING FREQUENCY

Once sampling stations have been located to ensure samples collected


a r e representative i n space, sampling frequency should be specified so
t h a t the samples a r e r e p r e s e n t a t i v e in time.
Sampling frequency a t each permanent sampling station w i t h i n a river
basin is a very important parameter which must be considered in the
design of a water q u a l i t y m o n i t o r i n g network. A l a r g e p o r t i o n of the costs
o f o p e r a t i n g a m o n i t o r i n g network is directly r e l a t e d to the frequency of
sampling. However, the reliability and utility of water quality data
d e r i v e d from a m o n i t o r i n g network i s l i k e w i s e r e l a t e d to the frequency of
sampling. Addressing this anomaly Quimpo (1968) summarized the
s i g n i f i c a n c e of sampling frequency a n d stated t h a t :
On the one hand, b y s a m p l i n g too often, the i n f o r m a t i o n
obtained i s r e d u n d a n t and t h u s expensive, a n d on the other
hand, sampling too i n f r e q u e n t l y bypasses some i n f o r m a t i o n
necessitating an extended p e r i o d of observation.
Significant as sampling frequency is to detecting stream standards
v i o l a t ion , m a i n t a i n i n g e f f I uent standards, a n d e s t i m a t i n g temporal changes
i n ambient water q u a l i t y , very l i t t l e q u a n t i t a t i v e c r i t e r i a which designate
a p p r o p r i a t e sampling frequencies h a v e been a p p l i e d to the design of water
21 2

quality monitoring networks. In many cases, professional judgment and


cost c o n s t r a i n t s p r o v i d e the b a s i s f o r s a m p l i n g frequencies. All too often,
frequencies are the same at each station and based upon routing
capabilities, once-a-month, once-a-week, etc. and although possibly the
only practical means to implement a sampl i n g program considering the
s t a t i s t i c a l b a c k g r o u n d o f d a t a collectors, there do e x i s t many q u a n t i t a t i v e ,
s t a t i s t i c a l l y meaningful procedures to specify s a m p l i n g frequencies a t each
station (Sanders and Adrian, 1978). The methods include specifying
frequencies as functions of the cyclic variations of the water quality
variable (Nyquist frequency), the d r a i n a g e b a s i n area and the r a t i o of
maximum to minimum flow (Pomeroy and Orlob, 19671, the confidence
i n t e r v a l o f the a n n u a l mean (Ward et al, 1976; L o f t i s a n d Ward, 1978),
the number of d a t a p e r y e a r f o r hypotheses (Sanders and Ward, 1978), and
the power of a test measuring water quality intervention (Lettenmaier,
1975).
A l l of the aforementioned procedures can b e a p p l i e d to the design of a
water q u a l i t y m o n i t o r i n g network w i t h each r e q u i r i n g a d i f f e r e n t level of
statistical sophistication insofar as data requirements as well as
assumptions app I y .
One of the simplest approaches is to assume that the water quality
variable concentrations are random, independent and identically
distributed (iid) and determine the number of samples per year as a
f u n c t i o n o f an a l l o w a b l e ( s p e c i f i e d ) confidence i n t e r v a l of the mean a n n u a l
concentration ( t h i s i s analogous to the procedure f o r d e t e r m i n i n g how many
analyses of a water sample should be made to determine a reasonable
estimate o f the mean water q u a l i t y v a r i a b l e c o n c e n t r a t i o n ) .

n = [ aizS] (13.3)

where n i s the number of e q u a l l y spaced samples collected p e r y e a r , taI2


is a constant which is a function of the l e v e l of significance and the
number of samples, S is the standard deviation of the water quality
concentrations a n d R i s s p e c i f i e d h a l f - w i d t h of the confidence interval of
the a n n u a l mean.
Using the same assumption, t h a t the water quality variable i s iid, the
number of samples p e r year can be s p e c i f i e d a s a function of the data
a n a l y s i s procedure as well. For example, if annual means were to be
tested f o r s i g n i f i c a n t changes u s i n g the d i f f e r e n c e in means, then to detect
a n assumed level of change, t h e number of samples c a n be specified.
A more sophisticated procedure, representing a higher level of
21 3

0.9

0.8

R vs. Number of Somples per Yeor


I Wore
0.7 2 Conn. at Thompsonville
3 Deerfield
4 Conn. ot Montopue City
5 Millers
0.6 6 Conn.ot Vernon
7 Westfield
8 Conn. ot Turners Falls
R
0.5

