WATER POLLUTION CONTROL RESEARCH SERIES • 16110 EAX 02/72
Basin Management
For Water Reuse
U.S. ENVIRONMENTAL PROTECTION AGENCY
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes the
results and progress in the control and abatement of pollution
in our Nation's waters. They provide a central source of
information on the research, development, and demonstration
activities in the water research program of the Environmental
Protection Agency, through inhouse research and grants and
contract^ vith F*?^*51^! } State, ?.r.d local agencies, research
institutions, and industrial organizations.
Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Chief, Publications Branch
(Water), Research Information Division, R&M, Environmental
Protection Agency, Washington, D.C. 20460.
-------
BASIN MANAGEMENT FOR WATER REUSE
by
ALAMO AREA COUNCIL OF GOVERNMENT'S
Three Americas Building
San Antonio, Texas 78205
for the
OFFICE OF RESEARCH AND MONITORING
ENVIRONMENTAL PROTECTION AGENCY
Proiect #16110 EAX
February 1972
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, B.C., 20402 - Price $2.25
-------
EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents necessarily
reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or
commercial products constitute endorsement or recommenda-
tion for use.
ii
-------
ABSTRACT
Computer programs were developed for the preliminary design and
costing of wastewater renovation by the lime-clinoptilolite-
carbon processes of advanced waste treatment; for activated sludge
treatment; and for pipeline conveyance of water. These together
with methods or algorithms of lesser depth for other processes
were used to cost water supply and waste treatment under conditions
expected in San Antonio in the year 2000 for two extreme alterna-
tives, one importation of surface water according to the Texas
Water Plan and conventional water treatment, waste treatment and
disposal by discharge; the other completly closed recycle,
discharging no waste water and reusing all the waste water after
treating it to make it reusable. The unit costs for these two
extremes were about 20<:/kilogallon of water used and the reuse
scheme was only 10% more costly than the conventional scheme, i.e.
well within the expected error of the estimates. It was shown
that the seasonality of the water consumption in the face of
non-seasonality of the sewage produced has an important bearing
on the design and cost of reuse systems.
This report is submitted toward fulfillment of Grant No.16110 EAX
under the partial sponsorship of the Environmental Protection
Agency.
111
-------
TABLE OF CONTENTS
Chapter Page
Conclusions 1
Recommendations 3
1 Summary of Project 5
2 Computer Program for Design and Costing of the
Lime-Clinoptilolite-Carbon Process for Waste
Water Renovation 45
3 The Logistics of Municipal Recycle Illustrated
by the San Antonio Supply in the Year 2000 121
4 Computer Program for Design and Costing of
Conveying Water by Pipeline 179
5 Computer Program for Preliminary Design and
Costing for Activated Sludge Treatment 229
6 Conventional or Reuse? A Cost Comparison for
Municipal Reuse in San Antonio in the Year 2000 263
7 Acknowledgments 285
8 References 28?
v
-------
FIGUP:
Page
1 Process Element Sequence in Municipal Recycle 8
2 Bexar County Annual Water Use 15
3 Long Term Water and Sewage Relations, San Antonio 16
4 Monthly Average Per Capita Water Withdrawal and
Sewage Delivered - San Antonio 1961-65 17
5 Monthly Water Patterns - 2000 A.D. San Antonio Basin 19
6 Seasonal Changes in Blend Composition -
San Antonio 2000 27
7 Cost of AWT by the Lime-Clinoptilolite-Carbon Process
San Antonio Sewage 1968 Utilization Factor, UBAR = 0.5
FCODAC =0.8 32
8 Cost of Conveying Water by Pipeline UBARE = 0.5
Effect of QBARE and Slope 34
9 Lime Stage Production Cost vs QBARE and Investment
vs QDOT 60
10 Lime Stage Production Costs C/Kgal, 1969 61
11 Clinoptilolite Process - Production Costs &
Investment vs QBARE 70
12 Sensitivity of Clinoptilolite Process to Various
Parameters 71
13 Activated Carbon - Investment & Production Costs 80
14 Effect of CODIN, CT, and CLF, etc on Total Production
Costs for Activated Carbon Plants 81
15 Cost of AWT by the Lime-Clinoptilolite-Carbon Process
San Antonio Sewage 1968 88
16 Sensitivity of AWT Costs to Utilization Factor and to
Fraction of COD removed by Accelators 89
17 Process Element Sequence in Municipal Recycle 122
18 San Antonio Recycle Mineral Quality as Influenced
by Lawn Watering and Sewer Losses 126
VI
-------
Figures, continued Page
19 Water-Sewage Loss Ratios, San Antonio 128
20 Monthly Average Per Capita Water Withdrawal and
Sewage Delivered - San Antonio 1961-65 130
21 Flow Diagram Municipal Recycle 133
22 Bexar County Annual Water Use 139
23 Long Term Water and Sewage Relations, San Antonio 142
24 Annual Rainfall vs Loss Ratio Relations,San Antonio 143
25 Long Term Annual Loss Ratio - SanAntonio 144
26 Blend Composition at Various Loss Ratios - SanAntonio 151
27 Seasonal Changes in Blend Composition-
San Antonio 2000 153
28 Monthly Average Per Capita Excess of Water Withdrawal
over Sewage Delivered. San Antonio 1961-65 155
29 Monthly Water Patterns 2000 A.D. SanAntonio Basin 156
30 Simplified Flow Diagram for Illustrating Mathematics
of Recycle 169
31 LP Diagram for Recycle (Schematic) 171
32 Contribution of Cost Elements to Conveyance Costs
Horizntal Lines, 1968, National 196
33 Cost of Static Lift Regardless of Flow Distribution 198
34 Cost of Conveying Water by Pipeline UBARE =0.5
Effect of Qbare and Slope 200
35 Contributions of Cost Elements to Conveyance Cost
Inclined Lines 201
VII
-------
TABLES
Page
1 Logistic of the New Supply 23
2 Logistics of the New Supply - 2000, Sewage Treatment
Plant Alternatives 24
3 Logistics of the New Supply -2000, Ground Water
Facility Alternatives 25
4 Logistics of the New Supply - 2000, Conveyance
Alternatives 25
5 Cost Summary 27
6 Components Remaining Unresolved 38
7 Operating Labor Estimate, 10 mgd Plant 48
8 Nonoperating Labor Estimate, Man hours/day 48
9 Total Labor Estimate, Man hours/day 48
10 Blue Plains Performance Data on Raw Sewage 51
11 Lime Clarification Stage Assumptions for Base Case
Performance 54
12 Clinoptilolite Stage Assumptions for Base Case
Performance 64
13 Energy Consumption, KWH/Kgal 75
14 Parameter Values Used in Base Case Activated
Carbon Runs 79
15 Typical San Antonio Sewage Composition Used in
Exemplary Runs 83
.16 Effect of QBARE on AWT Cost 87
17 Municipal Increment, San Antoriio and Western
Cities Average 147
18 Values Used in Composition Study WQMONSA 150
19 Seasonal Logistics of Water Supply, Bexar County 2000
Conventional vs Reuse 157
20 Logistics of the New Supply - 2000 160
21 Logistics of the New Supply -2000, Sewage and
Water Treatment Plant Alternatives 162
Vlll
-------
Tables, continues
Page
22 Logistics of the New Supply -2000, Ground Water
Facility Alternatives 163
23 Logistics of the New Supply -2000, Conveyance
Alternatives 164
24 Types of Conveyance Situations 186
25 Characteristics of Optimized Conveyance in
Horizontal Pipelines 195
26 Effect of Conveyance Rate on Cost of Static Lift 199
27 Costs of Conveyance in Future Years, Horizontal
Line, 100 mgd 202
28 Reservoir Costs 235
29 Comparison TCB (1967) vs LK-R (1969 Cuero-Cibolo-
Hildebrand 264
30 Investment in Recent San Antonio Well Stations 268
31 Quantities to be Demineralized (45% Removal) For
500 mgpl 270
32 Cost Summary 274
IX
-------
CONCLUSIONS
In the design of municipal water reuse systems a very
important factor is the seasonality of the water consumption
in the face of the relative constancy of the sewage production.
This situation means that there is a seasonal variation in the
amount of new makeup water required in the system. This
seasonality makes for a lower utilization factor and thus a
higher cost than would be estimated from constant annual
averages. Furthermore the seasonality brings about a
continually fluctuating quality in the blend water being used
or a continually fluctuating requirement for demineralization.
The magnitude and seasonal nature of these effects depend upon
the makeup quantity and the relative quality of the makeup
water and the required blend as well as on the characteristic
impurity increment added to the water by a single municipal
use. In San Antonio the conditions are such that without
discharge and without demineralization the water would be
better with respect to inorganic contaminants in the summer
than in the winter, levelling off at about 1000 mgpl (milli-
grams per liter) in the winter and about 500 mgpl in the
summer. To meet a typical quality standard of 500 mgpl would
require no demineralization during two summer months and de-
mineralization to the extent of removing 45 - 70% of the
contaminants on all of the recycled water during two winter
months.
The cost of supplying the projected San Antonio water supply
and waste treatment at the quality and cost levels of 1969
and at the quantity levels estimated for the year 2000 by the
conventional means of importation and sewage treatment and
discharge according to the current Texas Water Plan would be
within 10% of the cost of not discharging any wastes but
treating all sewage by advanced waste treatment and reusing
the product water. The 10% difference is well within the
estimating error which means that the complete reuse cost is
comparable with the conventional importation and discharge
cost. (Certain alternatives to the Texas Water Plan are said
to have lower costs for a conventional system.)
The San Antonio conditions are favorable for the conventional
scheme because the makeup water would be partly from local
ground water which is relatively inexpensive. (The
conventional scheme requires more makeup than the reuse
scheme.) In cities not having access to ground water the
reuse scheme would have that additional advantage.
-------
RECOMMENDATIONS
1. In their planning for future water and waste water
management, municipalities should utilize preliminary
estimating and comparison methods, as partly worked out
herein and elsewhere to assess the relative economics
of advanced treatment and recycling of waste waters in
comparison with the conventional methods of seeking
additional water sources and discharging waste to streams.
2. Most of the seasonality in water consumption comes from
irrigation of urban and suburban lawns. It is not meant to
derrogate the aesthetic and the psychological value of lawns,
but using the methods of this report some exemplary studies
should be made to determine the dollar cost to the community
of having (and irrigating) lawns. The seasonality, of course,
would be reduced as well as the total water consumption and
thus the community costs would be reduced. Comparison of
the costs for the actual consumption pattern with the
corresponding costs for a hypothetical pattern from which
the lawn losses had been eliminated would provide the
information on just what lawns are costing.
3. Work on the methodology partly developed in this project
should be continued so as to provide a complete methodology
that is applicable for any municipality having any particular
set of conditions and possibilities. The most pressing need
in the methodology is to complete the work on the recycle
problem, namely the method for determining in a given system
what quantity must be demineralized and what quantity
discharged in order to maintain a blend which just fails
to violate any of the quality constraints imposed. This
project studies the two extremes, namely (1) demineralize
and recycle none, discharge all, and (2) recycle all,
discharge none. Actually, it is very likely that the lowest
cost will be achieved somewhere between these two extremes,
particularly if a conventional sewage treatment plant is
already constructed and in operation.
4. In the present study lacking any other information, it
has been assumed that the municipal increment, namely the
concentration increment in the sewage attendant upon one
municipal use, does not have seasonality. Possibly this
is correct, but if it is not correct the recycle algorithm
would have to be quite severely modified. Thus, before a
recycle algorithm could be generally useful, it would be
necessary to demonstrate whether in general the municipal
increment has seasonality. This should be done by sampling
and analysis program in a number of cities carried out over
a complete year.
-------
5. Because of its bearing on the total cost in reuse, a
study should be made in a number of cities to determine
the extent and duration characteristics of the excesses
of sewage flow over water used. In a reuse scheme if
there are any days on which the sewage flow is greater
than the water use the excess water must either be wasted
or stored. If it is to be wasted, treatment and discharge
must be provided? if it is to be stored, storage must be
provided; in either case representing a cost not covered
in the methodology of the present study.
6. A number of other recommendations for further work are
contained in the appropriate sections of Chapter I including
the construction of practical design and costing computer
programs for demineralization, for canal conveyance and for
ground water facilities.
-------
CHAPTER 1
SUMMARY OF PROJECT
This project explored what is required and what happens if
some or all of the sewage collected from a municipality is
treated by advanced waste treatment (AWT) and put back in
the water distribution system. By what happens and what is
required is meant the changes in the quantities of water
withdrawn from the sources, passing through the city and
treated by various means, the changes in the concentrations
of contaminants in the water and the means taken to keep
these within specified limits, the size and costs of the
facilities for accomplishing this recycle system, and the
annual costs of such a scheme as compared with the annual
costs of the conventional once-through scheme of supplying
the same demands. Portions of the results are general and
usable for any municipal situation. Others are specific
for San Antonio in the year 2000 which was used as a
concrete case around which to frame the project.
This chapter is a summary of the remaining five chapters.
Chapters 2, 4, and 5 are computer programs for designing
and costing advanced waste treatment processes for
renovation', pipeline conveyance, and activated sludge
treatment. Chapter 3 discusses the pattern and logistics
of municipal recycle showing that the design and operation
is considerably more complex than hitherto revealed. An
important role is played by the losses that occur between
the water distribution and the sewage delivered. In
addition, these losses have a high seasonality which com-
plicates the design. Chapter 3 also supplies the numbers
for the year 2000 for San Antonio for two extreme cases:
discharge all and reuse none, and: discharge none and reuse
all. Chapter 6 develops the investment and operating costs
on an annual cost basis for these two cases, in 1969 dollars.
The outcome is that complete reuse is almost as cheap as
the conventional once~through system. Indeed, the difference
is only 10% which is so close that any real decision would
have to await a more refined study.
-------
In the year 2000 a design water usage of 341 mgd (million
gallons per day) average is predicted for San Antonio, with
either system. For the conventional system of ground water
plus surface water, treatment of the surface water,
conventional sewage treatment and discharge, the required
investment is 264 m$ (million dollars) and the total annual
production costs 24.4 m$/yr (million dollars per year),
19.5
-------
LOGISTICS OF MUNICIPAL RECYCLE
General Pattern and the Recycle Problem
Discussions on municipal recycle customarily stop at: "Let's
treat the sewage to make drinking water out of it and put it
back into the mains." Chapter 3 explores in some detail what
is required and what happens if this is done and reveals
incidentally why so many discussions come to a stop at that
point.
The generalized pattern of municipal use and reuse is shown
in Figure 1.
The water in a conventional municipal water supply and waste
disposal system is receiving a continual input of contamin-
ants: the contaminants in the source water, the contaminants
introduced through use, and the contaminants introduced in
the course of treatment for discharge. In the typical U.S.
city in addition to the increment of several hundred mgpl of
organics resulting from municipal use, the water also
receives an increment of the order of 300 mgpl of inorganic
ions.
In the conventional once-through system these contaminants
are removed in the sewage treatment, in the discharge, in
the losses from the sewage collection piping and in the
losses representing water that is used but not returned to
the collection system, the last in a major way from the
watering of lawns but also from fire fighting, street
flushing, etc.
In a recycle for reuse scheme some or all of the collected
sewage is diverted from discharge and treated to make it
acceptable for reuse. This closing of the system introduces
two hitherto untackled problems - a quantity problem and a
concentration problem which are themselves individually
complicated and which besides are inter-related with each
other in a complicated way.
As to the quantity problem it is not possible to obtain all
the needed water from recycle because some of the water used
does not find its way to the treatment plant but is lost in
lawn Losses and pipe losses. At least this much must come
from makeup water, ordinarily, of course, from the source
which had been in use in the conventional system. This in
-------
PROCESS ELEMENT SEQUENCE IN MUNICIPAL RECYCLE
Water Sources
oo
Water Treatment
Return
I I
Jr
{Contaminant
Increment
V
Waste Collection
AWT
V
Existing or conventional
Waste treatment
V
Explicit
demineralization
V
Discharge and disposal
Lawn Loss
<-
—Infiltration
Pipe
Loss
Output
V
-------
itself would not be much of a complication except for the
unfortunate fact that while the quantity of sewage collected
is almost constant throughout the year the quantity of water
used has a high seasonality which obtains almost soley
because of the high seasonality of lawn irrigation, high
in the summer and low in the winter. This means that the
system for supplying the makeup water must face a highly
variable load and therefore must operate under the disad-
vantage of a low utilization factor, that is it must have
the capability to meet a high maximum day demand but must
in the year's average spread the costs over a much smaller
quantity. Indeed where the present supply, now in the
recycle scheme to be used for makeup, is close to its
physical constraint limit it means that the new additional
makeup system must be installed simply to supply a peak
demand during a few months of the year.
This makes for a complicated and expensive system but this
complication is nothing compared with the concentration
complication.
One of the objectives that must be satisfied in reuse is to
have the blend of the makeup and the return be suitable for
use; that is, there is some specified constraint on each
contaminant in the blend. If the return water had zero
contaminants and if the makeup concentrations were all lower
than the blend constraints then any quantity of return would
be allowed. However, if some of the makeup concentrations
are higher than the blend constraints then the return must
be considered as "dilution water" for that contaminant and
the blend constraint could only be met if the quantity of
return water were sufficient. But the quantity of return
water is limited. In San Antonio in the summertime the
return water cannot amount to more than about 40% of the
water used. This means that it would not be feasible to
reduce the contaminant level in the makeup to half of its
value. Of course, this is not a situation that would occur
very often. In most cases presumably the municipality
would be willing to accept a slightly lowered quality in
the blend water compared to the makeup which they had
historically been using. However, the circumstance could
arise if it were necessary to seek a supplemental source
if that source were highly mineralized.
-------
But it would be very expensive to produce a return water
with zero contaminants. Instead one would attempt to remove
from the sewage only so much contaminants as is required to
meet the blend constraint. In that case it is usually
rather the makeup water which serves as the dilution water
for the return. If all of the sewage collected is used as
return water then in the summertime when the use of makeup
water is high the return water would not have to be purified
to the degree that it must be in the wintertime when the
makeup water for dilution is low. Indeed, as actually happens
in the San Antonio case it might not be necessary to demin-
eralize at all in some of the summer months. Thus not only
is the makeup water system faced with a low utilization
factor but also the demineralization system is similarly
disadvantaged.
Some demineralization processes may be adjusted so that they
only partially remove some of the contaminants to the extent
desired. Other demineralization processes remove some con-
taminants to a high degree and cannot efficiently be made to
do less. Where the inherent removal in a demineralization
process is greater than needed not all of the return need be
demineralized. If such is true for all of the contaminants
then for one contaminant the quantity that has to be
demineralized has to be greater than for any other and only
this quantity need to be demineralized. The rest can be by-
passed.
Some demineralization processes remove certain contaminants
very poorly. For such contaminants or indeed for any contam-
inants r those for which the input quantities are greater than
the sum of the output quantities in the lawn loss, the pipe
loss, and the demineralization process will build up in the
recycling water and eventually exceed any blend constraint.
Of course, one might shift to a demineralization process such
as distillation and mixed bed ion exchange polishing which
would remove virtually all of the contaminants. However, in
general, presumably a cheaper way to handle this contaminant
build up is to purge some of the sewage by discharge. While
the conventional scheme involves "discharge all and reuse
none," and the complete reuse scheme involves "reuse all and
discharge none," this represents a compromise between the two,
in the interests of ultimately achieving a cost lower than
for either extreme.
10
-------
There may, however, be constraints on the discharge also.
There are already constraints on discharges with respect
to organic contaminants and total suspended solids. In some
places there are constraints on the discharge with respect
to phosphates which in general appear to be handled not by
phosphate removal processes but by prohibiting phosphate
from appearing in the municipal increment.
For the common water minerals, NaCl, Ca(HC03)2, etc., it is
unlikely that any discharge constraint would be violated by
a municipal sewage if the municipal water itself was fit to
drink in the first place. However, if demineralization is
practiced and one is attempting to dispose of the demineral-
ization residue in the discharge then it is quite possible
that discharge constraints could be violated. In this case
there would be still another constraint on the amount that
could be demineralized.
The foregoing discussion has been presented to assist the
reader to visualize the complexity of the problems involved
in municipal recycle and to recognize that the problem is
far more complex than usually associated with the carefree
philosophy "let's treat the sewage to make drinking water
out of it and put it back into the mains." The formal
statement of the logical problem is as follows:
In a system of recycle for reuse having water for
use as a blend of makeup water and recycled water,
having a contaminant increment attendant upon use,
having losses in use not returned to the treatment
plant (lawn losses), having losses in transit of
waste not returned to the treatment plant (pipe
losses), having conventional sewage treatment,
having advanced waste treatment with some attendant
demineralization, having explicit demineralization
and allowing some by-pass thereof and having dis-
charge and disposal,
GIVEN:
makeup water quality in N contaminants, criteria
(maxima) for water quality in use in N contaminants,
municipal concentration increment in N contaminants,
quantity of pipe losses, quantity of lawn losses,
any set of treatment and advanced treatment processes,
any set of explicit demineralization processes,
criteria for water quality of the discharge in N
contaminants, and any set of disposal processes,
11
-------
DETERMINE:
for any given quantity of effluent discharged what
quantity of the recycle must be demineralized (and
what quantity by-passed) in order to maintain in
the blend water and the discharge a steady state
concencration meeting the criteria in N contaminants.