0.4

0.3

0.2

I 1 I I I
0. I
10 20 30 40 50
Number of Somples per Yeor

Fig 13.2 A p l o t n u m b e r o f s a m p l e s per y e a r of the expected


h a l f - w i d t h of t h e c o n f i d e n c e i n t e r v a l of m e a n log
f l o w , R , v e r s u s n u m b e r of S a m p l e s for S e v e r a l R i v e r s
in t h e C o n n e c t i c u t R i v e r B a s i n
214

statistical analysis, would be to recognize that water quality veriables


may not be i i d , b u t h i g h l y dependent, not i d e n t i c a l l y d i s t r i b u t e d , having
seasonal v a r i a t i o n , a n d determine s a m p l i n g frequency as a f u n c t i o n of the
variability of the water quality variable time series after trend and
p e r i o d i c components h a v e been removed. Unfortunately, other than mean
daily discharge, data bases of water quality variable of sufficient
number, r e l i a b i l i t y a n d l e n g t h a r e g e n e r a l l y not a v a i l a b l e f o r a p p l i c a t i o n
of t h i s procedure.
Once a uniform sampling frequency criterion is selected it can be
utilized to objectively distribute sampling frequencies within a water
quality m o n i t o r i n g network. For example, the expected h a l f - w i d t h of the
confidence interval of the annual mean (for specifying sampling
frequencies) approach c a n be a p p l i e d basin-wide in a consistent fashion
by specifying equality of these expected half-widths at each sampling
station. Thus, stations where water quality varies tremendously will be
sampled more frequently than stations where the water quality varies
little. With reference to Figure 13.2 which is a plot of the expected
half-width of the confidence interval of mean log r i v e r flow versus the
number of samples per year, the number of samples collected at each
s t a t i o n w i t h i n the r i v e r b a s i n f o r a g i v e n R a r e determined b y d r a w i n g a
horizontal line through R and reading the number of samples on the
abscissa axis below the intersections on the horizontal line with each
curve. Figure 13.2 may also be used i n an i t e r a t i v e fashion to specify
s a m p l i n g frequencies a t each s t a t i o n when a t o t a l number o f samples from
the b a s i n i s specified. For example, if only N samples per year were
collected a n d analyzed, a v a l u e of R i s assumed and a line is drawn
horizontally; the number of samples s p e c i f i e d by the i n t e r s e c t i o n of the
curves a r e summed a n d compared to N. If the sum were not e q u a l to N
then another estimate of R would be made u n t i l the sum of a l l the samples
i s equal to N.
I t should be noted t h a t the expected h a l f - w i d t h o f the a n n u a l mean i s
not the o n l y s t a t i s t i c that can be used to specify s a m p l i n g frequencies;
the expected h a l f - w i d t h d i v i d e d b y the mean i s a measure o f r e l a t i v e e r r o r
a n d may be more a p p r o p r i a t e when assigning sampling frequencies in a
b a s i n where water q u a l i t y v a r i e s tremendously from r i v e r to r i v e r .
When developing s a m p l i n g frequencies, one must keep i n m i n d two v e r y
important cycles which can have immense impact on water quality
concentrations, the d i u r n a l c y c l e a n d the weekly cycle. The effect of the
d i u r n a l cycle (which i s a f u n c t i o n of the r o t a t i o n of the e a r t h ) can be
e l i m i n a t e d b y s a m p l i n g in e q u a l time i n t e r v a l s f o r a 24-hour period and
215

the effect of t h e weekly c y c l e ( w h i c h i s a f u n c t i o n of mans' a c t i v i t y ) can


be eliminated by specifying that sampling i n t e r v a l s for a network cannot
be m u l t i p l e s of seven, and occasional sampling on weekends would be
necessary.
Perhaps the major impact between network design in terms of v a r i a b l e s
to be monitored, sampling location, and sampling frequency and the
operational monitoring functions is in the area of data analysis and,
consequently, ultimate v a l u e of the monitoring network information. Any
sampling program that i s to generate conclusive r e s u l t s from o b s e r v i n g a
stochastic process ( w a t e r q u a l i t y concentrations) must be well planned and
s t a t i s t i c a l l y designed. S t a t i s t i c a l l y designed implies that the sampling is
p l a n n e d ( i n p r o p e r locations and numbers) so t h a t the s t a t i s t i c a l analysis
techniques chosen w i l l be a b l e to y i e l d q u a n t i t a t i v e information. Thus, the
d a t a a n a l y s i s techniques ( l e v e l and t y p e of s t a t i s t i c s ) to be used must be
defined in order to know how to compute proper sampling frequencies,
locations, etc.