The answers to this problem comprise the quantity to be
demineralized and the quantities to be discharged. For any
given set of conditions, water use quantity, losses, munici-
pal concentration increment, characteristics of treatment and
demineralization processes, etc., there may be or there may
not be a feasible solution. If there is a feasible solution
there.is a set of discharge quantities that will allow a
solution which just fails to violate the criteria. For any
of the allowed discharge quantities there are two deminerali-
zation quantities which will satisfy the criteria. Of these
two demineralization quantities the higher one will give a
total cost for the system which is higher than the cost for
the lower one. (For any given demineralization quantity there
is also a pair of discharge quantities which will satisfy the
criteria but a simple statement cannot be made as to which of
the two will be cheaper.) These pairs of economic discharge-
demineralization quantities provide a set of solutions which
will meet the criteria. One of these pairs will have a cost
lower than all others and this is the optimum solution.
This optimum point may lie anywhere from one extreme to the
other. The one extreme is to discharge all the sewage
collected and reuse none and therefore demineralize none.
The other extreme is to discharge none of the sewage
collected and reuse all in which case of the amount reused
the amount demineralized may be none, some, or all. In
between the extremes lie the various combinations of dis-
charge some, return some, and of that returned demineralize
none, some, or all.
12
-------
Mathematically the possibilities for the solution are
represented as follows:
D = Quantity discharged
n _ Quantity discharged _ . R
Quantity collected
- Quality returned
Quantity collected
_ Quantity demineralized
Quantity retruned
Z = f(R, )
1> R > 0
1> Z> 0
R is the fraction of the quantity collected that is returned,
Z is the fraction of the quantity returned that is demineral-
ized. Z is a function of R and the constraints and other
quantity parameters. R and Z may lie in the range between
zero and one but such that the pair satisfy the functional
relation.
The problem is akin to a linear programming problem but
departs from linearity in a number of ways such that linear
programming cannot be used for solving it. The 'project did
not have sufficient time to work out the computer algorithm
for solving the problem and the completion of the work on the
problem remains as a recommendation for further investigation.
Instead, the project investigated the two extremes which are:
1. Conventional supply primarily by import and
conventional waste treatment and discharge
with no reuse and no demineralization, and
2. Complete reuse with required demineralization
and with no discharge.
Chapter 3 develops the quantity pattern for these two
extremes. Chapter 6 summarized beyond developes the costs
thereof.
13
-------
Numbers for San Antonio, Year 2000
Figure 2 shows a long term trend in water use in Bexar County
as represented by withdrawals from the Edwards. There appears
to be an upward trend for the first 20 years but after the
drought in 1957 the per capita consumption falls back to the
1940 level. It is projected that the usage in the target year
2000 A.D. (anno Domini) will not differ greatly from the usage
in the 60-70 decade.
Figure 3 shows the long term trend for sewage collected and
the difference between water used and sewage collected, being
the total loss. Over the past 15 years it appears that the
gpcd (gallon per capita per day) for sewage collected has been
increasing and the loss has been decreasing. Indeed, one could
draw trend lines showing that in four or five years the sewage
collected would amount to more than the water used. Such a
trend, although not impossible and resulting from heavy infil-
tration or gradual inclusion of storm flows in the sewer system,
is thought to be spurious and as will be seen the bases.for the
2000 design are the figures from the 61-65 seasonal study to
be described.
The decrease in the loss does not come from any change in
rainfall. The annual rainfall over the period of interest
has not been increasing particularly and the seasonal study
to be described showed that the monthly loss ratio has only
a second order correlation with the monthly rainfall, the
main predictive parameter being the monthly temperature.
For the design basis the maximum day for water use and the
maximum day for sewage collected were taken in ratio to the
maximum month, both expressed as mgd (million gallons per
day). The 90 percentile values for these ratios are 1.30
and 1.85, respectively. This means that there is only a 10%
chance that the ratio in any single year might be greater
than these values and these were used in the system design.
A study was made of the monthly per capita water withdrawal
and sewage delivered in San Antonio for 1961-1965 which some
of the results are presented in Figure 4. It is seen" that
the sewage flow has little seasonal variation but the water
use has an extreme peak in July and August. As the basis
for the 2000 design the monthly average per capita water
withdrawal was taken as approximately the upper envelope
of the water curve in Figure 4, and the makeup was similarly
taken from a curve showing the excess of per capita water
withdrawal over sewage delivered.
14
-------
BEXAR COUNTY ANNUAL WATER USE
ex IRRIGATION
Ul
300
250
200
150
100
50
0
40
Municipal, Industrial
Country Clubs, Schools
Domestic, Stock, Estates,
ex Schertz
gpcd, ex irrigation
mgd, ex irrigation
50
YEAR
60
68
Figure 2
-------
LONG TERM WATER AND SEWAGE RELATIONS
SAN ANTONIO
Period of the
Seasonal Study
LOSS, (W-S)
Sewage probably too low and
loss too high in this range
because population served
probably too high
1940
YEAR
-------
MONTHLY AVERAGE PER CAPITA WATER WITHDRAWAL AND SEWAGE
DELIVERED - SAN ANTONIO 1961-65
360
300
MONTHLY AVERAGES
as gpcd
200
120
100
J-
64
63
62
61
5 7
MONTH NUMBERS
I '
Figure 4
L2
17
-------
The existing facilities in San Antonio which will be part
of the 2000 system are the three existing or under construc-
tion sewage treatment plants and the existing well stations.
The latter have a firming factor, that is an average ratio
of installed capability to firm capability, of 1.4.
The municipal increment is the difference in concentration
between the water used and the sewage delivered. Figures
for the apparent increment in San Antonio were available for
five of the major ions and 13 of the minor elements. The
rest of the major components in the municipal increment were
filled in by the averages of figures in other western cities.
Figure 5 taken from the data just described shows the monthly
water patterns in 2000 in San Antonio. It should be remem-
bered that the design data are approximately the upper
envelopes of a five year monthly series. On this basis the
top curve labeled 6 shows the monthly water usage peaking at
515 mgd in August and averaging 341 mgd. The next lower
curve, labeled 7, shows the ground water withdrawal under
the conventional supply scheme peaking at 339 mgd which is
the allowable limit on monthly withdrawal for the aquifer,
and averaging 192 mgd which is the allowable limit for
annual ground water withdrawal from the aquifer. To make
up the difference, surface water would be required as shown
in the curve labeled 8 peaking at 176 mgd in August and
averaging 149 mgd. Under the reuse scheme the ground water
withdrawal would be as shown in curve 3 peaking at 336 mgd
and averaging 164 mgd which may be compared with the 163 mgd
now being withdrawn in 1970. No surface water would be
required.
Projection of the population to years beyond 2000 indicates
that in the reuse scheme the peak month constraint is just
barely met since it would be violated about the year 2005.
The annual constraint would be exceeded about the year 2017.
Starting in 2005 there would have to be provided some surface
storage for ground water pumped in the winter and spring
months and stored to avoid exceeding the peak allowable in
July and August. The storage period and quantity would have
to become larger and larger as the population grew. Begin-
ning in 2017 no amount of storage would suffice and it would
be necessary to supplement the ground water supply. However,
the target year in this project is 2000 and in this design no
constraints are violated in that target year.
18
-------
500
400
350
300
250
t
mgd
(?)
©
®
(D
X
,s
'
•
•.
\
'.
•.
-^f^
surfai
MONTH!
WATER PATTERNS
2000 A.D.
San Antonio
/
/
/
/
/
(^
Basin
•
'
Ground Water,
conventional
Average 192-s^
\
\^s
~ ^
e
B»*
^y.-'
^ ,"
.-.
•' \
'
/
y
^^
/
Monti-
y
j
i
t:
t:
1 :
II
:
/
/
/
/
.
/
t
t
f
f
f
/ \
^X
^^
)Ground water, reu
Average 164
'-.
\
\
\
\
\
\
A
\\
\\
u
•\
\\
\\
•\
\\
\\
:\
\\
•\
\\
• \
;
•
-• — _
^^>.
se^*
De mineralization
\
/
^
\
Number
i *
123456
^ /
>^:
/
/
Cc
>nvent
tonal
vs.
Reuse
V
\
\
\
\
\
'. \
•- \
1 1. \
^ \
\
1
t
_.-
3tal ii
itake
©
3/i 1 AiFpracp
l\
\
\
\
•
;
•
N\v~
;' ^-X
7
*,
\
\
'
Allow
month
able /
ly pea
/
&
ground water
f
"
\/
/\
•
Allowable
annual
B
round water
®
— Surface water
conventional
149 Average
X
>s
'.
*.
®
©
Figure 5
339
19?
'£ 9 10 11 12
19
200
150
100
-------
Of course, these projections depend upon the constancy of the
gpcd water use and sewage delivered. Under a reuse scheme
any steps taken to reduce the gpcd water intake and increase
the gpcd collected would be favorable toward postponing these
critical dates.
Figure 6 shows the seasonal changes in the blend composition
in the year 2000 for the recycle scheme with no demineral-
ization and no discharge. It is seen that with no discharge
and even with no demineralization the blend would have a TDI
(total dissolved ions) of less than 500 mgpl in the two
summer months but would rise to higher than 1000 mgpl in
the winter. In Chapter 6 summarized beyond there is worked
out roughly the quantities of demineralization (by electro-
dialysis) that would be required to lower the TDI to 500 in
every month. This is shown as the dotted lowermost curve on
Figure 5.
Table 1 shows the quantities involved in the 2000 system
giving the average, the peak monthly, and the peak day for
each element of the system. The triple figures in the
conventional column show the requirement if the peak day
load is thrown entirely to the ground water, entirely to
the surface water, or spread equally between them. The
figures for demineralization are obtained as a rough
approximation by determining the amount of demineralization
that would be required to reduce the return flow from its
concentration with no demineralization to a concentration
which would result in a blend concentration of 500 mgpl
for TDI. This somewhat overstates the quantities to be
demineralized for if the blend were reduced to 500 TDI
then the return liquor would not reach such a high concen-
tration in the recycling. To determine the quantities to
be demineralized under true steady state conditions not
violating the blend constraints would require the solution
to the recycle problem which has not been accomplished under
this project.
20
-------
1000
500
XTDI
100
.1000
.500
100
10 12
-------
TABLE 1
LOGISTICS OF THE NEW SUPPLY - 2000
Note: Peak day is the 90% level —i.e. expected to be
exceeded in only 10% of the years.
Ground water withdrawal, average
Peak monthly
Peak day
Utilization factor
AWT Reuse
Scheme
mgd
164
336
435
.377
Conventional
Import Scheme
mgd
192
339
339/440/494
.S67/.436/.389
Surface water withdrawal, average
Peak monthly none
Peak day withdrawn
Utilization factor
Lawn and pipe losses
164
149
176
331/230/176
.429/.649/.S45
Total withdrawal
Peak monthly
Peak day
Over- all utliziation factor
Water treatment , average
Peak monthly
Peak day
Utilization factor
164
336
435
.377
not
used
341
515
670
.510
149
176
331/230/176
.429/.649/.S4
5
164
Sewage treatment or AWT, average
Peak monthly
Peak day
Utilization factor
(AWT) (Conventional STP)*
177 177
189 189
350 350
.506 .506
Demineralization, average
Peak monthly
Peak day
Utilization factor
(rough)
117
261
486
.241
not
used
Disposal to the Gulf, average (rough) 7
Peak monthly
Peak day 14
Utilization factor .5
not
used
Discharge to San Antonio River none discharged
177
Storage required
yes
no
Sewage treatment plant
23
-------
When this is done the amount (of demineralization residue)
to be disposed of to the Gulf will also be decreased.
The term "storage required" refers to the need for a
balancing storage to store the return water on those days
on which the amount of sewage collected is more than the
amount of water used. This cannot yet be computed because
it requires a day-by-day study of the water and sewage flows
to determine the statistics and pattern of the excesses.
Table 2 shows the sewage treatment plant alternatives,
conventional sewage treatment for the conventional scheme
and advanced waste treatment for the reuse scheme.
TABLE 2
LOGISTICS OF THE NEW SUPPLY - 2000
Sewage Treatment Plant Alternatives
mgd
AWT Conventional
Reuse Import
Scheme Scheme
Existing or U.C. STP (3 plants)
Peak day capability not 116
Average used 59
New capability required (AWT) (STP)
Peak day 350 234
Peak monthly 189
Average 177 118
Utilization factor .506 .506
Discharged to San Antonio River none 177
Table 3 shows the required ground water facilities, which do
not differ much between the two schemes. Table 4 shows the
conveyance alternatives. The AWT reuse scheme in the single
AWT plant embodiment would require the conveyance back from
the AWT plant illustratively at the Rilling site to the water
distribution system illustratively taken as the Hildebrand
tank in the north part of the city. The conventional scheme
would require the conveyance from one of the several alter-
native supplemental sources here taken as the Cuero-Cibolo
source. This source would require the reimbursement of the
Guadalupe Basin with water from Goliad Reservoir.
24
-------
TABLE 3
LOGISTICS OF THE NEW SUPPLY - 2000
Ground Water Facility Alternatives
mgd
AWT Conventional
Reuse Import
Scheme Scheme
Existing GW facilities
Peak day firm capability 333 333
Average 125 145
New facility required
Peak day, firm 102 107
Average 39 47
Utilization factor .377 .436
TABLE 4
LOGISTICS OF THE NEW SUPPLY - 2000
Conveyance Alternatives
mgd
AWT Conventional
Reuse Import
Scheme Scheme
Cureo and Cibolo to Hildebrand
Peak day 0 230
Average 0 149
Goliad to Victoria
Peak day 0 141.4
Average 0 113.1
Rilling to Hildebrand
Peak day 350 0
Average 177 0
25
-------
COSTS OF THE TWO EXTREME ALTERNATIVES
Chapter 6 develops approximate costs for those cost
elements other than AWT, pipeline conveyance and activated
sludge treatment for which computer programs are developed
in chapters 2, 4, and 5. Some of these Chapter 6 costs
for the other elements are generally applicable though of
an approximate nature without the flexibility of the full
computer programs. Others of the Chapter 6 costs are
simple ad-hoc costs developed specifically for San Antonio.
The assignment to these categories is as listed:
General Ad-hoc
Surface water reservoirs X
Ground water facilities X
Water treatment (X)
Demineralization
{by electrodialysis) X
Sewage conveyance X
Table 5 gives a summary of the element and system costs for
the two extreme alternatives. About three-quarters of the
conventional costs are for the source water and its treatment
The differential over the comparable cost of source water for
the AWT reuse case is about 17 m$/yr. However, the only
additional cost in the conventional system is for sewage
treatment and the cost of advanced waste treatment and asso-
ciated elements slightly overbalances the differential.
26
-------
TABLE 5
COST SUMMARY
Conventional
AWT Reuse
Design
Capability
mgd
Surface water delivered 230
Water treatment 230
Total SW delivered and treated
Ground water 440
Total water delivered and treated
Sewage treatment 350
Advanced waste treatment
Return conveyance
Balancing storage
Demineralization (very rough)
Disposal to Gulf (very rough)
Average New Annual
Production Capital Cost
-------
COMPUTER PROGRAM FOR RENOVATION COSTS
Chapter 2 covers the development of a computer program for
the design and costing of the lime-clinoptilolite-carbon
process for renovating municipal waste waters for reuse.
The over-all process points toward an advanced waste treat-
ment system whereby water meeting potable standards is
produced from waste waters.
The major steps are:
1. Lime clarification with recalcination of the
lime...reducing the suspended solids, the
organic content both suspended and dissolved,
and the phosphate content of the waste.
2. Clinoptilolite ion exchange regeneration and
stripping ...removing the ammonia nitrogen
from the waste and concentrating it in a small
volume of ion exchange regenerant from which
it is air stripped.
3. Activated carbon absorption...removing the
organics residual from lime clarification.
These steps are followed by a final chlorination to assure
disinfection of the water.
In this process all the wash liquors generated in the later
stages are returned to the lime clarification step so that
no liquids leave the system other than in the product water
and the only solid for disposal is the moist inert residue
from the lime slaker. The organics are burned off in the
lime recalcination kiln and the ammonia leaves as a gas
from the stripping tower.
Selecting this process for the basis of the present study was
a calculated risk. The process had been piloted on a moderate
scale but as applied to secondary sewage treatment plant
effluent, not to raw sewage. Applied to raw sewage the pro-
cess was under study during the design period, at the Blue
Plains, District of Columbia, Experimental Plant operated by
EPA. It appeared, and still appears so, that this process,
the so-called independent physical-chemical process, was a
likely contender for water renovation as against the conven-
tional biological processes with supplementary (tertiary)
treatment, in part because it was more readily controllable
for varying production rates and not subject to the erratic
fluctuations that plague the biological processes. Further-
more, renovation by conventional biological treatment followed
28
-------
by tertiary treatment would involve exactly the same process
steps, -lime, clinoptilolite or other forms of ammonia removal,
and carbon, and the cost of these tertiary steps would be
only very slightly less than the cost of the same steps
applied directly to raw sewage. In other words, the cost of
renovating secondary effluent would have been only slightly
less than the cost of renovating raw sewage and in addition
there would have to be borne the basic cost of the conven-
tional biological process itself. It was decided therefore
to use the independent physical-chemical process even though (1)
it had not been fully proven, and (2) the basic performance
data were not yet complete.
Under these circumstances it is readily conceded and should
be emphasized that the design and computer program is highly
preliminary and is to be considered not an end in itself but
only one of the steps toward answering the question: "With
what we know about it now, how does reuse compare economi-
cally with a conventional system?"
Not enough i;s known about the equilibrium, stoichiometric,
and rate processes involved to construct a true mathematical
model, i.e., a model in which any parameter change will be
reflected in a corresponding performance change. For example,
in the activated carbon stage the COD (chemical oxygen demand)
of the effluent is set by the value assigned to the parameter
COD89, in the exemplary work set at 8 mg/1 (milligrams per
liter). The COD actually reached in the effluent for a fixed
feed composition will in a real plant vary with the equipment
configuration, the contact time, and the carbon loading
factor. Lacking knowledge of the relationship among these
the program merely provides for setting the contact time and
carbon loading factor...or in other words it allows the user
to make his own selection of carbon loading factor and
contact time which he thinks will produce the desired COD
in the effluent.
In using the program the user may select any combination of
feed water quality parameters and any reasonable combinations
of the other decision parameters. Using these, the program
will design and cost' the equipment so that the plant
capability is QDOT, the design capability mgd, and will cost
the operation when it is operated in that equipment at QBARE,
the expected average production rate, mgd.
29
-------
Such a program, of course, can be used as a research and
development tool to direct research toward cost improvement.
Very little of this was done with the program, however,
since the present purpose was simply the costing of the
process under the current technological level. In the cost-
ing the various performance parameters were set at levels
expected from pilot plant work or from actual operation of
similar equipment.
Economic factors were set at plant life, 20 years; tax,
insurance and interest rates .01, .01, and .045 annual
fraction of investment, respectively; energy price IC/Kwh
(kilowatt hour); base labor price $3/man hr; and payroll
extras factor .45 fraction of payroll. The feed liquor
used is a typical San Antonio sewage composition with a COD
of 500, suspended solids of 220 mg/1, and inorganic ions
approximately as found in the sewage.
The next page comprises the printout of results of the AWT
renovation process for San Antonio in the year 2000 con-
sidered as a single central plant. The costs are 1969,
San Antonio. The QDOT capability is 350 mgd, meaning that
the plant has a capability to treat as much as 350 mg in a
single day. The average production is 177 mgd, for a
utilization factor of .506. The entries may be explained
by example as follows. For the lime stage of the process
the investment is 20,454K$ which is 23.77% of the total
investment. The unit investment in this lime stage facility
is 5.844C/gpd of capability or 11.556<:/gpd of actual pro-
duction. The production cost for the lime stage, that is
the operating costs plus the amortization is 3,479K$/yr
which is 5.386C/Kgal produced and this represents 22.67%
of the total production cost. The total investment for the
entire renovation plant is some 86 million dollars and the
annual production cost 15.3 million dollars. Over-all amorti-
zation costs are 12.9
-------
AkT PROCESS ----
frBARE 177.00 9D01 350. 00 UHA1": . S0!>71 LI
INVESTMENT COSTS, K?
PROCESS
PRELIiX.
LIKE
CLINOP.
CAhfcOiX'
CBLOFr.
BUILDINGS
DISPOSAL
ENGH.
TOTAL
PftODUCTIOfc
PBOCBiSS
PBELItt.
LIKE
CLINOP.
CARBON
CHLOh.
BUILDINGS
DISPOSAL
ENGR.
OPE. LABOR
KS
6^5.477
22410 .410
19517.989
37438.609
690.852
8074.334
0.
3516.747
' 92334.416
COSTS
KS/YEA
123.624
3759.420
4088.432 '
6116.832
342.300
782.210
78.373
340.689
551 .903
PCT. OK
TOTAL,
• If'
£4. £7
P 1 . 1 A
40.55
.75
8.74
o.
3.K1
100.00
rt CK^T5/>
'.191
5.819
6.328
9 . 4 6 1?
.530
1.211
.121
.527
.854
CP-ivTS/i
ODO'l
. 196
6.403
5.577
10.697
. 197
2.307
o.
1 .005
26.381
?GAL PCT.
.76
23.23
25.26
37 . u 0
2.12
4.83
• 4fi
2.1 1
3.41
GPI- OF
vi;A7
-------
COST OF AWT BY THE LIME-CLINOPTILOLITE-CARBON PROCESS
San Antonio Sewage 1968 Utilization Factor, UBAR = 0. 5
FCODAC =0.8
100
UNIT INVESTMENT
<£/gpd of capability
gpd vs capacility
00
I -
or
- Unit Production Cost
<£/Kgal vs product!
base case
10
.1
10 100 1000
DESIGN CAPABILITY OR AVERAGE PRODUCTION, mgd
Figure 7
-------
COMPUTER PROGRAM FOR PIPELINE CONVEYANCE COST
Chapter 4 covers the development of a computer program for
the design and costing of pipelines and the conveyance of
water or other fluids through them. Water conveyance is one
of the important elements bearing upon the economic competi-
tion between conventional water supply and waste treatment
typically by importing water from remote sites, and renovation
for reuse by advanced waste treatment.