D ISCUSS ION

The above section has pointed out many problems due to not d e s i g n i n g
a m o n i t o r i n g system in a systems context. Perhaps the major concern is
that all aspects of a monitoring program should match in terms of
accuracy. For example, i t would not be wise to use time series analysis
on nonrepresentative, grab sample d a t a . The system would be providing
excessive accuracy i n one segment compared to the accuracy in another
segment .
I n a s i m i l a r manner, i t may be u n r e a l i s t i c to encourage use of more
sophisticated sample collection and laboratory a n a l y s i s techniques if the
d a t a i s not to receive a thorough s t a t i s t i c a l a n a l y s i s .
It i s difficult to test hypotheses, make decisions and i n i t i a t e action
u s i n g water q u a l i t y d a t a which a r e collected o n l y in the daytime a n d not
flow weighted, several times a year, from locations which are not
completely mixed a n d u s i n g l a b analyses procedures which may h a v e more
variation in their results when analyzing the same sample than the
ambiant v a r i a t i o n of the water q u a l i t y v a r i a b l e in the r i v e r .
Perhaps an even l a r g e r concern to those in m o n i t o r i n g network design
i s the use of water quality standards that generally ignore statistics.
T h i s lowers the v a l u e of a n y information from a compliance v i e w p o i n t , to
t h a t of spot checks. I n c o r p o r a t i n g water q u a l i t y means a n d v a r i a t i o n into
standards would greatly facilitate incorporating more statistics into
216

m o n i t o r i n g . T h i s would h a v e t h e effect of t y i n g network design to d a t a use


in a much more concrete, s t a t i s t i c a l manner t h a n i s now possible. I t would
a l s o encourage use of the system approach to network design as there
would be a statistical thread moving through the entire monitoring
operat ion.

REFERENCES

Lettenmaier, D.P., 1975. Design of M o n i t o r i n g Systems f o r Detection of


Trends i n Stream Q u a l i t y . Technical Report No. 39, Charles W. H a r r i s
H y d r a u l i c s L a b o r a t o r y , U n i v e r s i t y of Washington, Seattle.
L o f t i s , J.C. a n d Ward, R.C., 1978. S t a t i s t i c a l Tradeoffs i n M o n i t o r i n g
Network Design, presented a t AWRA Symposium Establishment of Water
Q u a l i t y M o n i t o r i n g Programs. San Francisco, C a l i f o r n i a .
Pomeroy, R.D. a n d Orlob, G.T., 1967. Problems of S e t t i n g S t a n d a r d s o f
S u r v e i l l a n c e f o r Water Q u a l i t y Control. C a l i f o r n i a State Water Q u a l i t y
Control Board P u b l i c a t i o n No. 65, Sacramento, C a l i f o r n i a .
Quimpo, R.G., 1968. Stochastic A n a l y s i s of D a i l y R i v e r Flows. Journal o f
H y d r a u l i c s , ASCE. 94(HY1) p43-47.
Sanders, T.G., A d r i a n , D.D. a n d Joyce, J.M., 1977. M i x i n g L e n g t h f o r
Representative Water Q u a l i t y Sampling. Journal Water P o l l u t i o n Control
Federation. 49 p2467-2478.
Sanders. T.G. a n d Ward, R.C., 1978. R e l a t i n g Stream Standards to
Regulatory Water Q u a l i t y M o n i t o r i n g Practices. Presented a t the AWRA
Symposium “Establishment of Water Q u a l i t y M o n i t o r i n g Programs, San
Francisco, Ca I i f o r n i a .
Sanders, T.G. and Adrian, D.D., 1978. Sampling Frequency f o r R i v e r
Q u a l i t y M o n i t o r i n g . Water Resources Research. 1 4 ( 4 ) p 569-576.
Sanders, T.G., Ward, R.L. L o f t i s , J.G. Steel, T.D, Adrian, D.D. and
Yevjevich, V., 1986. Design of Networks f o r M o n i t o r i n g Water Q u a l i t y ,
2nd E d i t i o n , Water Resources P u b l i c a t i o n s , Colorado.
Sharp, W.E., 1973. A T o p o l o g i c a l l y Optimum R i v e r Sampling P l a n f o r South
C a r o l i n a . Water Resources Research I n s t i t u t e Report No. 36, Clemson
U n i v e r s i t y , Clemson , South Carol i n a .
Ward, R.C., Neilsen, K.S. a n d Bundgaard-Nielsen, M., 1976. Design of
M o n i t o r i n g Systems f o r Water Q u a l i t y Management. C o n t r i b u t i o n f o r the
Water Q u a l i t y I n s t i t u t e , Danish Academy of Technical Science, No. 3,
Horshdm, Denmark.
21 7