A computer program takes the specified characteristics of the
conveyance situation, designs a pipeline which will minimize
the cost of conveyance in that situation, and returns the
design data and the cost breakdown. The line is designed in
segments (up to three) as may be specified, each section being
optimized for conveyance of an average amount of QBARE in a
facility which has the firm capability of conveying an amount
QMAX. The program generates the necessary cost indexes for
the state, region (21 in the nation) and future year, as well
as energy price corresponding to the state, the future year
and the expected Kwh/yr energy consumption, A subroutine
generates the proper friction factor from the Moody diagram.
The program determines whether the line shall be pumped,
boosted, or gravity, and for the pumped and boosted selects
the proper psi pressure classes of pipe.
To the cost of conveyance in a horizontal line the fixed
charges on the pipeline itself contributes some 70-80% of
the total. The pipeline price relation therefore is very
important in costing. In this study the basis is a historical
correlation of the installed costs of 825 pipelines which were
correlated against diameter and by region of the country.
There are large regional differences in the prices of pipeline
which must be taken into account in estimating the cost of
conveyance. Special regionalization factors are provided for
doing this.
Figure 8 shows the cost of conveyance by pipeline at a utili-
zation factor, UBARE, of 0.5. (The firm capability is two
times the average conveyance rate.)
The study showed that the capital charge on pump stations
contributes 5-10% of the cost in horizontal lines and the
energy about 10%. For pumped lines the optimum diameter is
practically independent of the slope or static head. The
conveyance cost is directly proportional to distance down to
distances of one to three miles at least.
Using the future year cost index feature which is incorporated
in these computer' programs, it was shown that both the invest-
ment and the conveyance cost will approximately double, in
current year dollars, between 1970 and 2000.
33
-------
Average Conveyance,
mgd.
Conveyance
Cost, .
t/Kgal mi \
(1968, National) \
(Field boundaries) '/\
approximate) i / I
III
COST OF CONVEYING
WATER BY PIPELINE
UBARE = 0.5
Effect of QBARE and SLOPE
.01
-50 -40 -30 -20
-10 0 10 20
Line Slope, ft/mi
30 40 50
Figure 8
34
-------
The cost of conveyance in the Cuero-Cibolo-San Antonio link
of the Texas Water Plan had been the subject of an engineer-
ing study by a consulting engineering firm.. That study laid
out an actual route and designed a pipeline and pump station
system under a set of ground rules laid down by the client.
When the present computer program was run with the same basic
data the corresponding 1969 costs of the two studies were
within 7% of each other over the range of the engineering
study which was 100, 200 and 300 thousand acre feet per year.
This illustrates how the computer program without using any
input data other than the quantities, elevations and distances
can reproduce costs from rather detailed engineering studies
which include field studies, map routing, topographic and
geological profiles and item-by-item preliminary cost esti-
mating. (For the exemplary quantity above most closely
corresponding to the actual San Antonio case costed, the
difference was only two parts in 500.)
COMPUTER PROGRAM FOR ACTIVATED SLUDGE TREATMENT COSTS
Chapter 5 takes an existing computer program for design and
costing of activated sludge plants, modifies it to handle
some of the actual design problems and uses it to develop
capital'and operating costs5for the conventional sewage
treatment components in thesSan Antonio scheme. For general
application it would be necessary to have programs for other
treatment types also. For example, trickling filters are
more economic than activated sludge plants under some condi-
tions, particularly at small sizes. However, activated
sludge was chosen as a conventional treatment type for this
project because (a) the established computer program was
available, (b) activated sludge was a preferred type of
treatment for larger capability plants as applicable in the
San Antonio study, and (c) activated sludge is a treatment
type of the existing San Antonio plants which are involved
in the basin management.
The program permits any feed, any product, and any size
(within limits). This means that the user may specify the
raw sewage composition and the desired effluent BOD (bio-
chemical oxygen demand). The program will then design a
plant to achieve this of a size capable of handling the
maximum day and will cost the treatment of the QBARE average
day in the plant designed. It is not permitted to specify
the total suspended solids desired in the effluent but this
usually comes put of a magnitude quite close to the effluent
BOD. The program will correctly handle a specified effluent
BOD of ten mg/1 but will not handle one as low as five, the
actual limit being some undetermined value between these two,
35
-------
COMPONENTS OF THE PROBLEM REMAINING UNRESOLVED
The objective of this project was to develop general
procedures by which a comparison could be made and an
optimization achieved between the total costs of a future
water supply-pollution control system which is at or some-
where between the two extremes:
1. Conventional supply primarily by import and
conventional waste treatment and discharge,
and
2. Complete reuse with no discharge
At the inception the problem was admitted to be complicated
and possibly massive but it was thought that those problems
which could be foreseen could be solved by the application
of known techniques. For example, it was thought that the
"placement problem," the problem of how many AWT plants of
what sizes and where as contrasted with a single AWT plant
treating all the waste, could be treated as a network
problem once the basic cost relations had been worked out.
However, as the basic data and relationships revealed them-
selves it was discovered that municipal recycle contained
some problems that had not been suspected and attacked before.
These included the central role of the loss ratio in the
compositional changes on recycle, the seasonality of the loss
ratio and the day-to-day imbalance of the water-sewage relation
and finally the inter-relationship of the independent para-
meters, quantity demineralized and quantity discharged, as
they affected the composition of the blend, the "recycle
problem " . Because of the magnitude and the difficulty of
these unsuspected problems, there still remain a number of
components of the over-all problem which are unresolved. The
types of these fall into several categories:
1. Problems for which techniques are known but
which are massive and set aside in favor of
problems heretofore unattacked. Examples:
the placement problem, demineralization
design and costing.
2. Problems requiring the collection of additional
basic data primarily in other cities, for the
present well enough known for San Antonio.
Examples: seasonality of municipal increment,
seasonality of loss ratios.
36
-------
3. Problems representing special details for the
San Antonio case or for which approximate
solutions are allowable, or which probably
will not be generally encountered. Examples:
the alternate (Colorado-Applewhite) source
for San Antonio, canal conveyance., design and
cost for ground water facilities.
4. The recycle problem, the algorithm which deter-
mines the optimum quantities for demineralization
and for discharge to meet the blend constraints,
a problem which our studies to date show is
solvable but which we have not yet completed.
Table 6 lists these components remaining unresolved and shows
whether or not they are needed for a decision in the San
Antonio case and in the general case. Under the general case
the term "probably not" signifies that it probably is not
necessary to work upon these problems for the general case
because: canal conveyance will probably be rarely met with
in actual practice and the costs of ground water facilities
can be approximated for individual cities on the basis of
historical facilities cost in that location without a large
parametric design program.
Under the San Antonio case the term "probably not" means some-
thing else. It takes into account what we already know about
the economics of the San Antonio situation, namely, that reuse
appears to be slightly uneconomic compared with the conven-
tional import scheme, and therefore any additional elaboration
or additional costs taken into consideration for the reuse
scheme would not change the decision since in all such cases
they could only increase the reuse cost. Only one unresolved
problem stands a chance of reducing the reuse cost for San
Antonio, this being the placement problem. The major portion
of this problem is being worked out on another project using
the STORET sewer design scheme. Taken one by one: the
Colorado-Applewhite source including the canal conveyance
from the Colorado source will probably result in a lower cost
for the conventional supply than here computed and this will
not change the decision. The recycle algorithm and the
storage problem could only increase the cost of the reuse
scheme and thus would not alter the decision. The demineral-
ization costs have been briefly reviewed by Ionics Inc., the
vendor, with the comment "your figures on ED costs seem
remarkably up-to-date." The costs for the actual San Antonio
water would probably differ little from those shown.
37
-------
TABLE 6
COMPONENTS REMAINING UNRESOLVED
Needed for Decision in
Component of the problem remaining San Antonio General
unresolved Case Case
Colorado-Applewhite source, replacing probably
Cuero-Cibolo not
no
The "Placement Problem"; area AWT
instead of single AWT at Rilling
site
The "Recycle Problem"; completion of
recycle algorithm for Q demineral-
ized and Q discharged
yes
probably
not
yes
yes
The "Storage Problem"; day-by-day
differences water-sewage
probably
not
yes
Demineralization design and cost
program
probably
not
yes
Canal conveyance design and cost
program
probably
not
probably
not
Seasonality of municipal increment
and completion of basic data probably
San Antonio and other cities not
yes
Loss ratio relations other cities-
seasonality
Design and cost for ground water
facilities
no
no
yes
probably
not
38
-------
To assess as "probably not" the need for studies of the
seasonality of the municipal increment and the magnitude
of the increment for those contaminants not specifically
known for San Antonio is indeed a presumption. Even if
demineralization would not be required because of some of
the contaminants whose increment quantities are not known
for San Antonio, still this would not change the fact that
demineralization is required for some of those that are
known. But as for seasonality it might indeed be so that
the municipal increment does have a seasonal variation and
this might be in the direction of being low in the winter-
time and high in the summer and providentially might allow
the scheme to sneak by without demineralization. But this just
would be more good luck than one could hope for.
39
-------
REMAINING COMPONENTS OF THE GENERAL PROBLEM
The project took as many components of the San Antonio
problem as could be handled within the time and funds
available and worked upon them. It attempted to provide
general solutions for as many of these as time and funds
allowed. Those remaining unresolved from such an attempt
have been described in the previous section.
There remain a number of facets of the problem outlined in
the original concept and the various proposals which were
not worked upon at all.
One of the use alternatives specifically possible in San
Antonio and presumably usable generally where the situation
allows, was the exchange of conventional effluent for source
water used for irrigation. The San Antonio situation is
that some 20 mgd average are withdrawn from the underground
aquifer and used for irrigation about 15-20 miles west from
the present sewage treatment plant, some of the area actually
being in another river basin. If the irrigators would use
the treatment plant effluent and cease the withdrawal from
the aquifer then this additional quantity would be available
for withdrawal from the aquifer for higher order uses, i.e.
for potable water for the city. This would have the effect
in the present scheme of increasing the limits on ground
water withdrawal thus allowing the withdrawal of more ground
water and the importation of less surface water, for a net
economic benefit. On the cost side of the balance would be
the cost of conveying the effluent to the irrigators. There
would, of course, be the administrative and political problems
involved in such a transfer but these are outside the scope
of this project.
The present study provides the means for exploring the
economics of such a transfer. The pipeline conveyance
program would show the costs of the conveyance and the
ground water and surface water costs could be adjusted
for the new quantities. However, this was not done on the
present project. In addition, some of the local agencies
had funded an engineering study of the same scheme and
although the final report thereon was not available the
conclusion had been that such a transfer was uneconomic.
It would be interesting to check this conclusion with the
mechanism provided by the present study.
40
-------
The situation does not fit either of the two extremes costed
herein since under the reuse scheme it would involve operating
one of the conventional treatment plants to produce the
irrigation water. Under the conventional plan some of the
plant effluent would merely be conveyed to the irrigators
thus reducing the amount of discharge.
EPA was particularly interested in the economics of segre-
gating different qualities of water for different uses.
Three broad categories of uses are municipal, irrigation,
and industrial. A specific use for conventional sewage
treatment plant effluent as irrigation water has just been
mentioned. The commodity cost of the water in this case
would be zero since it is merely to be discharged anyway.
However, if one considered an irrigation use in the reuse
scheme the commodity cost would be greater since every gal-
lon of water diverted from reuse to irrigation would require
the generation of a gallon of water from the source. In the
San Antonio case which is impinging on the peak ground water
constraint this would mean additional surface water at a cost
of the order of 30C/Kgal. Balanced against that would be the
possibility of by-passing some of the AWT steps, specifically
the clinoptilolite-ammonia removal and the activated carbon
treatment at a saving of something of the order of 18C/Kgal.
The water would have about 20 mg/1 of NH3-N (ammonia nitrogen)
and a COD of 82 which presumably would be acceptable for
irrigation. This would also mean that the amount of makeup
water would be increased thus the requirement for demineral-
ization would be decreased for an additional saving. All this
assumes, of course, that the use points for the irrigation
water are relatively few and concentrated, else the cost of
conveyance to the use sites would be excessive.
The reuse of lower grade water for industrial purposes was
not explored in the project because San Antonio, with one
exception, does not have any large industrial water users
such as a paper plant, a petroleum refinery, etc. If such
water users occur in a community there are a number of
possibilities for reuse, or, indeed, for successive use of
waste water. Probably the ammonia removal and carbon treat-
ment in advanced waste treatment could be dispensed with for
industrial cooling water with some process adjustments.
Although clinoptilolite-ammonia exchange is used for ammonia
removal another method of ammonia removal is simple air
stripping of the entire liquor in equipment identical with
cooling towers. Thus, passage through an industrial cooling
tower in several recycles therein might indeed accomplish the
ammonia removal which now costs something of the order of
7C/Kgal in the AWT process. The blowdown liquor could then
be returned to the AWT process to the lime stage or to the
activated carbon stage as the economics dictated. Involved,
41
-------
of course, would be the replacement of the water lost by
evaporation in the cooling process. For a study of the
economics of this the project provides the conveyance costs
but the AWT program would have to be modified to include the
extraneous cooling use. Also, the project, of course, does
not provide any information on the additional cost of using
such water in the industry's cooling tower itself.
There are also possibilities for using various grades of
water for industrial process use, but such uses and the
requirements for them are highly specific to the particular
industry. San Antonio has no such industries and it was
outside the philosophy of the project to generate schemes
for hypothetical situations which did not have a real
embodiment on which they could be hung.
The one industrial use embodied in San Antonio is the
cooling water requirement for the power plants of the City
Public Service Board. The project was not able to determine
to what extent these requirements were included in the use
projections of the Texas Water Plan and, indeed, this was
one of the developments that led to the independent explor-
ation of the San Antonio demand situation. In any event the
major interest lies in the Braunig and Calaveras power plants
which will supply the bulk of the future demand and which
operate on their own cooling lakes with water drawn from the
San Antonio River downstream of the City's discharge point.
The 1985 usage of these two plants will be approximately
45 mgd average, this being water consumption. This figure
will never be exceeded because any plants for supplying San
Antonio beyond the 1985 date will not be constructed within
the basin. The 1985 water usage in San Antonio will be about
200 mgd average so that typically the cooling lake require-
ments are about one-fourth of the water use of the munici-
pality served. In the year 2000 San Antonio design the sewage
collected averaged about one-half the water use so that the
water use power plant cooling consumptive use would be about
one-half the sewage collected. If in general it were neces-
sary to supply the power plant cooling use solely from the
sewage collected, this would seriously cut down the extent to
which other reuse could be practiced. Indeed, reuse in such
a case could only supply one-quarter of the total water needs.
The problem was not faced in the present project because the
expected 45 mgd are withdrawn from the flow of the San Antonio
River downstream from the City discharge point. The average
flow in the river is about 180 mgd of which about 90 mgd are
return flow from the collected and treated sewage. Thus, the
average flow of the river in excess of the sewage discharged
is about 90 mgd. The power plant cooling requirement is only
45 mgd and the cooling lake comprises damping storage for the
peaking. Because of this it was judged that the power plant
42
-------
cooling requirement in San Antonio is not part of the reuse
picture. This dictated the development of the project's
own demand figures which made it of no consequence whether
or not the Texas Water Plan figures included or did not
include the power plant cooling water.
San Antonio discharges excess sludge and also bypasses some
sewage flows into an 800 acre lake, Mitchell Lake, which
serves as a holding basin and oxidation pond. The original
concept included the study of the use of Mitchell Lake as an
oxidation pond, or the separation of Mitchell Lake from the
sewage system so that it might be rehabilitated for recre-
ation, storage or other uses. The Mitchell Lake problem also
received a low priority and was not worked upon. The lake
and the plant already for many years have supplied irrigation
water and this irrigation water would have to be replaced if
the status quo were to be maintained with a reuse scheme.
The project as it stands then condemns these irrigation uses
to oblivion. The alternative would be, in the reuse scheme,
to operate the sewage treatment plants to produce the
approximately 12 mgd of effluent now used for irrigation
south of San Antonio, at an additional (increment) cost of
about GC/Kgal.
The original suggestion for this project contained the phrase
"collect every drop of water available to the basin in wastes
and storm drainage." This project does not do that. It
specifically takes nothing from the natural flow of the San
Antonio River in the local area except to withhold the
contribution of the present sewage treatment plant effluent.
To have done so would have involved the placement and con-
struction of a storage reservoir somewhere immediately south
of San Antonio. This would have been a major alteration of
the Texas Water Plan for this area although obviously this
would not have been any consideration for studies in other
areas. If the scheme had used the entire flow of the San
Antonio River, say at Elmendorf, it would have dried up the
river for the downstream users including the power plants.
Also, the total flow of the river is actually not sufficient
to supply the water needs so that reliance on some other
reservoir and conveyance from that point would be required.
The present embodiment leaves enough water in the San Antonio
River to supply the downstream users including the power
plant lakes and the irrigators.
43
-------
CHAPTER 2
COMPUTER PROGRAM FOR DESIGN AND COSTING
OF THE LIME-CLINOPTILOLITE-CARBON PROCESS
FOR WASTE WATER RENOVATION
This chapter covers the development of a computer program for
the key element in the project, namely, for the design and
costing of the lime-clinoptilolite-carbon process for reno-
vating municipal waste waters for reuse. The over-all process
points toward an advanced waste treatment system whereby water
meeting potable standards is produced from waste waters.
The major steps are:
1. Lime clarification with recalcination of the
lime...reducing the suspended solids, the
organic content both suspended and dissolved,
and the phosphate content of the waste.
2. Clinoptilolite ion exchange regeneration and
stripping...removing the ammonia nitrogen from
the waste and concentrating it in a small volume
of ion exchange regenerant from which it is air
stripped.
3. Activated carbon adsorption...removing the
organics residual from lime clarification.
These steps are followed by a final chlorination to assure
disinfection of the water.
In this process all the wash liquors generated in the later
stages are returned to the lime clarification step so that no
liquids leave the system other than in the product water and
the only solid for disposal is the inert residue from the
lime slaker. The organics are burned off in the lime recal-
cination kiln and the ammonia leaves as a gas from the
stripping tower.
Selecting this process for the basis of the present study was
a calculated risk. The process had been piloted on a moderate
scale but as applied to secondary sewage treatment plant
effluent, not to raw sewage. Applied to raw sewage the process
was under study during the design period, at the Blue Plains,
District of Columbia, Experimental Plant operated by EPA. It
appeared, and still appears so, that this process, the so-
called independent physical-chemical process, was a likely
contender for water renovation as against the conventional
biological processes with supplementary (tertiary) treatment,
in part because it was more readily controllable for varying
production rates and not subject to the erratic fluctuations
that plague the biological processes. Furthermore, renovation
by conventional biological treatment followed by tertiary
45
-------
treatment would involve exactly the same process steps--lime,
clinoptilolite or other forms of ammonia removal, and carbon,
and the cost of these tertiary steps would be only very
slightly less than the cost of the same steps applied directly
to raw sewage. In other words, the cost of renovating secon-
dary effluent would have been only slightly less than the
cost of renovating raw sewage and, in addition, there would
have to be borne the basic cost of the conventional biological
process itself. It was decided, therefore, to use the
independent physical-chemical process even though (1) it had
not been fully proven, and (2) the basic performance data
were not yet complete.
Under these circumstances it is readily conceded and should
be emphasized that the present design and computer program is
highly preliminary and is to be considered not an end in it-
self but only one of the steps toward answering the question:
"With what we know about it now, how does reuse compare
economically with a conventional system?"
Not enough is known about the equilibrium, stoichiometric,
and rate processes involved to construct a true mathematical
model, i.e., a model in which any parameter change will be
reflected in a corresponding performance change. For example,
in the activated carbon stage the COD of the effluent is set
by the value assigned to the parameter COD89, in the exem-
plary work, set at 8. The COD actually reached in the
effluent for a fixed feed composition will in a real plant
vary with the equipment configuration, the contact time, and
the carbon loading factor. Lacking knowledge of the relation-
ship among these the program merely provides for setting the
contact time and carbon loading factor...or in other words,
it allows the user to make his own selection of carbon loading
factor and contact time which he thinks will produce the
desired COD in the effluent.
In using the program the user may select any combination of
feed water quality parameters (the ions should be in stoichio-
metric balance) and any reasonable combinations of the other
decision parameters. Using these, the program will design
and cost the equipment so that the plant capability is QDOT,
the design capability mgd, and will cost the operation when
it is operated in that equipment at QBARE, the expected
average production rate, mgd. Obviously, at reasonable
utilization factors (UBAR = QBARE/QDOT) of 0.4 to 0.7, the
product quality will be different at QBARE than it is at
QDOT. For example, the longer contact time in the carbon
adsorbers should improve the COD quality of the product. The
program does not take this into account. It does, however,
reduce the operating costs approximately correctly, for
example, by reducing the cost of pumping energy corresponding
to the flow.
46
-------
GENERAL PROCEDURE IN DEVELOPING THE COMPUTER PROGRAMS
For the LIME subroutine as well as the CLINOPTILOLITE and
ACTIVATED CARBON subroutine , the general procedure was to
develop and debug these subroutines individually and run
some sensitivity analyses on a program version containing
extensive printouts. In these operating labor, engineering,
and buildings were set to zero since they were to be applied
on the entire plant as described beyond.
The sensitivity runs are not explored thoroughly and in some
cases use computations which were later superceded or
corrected. However, the sensitivity runs are not of parti-
cular interest to the parent project.
Buildings
The kiln, the accelators, the carbon regeneration furnace, the
preliminary treatment and the chlorination tanks would not be
housed. All other facilities would be housed. In the present
program the cost of buildings was taken as 10% of the sub-
total investment, which makes allowance for the outside
locations of some portions of the equipment.