AUTHOR INDEX
Kemp, P.H. 64
Kim, J.I. 70
Abulnour, A.M. 116
Kleinecke, D. 41
Adarns, B.J. 190
Adarnson, P.T. 76
Lance, J.C. 143
A d r i a n , D.D. 209,212
Larnbert, J.L. 66
A g a r d y , F.J. 66
Larnbourne, J.J. 66
American Water Works Association 37
L a n g e l i e r , W.F. 3, 5, 6
A r n o l d , R.W. 36
L a n y o n , R. 75
L a r s o n , T.J. 104
Baker-Duly, H.L.G. 123
L a u r i a , D.T. 165
B a l l , J.M. 70, 77
L e i g h t o n , J.P. 146, 149
B a r e n b r u g , A.W.T. 2
Lettenmaier, D.P. 212
Bauer, C.S. 143, 146, 149
Lewis, R.W. 119, 149
Beck, M.B. 202
L l o y d , P.J. 1
Bedient, P.B. 66
L o f t i s , J.C. 209, 212
Betz, 3
Loucks, D.P. 116
Bishop, A.B. 165
L u d w i g , L. 9
Boyd, G.B. 66
.
B r a d f o r d , W. J 64
L y n n , W.R. 116
B r e b b i a , C.A. 62
M a d i s h a , J.L. 75
Brownlow, A.H. 1
Mathew, K. 143
Bungaard-Nielsen, M. 210
McDonell, D.M. 56
Chan, W.Y.W. 167
McPherson, D.R. 41, 45
Chiang, C.H. 165
M i c h a i l , M. 143
CIRIA. 164
M i k a l s e n , K.T. 75
C o l w i l l , D.M. 66
Mrost, M. 1
Connor, J.J. 62
MOller, D. 190
Corbetis, S. 116
Cordery, I. 70
Neilsen, K.5. 210
Crabtree, P.R. 167
Neurnann, A. 190
Newrnan, P.W.G. 143
D e i n i n g e r , R.A. 36, 39, 51
D a n t z i g , G.B. 82, 163
O'Conner, B.A. 56
Orlob, G.T. 212
F r i e d , J.J. 55
Fuchs, L. 190 P a l i n g , W.A.J. 141, 143, 145
P e l l e t i e r , R.A. 1
G i l b e r t , R.G. 143 P e r r y , R. 66
Goodier, J.M. 63
Peters, C.J. 66
Green , I. R.A. 64
Petersen, 5.0. 190
G r i z z a r d , T.J. 70
P o l l s , I. 75
Grosman, D.D. 86
Pomeroy, R.D. 212
Porges, J. 2
H a d l e y , G. 162
P r a t i s h t h a n a n d a , S. 165
H a l l , G.C. 160
Helsel, D.R. 70
Quimpo, R.G. 211
Henderson-Sel l e r s , B. 24
H i l t o n , E. 27, 119
R a n d a l l , C.W. 70
Hine, A.E. 197
Rand Water B o a r d 155
Hinton, E. 149
Revelle, C.S. 116
Ho, G.E. 143
Rice, R.C. 143
Hoehn, R.C. 70
Rinaldi, S. 116
Holton, M.C. 75
Ryzner, J.W. 36
Hunter, J.V.I. 66

IBM 162 Sanders, T.G. 24, 209, 210, 212


S a r t o r , J.D. 66
Idelovitch, E. 143
S c h i l l i n g , W. 190
S h a r l a n d , P.J. 41, 45
Joyce, J.M. 210
Sharp, W.E. 210
21 8