Engineering
Engineering as an investment item was applied on the value
of the subtotal plant according to the relation shown in the
ENGR FUNCTION subprogram. This had previously been developed
from a number of engineering fee schedules.
Operating Labor
Operating labor was estimated by visualizing the operating
activities of a 10 mgd plant as in Table 7. Nonoperating
labor was applied according to the schedule in Table 8.
Operating labor at other mgd levels was obtained from the
10 mgd estimate by using the log-log slope typically found
in chemical process plants (1). The estimate for total
labor is shown in Table 9, the exponential relationship in
Line 405 of the Program.
47
-------
TABLE 7
OPERATING LABOR ESTIMATE
10 mgd Plant
Man hours per day
Preliminary 3
Lime 20
Clinoptilolite 18
Carbon 6
Chlorination 2
Disposal 8_
57
TABLE 8
NONOPERATING LABOR ESTIMATE
Man hrs/day
Total Non-
mgd Chemist Clerical Asst.Supt. Supt. Custodial Operation
"~ "" ' l '
.11 0 0 00 8
11 1 0 1 0 24
10 1 1 0 11 32
30 2 2 1 1 2 68
100 42 3 1 3 104
TABLE 9
TOTAL LABOR ESTIMATE
Man hrs/day
mgd Nonoperating Operating Total
.1 8 16 24
1 24 30 54
10 32 57 89
30 68 77 145
100 104 103 207
48
-------
When the nonoperating labor at applicable estimated pay
levels for each category was added to the operating labor
at each mgd level the average total labor rate was 1.12
times the $3.00 per hour operating labor rate. The factor
1.12 in Line 405 makes this adjustment.
Preliminary Treatment
Investment and operating costs for preliminary treatment
were taken from Reference 2.
Residue Disposal
The cost of hauling and sanitary landfill disposal of the
slaker residue is taken as 1.5j?/ton of (moist) residue.
Moisture content was taken as 30%.
The UNITS Subroutine
The UNITS subroutine and associated PARAB3, BAJO and state-
ments in theP rogram determines the optimum number of major
units to be used. The concept is as follows:
The plant design for fail-safe operation requires that the
plant fulfill its function even though one unit is out of
service. By "fulfill its function" is meant to be capable
of the design capability or some specified fraction thereof.
When an equipment unit is capable of being pushed beyond its
nominal capability as for example anAccelator, then this
fraction can be less than 1.0. In the case of the accelators
in this program the fail-safe principle is that they must be
capable of 65% of the design capability with one unit out of
service. The principle in general requires at least two units,
And if there are only two units each must have the design
capability and thus the total capability two times that with-
out the fail-safe principle. If three units are used then
each can have a capability of one-half the design capability
such that the total is only 1.5 times the design capability.
Carrying this to the extreme one would seek the largest
possible number of units limited only by the smallest practi-
cable unit available, so as to have the least excess capa-
bility. However, as the number of units is increased and the
size of individual units decreased the unit cost (e.g., cost
per gallon per day of capability) of each increases.
Accordingly as the number of units is increased on the one
hand the cost is lowered because of lowering the excess
capability but on the other hand the cost is increased because
of increasing the unit cost of each unit. Therefore, there
is an optimum number of units. The UNITS and associated
subroutines determine the number of units which will give a
minimum production cost.
49
-------
In the LIME subroutine UNITS determines the minimum
investments forAccelators and filters (separately),
since the minimum in production costs occurs when these
are minimized individually.
In a complete optimization scheme the number of units would
be simply another one of the parameters to be optimized.
In the present version, only the one dimensional'optimization
is used.
50
-------
LIME CLARIFICATION STAGE DESIGN
A flow sheet of the lime clarification stage with stream
numbers is shown on the next page. The preliminary treat-
ment includes a screen chamber including a comminutor, an
aero-degritter, overflow and by-pass chamber and Parshall
flume.
Basic Performance Data
Table 10 shows the performance of the system at Blue Plains
as operated up to April 1970 (3).
TABLE 10
BLUE PLAINS PERFORMANCE DATA ON RAW SEWAGE
Concentrations in mg/1 (3)
Removals as Fractions
BOD COD TOC* P04 NH-j-N
Raw sewage
After lime
Removal, lime
After filter
Removal, filter
After clinoptilolite - - - - 2.3
Removal, clinoptilolite - - - - .82
After carbon 3.7 8.0 3.7 .64
Removal, carbon .85 .85 .81 .32
Removal, over-all .97 .97 .97 .98 .83
* Total organic carbon
142
31
.78
24
.23
347
66
.81
54
.18
118
26
.78
20
.23
27
1.4
.95
.95
.32
13.6
12.8
.06
-
_
51
-------
10,
Preliminary | 20 = 19> 16> {4> 30> 26f 44
en
to
19
[15
_16 Centrifuge
17
kiln
18
! Milk of lime ]
±
clinoptilolite
regeneration
ge
/
r
-------
At the time this Program was developed there was available
a mathematical model of lime treatment (4). This work on
this difficult subject was not used in the present program
because the design was for Densators, because the perfor-
mance data was for secondary effluent rather than raw waste,
and because the present authors have come to distrust
solubility and solubility product relations as applied to
calcium carbonate and phosphates in waste treatment. It was
preferred to use empirical performance data such as in
Table 10 and 11.
In this program the COD is used as the parameter for the
organics. The removal in the Accelator stages is about 80%,
and about 18% of what remains is removed in the filter. The
phosphate is reduced to 1.4 mg/1. In laboratory experiments
at Cincinnati (5) the PC>4 was reduced to 1-3 mg/1. The
program uses 2.0, a concentration parameter rather than a
fraction removal parameter.
Table 11 shows the values taken for the controlling para-
meters in the lime clarification stage (6,7).
CaO Dosage
To avoid the confusion which exists in the literature over
lime terminology the following definitions are used in this
work:
CaO or calcium oxide - the chemical substance CaO
Ca(OH)2 or calcium hydroxide - the chemical
substance Ca(OH)2
Quick lime - the technical grade solid comprising
CaO as the major component and used for its
CaO content
Hydrated lime - the technical grade solid containing
largely Ca(OH)2 and used for its Ca(OH)2 or
equivalent CaO content
Lime - the technical grade solid, either quick lime
or hydrated lime
Milk of lime - an aqueous slurry containing particulate
Ca(OH)2 and other solids, Ca(OH)2 being a major
component; made by mixing water and lime
53
-------
TABLE 11
LIME CLARIFICATION STAGE ASSUMPTIONS
FOR BASE CASE PERFORMANCE
XMGH12
P0412
TSS12
TSOL13
FCODAC
FCODF
FINRTD
CAC26
CAC27
TSOL14
TSOL15
TSOL16
TSOL17
Mg in Accelator #1 overflow
P04 in Accelator #1 overflow
Total Suspended solids
Accelator #1 overflow
Total suspended solids
Accelator #2 overflow
Total suspended solids
Accelator #1 underflow
Fraction COD entering
Accelators which is removed
in the two Accelators
Fraction COD entering filter
which is removed in filter
Fraction of inerts entering
slaker which leave as
disposed residue
CaCO3 in Accelator #2 underflow
CaC03 in Accelator #2 overflow
TSS in thickener overflow
TSS in thickener underflow
TSS in centrifuge overflow
TSS in centrifuge underflow
pH of Accelator #1 overflow
pH of Accelator #2 overflow
CO2 required to achieve this
Removal of P04 in Accelator #2
CO2 in scrubber exit gas
Ratio backwash to filter exit
mg/1* 3 .
mg/1 2 .
mg/1 10.
mg/1 10.
mg/1 105
fraction .8
fraction .18
fraction .9
mg/1 1.5*105
mg/1 35.
mg/1 100.
mg/1 2.5*105
mg/1 4.7*104
mg/1 6.5*105
pH 11.0
pH 9.5
mg/1 30.
fraction 0.
% 25.
fraction .03
6
3,5
6
6
6
3
3
6
6
6
6
7
7
7
6
6
6
6
7
*Milligrams per liter
54
-------
Reference 8 presents a relationship for determining CaO dosage,
mg/1 from waste water alkalinity, mg/1 equivalent CaC03,
derived from their own investigations as well as several others.
The CaO requirement is that to bring the waste water to a pH _>
11.0. The investigations were performed with actual municipal
waste waters having alkalinities between 60 and 450 mg/1 and
for the Reference 8 sample a PO4 content of 7.2 mg/1. The
composition of the actual waste may vary greatly from those used
in the investigation. A modification may be necessary. A sixth
degree polynominal, that given in Lines 1105-1107 of the
Program, was fitted to the data. The equation gives negative
values for dosage if the alkalinity is less than 60 mg/1. This
is handled by a lower constraint of 50 mg/1 on the CaO dosage.
It should not be trusted for alkalinities greater than 450 mg/1.
Lime Slaker
The particular lime slaker used in this version of the lime
clarification step is the Infilco "Viscomatic" (9) which allows
separation of the impurities, MgO, hydroxyapatite (Ca50H(P04) 3) ,
and inerts, in a grit chamber and their removal by a flight
conveyor (6). The use of this slaker eliminates the necessity
for discarding lime in order to dispose of the impurities. In
the Program the slaker is operated to produce a milk of lime
with a water/CaO ratio of 7.72 corresponding to 1.05 pounds
CaO/gallon. (9)
Reference 9 provides some data points for the relation between
percent CaO in the feed to the slaker and percent Ca(OH>2 in the
residue from the slaker, respectively 0.8747, 0.282, .959, .664,
.959, .628. The relation must also pass through the points 0,
0, 1.0 and 1.0. A relation which approximately reproduces these
pairs is:
FCAOHD = FCAOS**10
The recycle around the slaker is brought to balance by iteration
between Lines 1665 and 1615, the convergence being considered
complete when two successive values for makeup lime differ by not
more than 1% from each other.
The slaker reactions are:
CaO + HO > Ca(OH)
MgO > Unchanged
Ca5OH(PO4)3 > Unchanged
55
-------
Recycle Around Accelators
The process uses Accelators because they are superior to
Densators for the intended service. The recycle of streams
around the Accelators is accomplished by iteration between
Line 1865 and Line 1201 in the Program and using material
balances which involve some approximations and the neglect of
some small streams in the water balance. The recycle is con-
sidered converged when two successive recycle streams (Q20)
are within 0.1% of each other.
The reactions in the first Accelator are :
Ca(OH)0 + Ca(HCO_)_ - > 2CaCO j +2H O
2 o z o, z
Ca(OH)2
3(P04)~~~ + 5Ca(OH)2 - > Ca5OH(PO4)3^ + 9(OH)~
leaving excess Ca (OH) 2 in the liquor at pH 11. Some of the
organics are hydrolyzed, others are brought down with the
solids.
In the first recarbonation and second Accelator the reaction is
Ca(OH) + CO - - > CaCOj, + H0O
2 9 v_.civjVAj'
2
to pH 9.5 with 35 mg/1 CaC03 equivalent in solution.