S h a w , V.A. 167
Shoemaker, C.A. 146, 149
Simpson, D.E. 64
Smeers, Y. 116
Smith, A.A. 119, 149
Soncini-Sessa, R. 127
South A f r i c a n Bureau of S t a n d a r d s 72
S p r i n g e r , N.K. 66
Steel, T.D. 209, 212
Stehfest, H. 127
Stephenson, D. 27, 66, 80, 81, 82,
115, 116, 117, 163, 175,
197, 200

T e r s t r i e p , 66
Thomann, R.V. 39
Timoshenko, 5 . 55
Tyteca, D. 116

U h l i g , H.H. 13
Van Staden, C.M.V.H. 2
Velz, C.J. 41

W a n i e l i s t a , M.P. 64, 146, 149


Ward, R.C. 209, 210, 212
Whipple, W. 66
Wang, L.K., 167

Yen, B.C., 190


Yevjevich, V . 209
Y u , S.L. 66

Zukovs, G. 190
21 9

SUBJECT INDEX
C o r r e l a t i o n 66
Acid 1 Corrosion 3, 13
Additives 6 Cost 79, 107, 146
Advection 21, 52 C r i t e r i a 211
Aerobic 9 Crop 17
A g r i c u l t u r e 17 Crump w e i r 65
Antecedent moisture 66 Crystal 7
Air 1 C y a n i d e 16
Alkalinity 3 Cycle 214
A l l o c a t i o n 79
A l l o y 13 D a t a 177, 204
Ammonia 9 Dead water 24
Anaerobic 9 Decomposition p r i n c i p l e 163
Analyses 195 D e s a l i n a t i o n 99, 115
A n a l y t i c a l 39 D e t e r i o r a t i o n 116
Apartments 167 D i f f u s i o n 36
A q u i f e r 141 Disc 128
Arsenic 17 Dispersants 7
A r t i f i c i a l r e c h a r g e 141 Dispersion 21, 166
D i s t i l l a t i o n 101
Backwater 193 DO ( d i s s o l v e d o x y g e n ) 37, 206
B a c t e r i a 9, 206, 207 Dissolved s o l i d s 206
B a r i u m 16 Downstream 193
B a s i n 209 D r y d a y s 66
Benefits 126 D r y weather 77
Bicarbonate 67
Biocide 9 Economics 99
Blend 89 E l e c t r i c a l corrosion 14
Blowdown 2 E Iect r o d ia I v s i s 105
BOD (biochemical oxygen demand) 37 Emulsion 10
Booster 152 Env ironmenta I 193
Bottleneck 173
Boundaries 62 Equipment 107
Bremen 193 E r r o r 91
B r i n e 104, 122 E s t u a r i e s 37
E u l e r 57, 59
Calcium carbonate 4 Evaporation 2
C a l i b r a t i o n 40 E x p l i c i t 39, 51
C a p i t a l 107, 157
Carbonaceous 38 F a l l o u t 66
Cathode 10 F a r a d a y s l a w 14
Catchment 64 Feedback 205
Cellulose acetate 104 F i e l d 45
C h a r a c t e r i s t i c 39 F i n i t e d i f f e r e n c e 55
Chelant 7 F i n i t e elements 62
Chemical 67 F i r s t f l u s h 70
C h l o r i d e 2, 67 F l o o d i n g 193
Chlorine 9 Flow 166
C i v i l e n g i n e e r i n g 107 Foam 8
Commerci a I 1 70 F o r m u l a t i o n 88
Cleaning 115 Fouling 9
Computer 20, 115, 128 Four p o i n t 51
Concentration 2, 71, 159, 212 F o u r i e r series 54, 169
C o n d u c t i v i t y 18 Freezing 103
Conduit 175 Frequency 21 1
Confidence 214
C o n s t r a i n t s 41, 86 Gain 23
Conveyance 141 G a l v a n i c corrosion 13
Cooling 20 Geochemical 1
Geohydrology 41
220

G r a p h i c s 118, 177 Nodes 119, 128, 160


Groundwater 98, 112, 143 Non c o n s e r v a t i v e 35
Gypsum 6 Numerical 23, 51
d i f f u s i o n 35
H i e r a r c h y 209
H i l l b r o w 68 O b j e c t i v e 41
H y d r a u l i c 51, 167 O i l 10
Hydrodynamic 56 O p e r a t i n g 157
H y d r o g r a p h 166 Optimum 79
O p t i m i z a t i o n 116, 152, 162
IBM 150 Ore 30
I m p l i c i t 55 Oxygen 10, 37, 40
I n d u s t r i a l 1, 104, 112, 172
I n f i l t r a t i o n 143, 176 Peak 146, 174
I n f l o w 177 PH 3
I n a c c u r a c y 52 Phenol 16
I n s t a b i l i t y 55 Phosphate 7
I n t e g e r Programming 141, 149 Photosynthesis 46
I n t e r e s t r a t e 107 P i p i n g 2, 146
I o n exchange 105 P l a n n i n g 149
I r o n 3, 16 P l a n t 122
I r r i g a t i o n 17 P l u g f l o w 21
I t e r a t i o n 165 P o l l u t i o n 1 , 64
Pol l u t o g r a p h 23
Johannesburg 167 Polymer 7
Polyphosphate 7
K l i p r i v e r 40 P o p u l a t i o n 166
P o t a b l e 15
L a b o r a t o r y 205 P o u r b a i x d i a g r a m 12
L a b o u r 108 P r o b a b i l i t y 167
L a n g e l i e r index 5 P r o d u c t i o n system 190
L a x a t i v e 16 P r o g r a m 122, 128, 136, 174, 179
Leach 1 , 26, 75 P u r i f i c a t i o n 143
L e a d 17
L e a k 166, 176 Rand Water Board 157
L e a p f r o g 51 Random 212
Least squares 42 Raw water 122
L e g i s l a t i o n 204 Reaction 14
L i n e a r p r o g r a m m i n g 43, 85 Recharge 144
L o a d f a c t o r 107 Recovery r a t i o 1 1 1
Loops 119 Reed beds 40
Regional 155
Maintenance 116 Regression 67
Make-up 26, 33 R e l i a b i l i t y 211
Manhole 167 Reservoir 23
Mass b a l a n c e 20, 35, 64, 72, 161 R e s i d e n t i a l 167
Master programme 163 Re-use 99
Mathematical models 20, 149, 158 Reverse osmosis 81, 104
Measurement 167 R i v e r s 37, 214
Membranes 105, 108 R o u t i n g 166, 175
Meta p r o d u c t i o n 191 R u l e base 191
M i n e water 26, 117, 123 Runge Cutte 61
M i n i m i z e 41 R u n n i n g 108
M i x e d f l o w 21 Runoff 67
M o n i t o r i n g 204 Ryzner i n d e x 3
M u l t i - s t a g e f l a s h d i s t i l l a t i o n 103
M u l t i step 61 S a l t s 102
S a n i t a t i o n 195
Network 146, 205 S a n i t a r y e n g i n e e r i n g 206
N i t r a t e 17, 72 Sample 68, 204, 205
221

Sampling frequency 206 Washoff 67


Scale 102
Waste t i p 65
Scaling 3 Waste water 99, 155
Sea water 101 Water resources 79
Sediment 8
Water s u p p l y 116
S e n s i t i v i t y 195
Waterways 190
Separable programming 81, 95 Water v a p o u r 2
S e n s i t i v i t y 95, 165 Welding 13
Sewage 144, 176 W i t w a t e r s r a n d 155
Sewer 72, 166, 190, 196, 198
Shadow v a l u e 165 Zeolites 107
Shops 169 Z i n c 16
Simulation 31, 51, 166 Zooming 56
Simplex method 89
S i n k 42
Slack 82, 160
Software 195
Solution 82, 160
Source 43
Standards 15, 141
Station 210
S t a t i s t i c a l 205
S t a t i s t i c s 215
Steady s t a t e 20
Stormwater 64, 77, 166, 176
Stream 159, 204
Stream gauge 66
Streeter Phelps e q u a t i o n 37
Sub-programme 165
Sub-division 173
Sulphate 5, 16, 30, 67
Surcharge 168
Suspended 206
System 80
Systems a n a l y s i s 24, 118

Tape 128
Taste 16
T a y l o r series 53
TDS ( t o t a l d i s s o l v e d s o l i d s ) 2, 95
Temperature 3, 206, 107
Terminal concentration 24
Time l a g 166
Topography 1%
Toxic 16
T r a f f i c 69
T r a n s p o r t a t i o n p r o g r a m m i n g 80
Treatment 141, 155, 157
T u r b i d i t y 206

Turbulence 8
Two step 39, 52

U n p r e d i c t a b l e 64
Upstream 193

V a a l r i v e r 155
Vapour compression 102
Vegetables 18
Ventilation 2
This Page Intentionally Left Blank

You might also like