The second recarbonation merely lowers the pH to 7.5 for the
clinoptilolite stage without further precipitation.
In the thickener and centrifuge i't is assumed that there is no
segregation effect on the various solid components so that the
relative composition of the suspended solids is the same in all
streams. Actually the centrifuge overflow would be found some-
what enriched in Mg(OH)2. The Mg content of the liquor would
in actuality affect the centrifuge performance. At Miami and
also at Austin where the Mg is low the 65% solids is achieved
in the centrifuge cake, but at Dayton with higher Mg only 45%
is reached. Of course, this also depends on the rpm speed and
the loading. Reference 8 uses recarbonation of the sludge to
redissolve about 20% of the precipitated Mg(OH)2 to improve
dewatering characteristics. These things are not taken into
account in the present Program.
Kiln
The reactions in the kiln are taken to be the calcination of
CaCO^ and Mg(OH)2 to CaO and MgO and the conversion of the
original suspended solids to ash by combustion. The fraction
of non-volatile suspended solids in the original suspended solids
is taken as 0.25 (8). The hydroxyapatite passes through the
kiln unchanged.
56
-------
LIME CLARIFICATION PROCESS COSTS
Accelator Investment
From Infilco price lists (6) installed costs of Accelators
were obtained as a function of mgd capability at 1.5 gpm/sf
rise rate. The size range covered was from .07 to 20.16 mgd,
taken as the minimum and maximum size units, respectively. The
relation between 1970 dollars investment and mgd capability was
expressed as a fifth degree polynominal in LNS (natural logar-
ithms) found in Line 2050 of the Program. The NTRAIN refers to
the number of trains of two Accelators each. This is obtained
from a subroutine UNITS which approximates the number and
capability of units such that the production cost will be a
minumum and the system can handle a specified fraction of plant
capability with one unit out of service.
Costs Related to_ Recalcination
Some basic data on actual costs of installed recalcination
facilities for lime softening plants including thickener,
centrifuge and kiln were available in References 7 and 10.
Other workers (10) had computed some of these on the basis
of 1.2 tons CaO per mg of production which obtains at Tahoe.
The available data were further explored with the result that
the 1.2 ratio was considered typical of softening plants as
well as the Tahoe waste treatment plant. Since the calcination
in the softening plants produced a 93% available CaO lime, the
ratio of pounds of kiln product to pounds of CaO in the lime is
1.07. The resulting data ranging from 6.4 to 161 tons per day
of product were fitted with the following equation:
VRCALC (K$, 1970) = 178.959*(TLBS18/2000)**.535022
where:
(TLBS18/2000) = tons per day of kiln product
K$ = thousand dollars
a ratio has this meaning: Probability is 68% that an actual
value will lie between a ratio times, and 1/cr ratio time the
equation value.
Other workers (10) studied the fuel cost for several plants and
used this to generate fuel costs in one of their tables. The
table was reconstituted by taking 1.28 (=1.2 x 1.07) tons per day
kiln product per mgd. The correlation is:
fuel costs ($/day) = 11.249 x (TLBS18/2000)**.767663*UBAR
57
-------
That reference gives an electrical energy cost for recalcination
of O.lC/Kgal which under the assumptions above computes to
0.791 $/ton kiln product. Note that the energy price is buried
in this cost.
Costs of Recarbonation
Investment in recarbonation equipment comprises the recarbon-
ation unit (including the scrubber), the grids and the concrete
basin. Infilco prices (11) and general design data (6) were
used in the costing. The design bases included the following.
The gas to the recarbonators will be 25% CC>2, the price list
specification for Ib CO2/24 hours being for a 12% gas. Thirty
mg/1 of CC>2 is supplied to each unit. The grid area required
is 25 sf/mgd. The basin is 10 feet deep with bottom and walls
one foot thick. Concrete costs are 100 $/cu. yd. installed and
grids cost 6 $/sf.
Costs of recarbonation units were available from 104 to 10,400
Ib 25% CC-2/24 hours corresponding to mgd's of .51 to 51. The 13
data points were fitted to a third degree polynominal in LNS
(Natural logarithms) which is found in Line 2076 of the Program
with a 35% increase for installation. For large units the 50
mgd value was extrapolated at a In-ln slope of 0.7 according to
the relation in Line 2073.
The same references supplied data over the same range on the
horsepower of the blower as a function of mgd to which the
following relation was fitted:
COMPHP = 1.30789*Q12**.668189 -0.45166
cr ratio =1.10
The conversion of this to energy costs for the two recarbon-
ators is contained in the expression in Line 2248 of the Program
along with the energy cost arising from pumping to the filter
and to one Accelator.
Cost of Filtration
References 11 and 6 supply data on costs of dual media filter
plants installed, over the range from .7 to 28.8 mgd, operated
at 2 gpm/sf. In the Program the filters are operated at GPMSFF
set to 4. and multiple units are used beyond 70 mgd at this rate,
(10410 sf). The relation between installed costs K$ and square
feet was expressed as a fourth degree polynominal in LNS, and is
found as the exponential in Line 2150 of the Program, having
been trended to 1970, N = llr a ratio = 1.038. The pumping
energy requirement is figured assuming a -A P of 40 psi and an
efficiency of .75.
58
-------
RESULTS OF INDIVIDUAL LIME CLARIFICATION RUNS
On the above basis individual runs on the design and costs of
the lime clarification stage alone were made using parameter
values shown in Table 11. Figure 9 shows the cost versus
QBARE relation, and Figure 10 shows the sensitivity around the
base case to certain of the parameters. The waste used was
typical of San Antonio waste.
Figure 9 shows a cost of 6.87 C/Kgal for the base case at
100 mgd, utilization factor UBARE =0.5. (UBARE = QBARE/QDOT)
The computations for the present project are being carried out
on the basis of present technology and the parameter values
chosen are representative of that level. However, Figure 10
shows that if the filters could be eliminated by placing the
filtering burden on the clinoptilolite and the activated carbon
beds the costs could be reduced by about 1.3 C/Kgal at 100 mgd.
The filters comprise 21% of the investment. At 2 gpm/sf
(gallon per minute per square foot) instead of 4, the filters
would comprise about half of the total investment, greater than
the cost of the Accelators and twice as great as the recal-
cination facility.
The sensitivity relations in Figure 10 show that the costs are
quite insensitive to the water quality parameters and also to
the fraction of inerts disposed of from the slaker. The process
is highly capital intensive and therefore very sensitive to
UBARE which therefore appears as the only controllable decision
parameter by which costs might be reduced.
59
-------
QDOT, mgd
20
200
100
10
Operating labor set to zero
Disposal not included
LIME STAGE PRODUCTION COST
Versus
QBARE and INVESTMENT vs QDOT
LIME STAGE
INVESTMENT
C/gpd of QDOT,
1969
LIME STAGE
PRODUCTION COST
C/Kgal, 1969
«. without
filters
100
10
To
QBARE, mgd
Figure 9
-------
LIME STAGE PRODUCTION COSTS
C/Kgal, 1969
— 6
- 5
P04
X
ALK
Sensitivity of Lime Stage to Various Parameters
QBARE = 100
Operating labor set to zero
Disposal not included
— 4
.5
40
40
I
.6 PINRTD .8
120 200 ALK
50 CaCO3
10 PO4 20
.3
I
UBAR
.9
360
80
25
L
.5
1.0
90
30
520
100
35
1_
.7
40
120
Figure 10
1.0
-------
CLINOPTILOLITE ION EXCHANGE STAGE DESIGN
A schematic diagram of the clinoptilolite ammonia exchange
process is shown on the next page. The feed liquor exit from
the lime clarification stage is fed to several clinoptilolite
ion exchange beds in parallel. When a bed has reached exhaustion
it is piped into the regeneration system. In the first regen-
eration step a regenerant liquor containing about 100 mg/1
NH3~N is circulated through the bed via Tank A until it reaches
a concentration of 600 mg/1. The bed is then piped into Tank B
containing 10 mg/1 liquor and recirculated until the liquor has
built up to 100 mg/1. Finally the exchanger is flushed with
product water to free it from residual brine and lower the
pH to 8.5. Then it is returned to the exhaustion line.
In the exhaustion stage the NH3 in the liquor exchanges for Ca
on the clinoptilolite. In the regeneration stage Ca(OH)2 is
the regenerant exchanging for NH3 on the clinoptilolite.
When Tank A has reached 600 mg/1 it is piped to the stripping
columns where it is stripped in two columns in series with air
which when necessary is heated to keep the liquor temperature
above 25°C. The stripped liquor contained in Tank C is trans-
ferred to Tank B of the regeneration line for use in subsequent
regeneration. A clarifier may be needed ahead of the stripping
towers, but it is not included here.
This process has been piloted (12, 13). Other work had been
done on ammonia stripping some of which was used in the present
design (14). However, most of the practical design conditions
represent the technology as of about April 1970 (15).
Basic Design Assumptions
Table 12 shows the design parameter assumptions for the clin-
optilolite stage. In addition to the NH^ the clinoptilolite
also must exchange other ions particularly K+ for which it is
selective. This K+ and the other ions must build up i.n the
regenerant liquor since only NH3 and any other gas-forming ions
are removed by the stripper.
62
-------
CLINOPTILOLITE AMMONIA EXCHANGE, SCHEMATIC
EXHAUSTION
i '
To Brine Dilution
and Ca(OH) Dilution
REGENERATION
_L -i
Clino Tank
A
100-600
nig/lNH3-N
•._-_,.]. ••<»•— _ .. ... —
Regen, Step 1
STRIPPING
Tank A
600 mg/1
NH3-N
Brine . .
Tank
M»n 1 iPa/OF-A i
^ J, 2 H
— > Fuel
/•
IGa(OH)
and Nad
1 Drain I
Tank 1
•l> V V
Clino
Clino Clino
^ « ^
. pH 8. 5 from
\ f nV
..... ..... 1 f-1/ni-n
^ '2
__L_- + — L_. (Residual
Clino
Tank Clino Brine)
B
10-100
mg/1 NH3-N
t t
Regen, Step 2 Regen, Step 3
(?
/
V
?
strip
Air
X
'
--X
tripper gas outy
~~ — ^^-^
v A
[ ^L
TankC
Strip LO mg/1
NH3-N
S*
/
eater \
Exhaust Gases
^
~>J Blower
r
Stripping
To Product (see above)
63
-------
TABLE 12
CLINOPTILOLITE STAGE ASSUMPTIONS
FOR BASE CASE PARAMETER
Symbol Description
AN79 NH3-N in effluent
CLLF Clinoptilolite exchange
capacity
FCLPR Clinoptilolite attrition
loss per regeneration
GPMSFX Maximum allowable flow
in exchanger
GPMSFS Liquid loading in stripper
CFAPG Air/1 iauor flow ratio in
stripper
DIAMAX Maximum allowable diameter
of exchanger
DEEMAX Maximum allowable depth
of Clinoptilolite bed
EPUMP Efficiency of pump, wire
to water
EBLOW Efficiency of blower, wire
to air
Fraction of NH3~N removed
in each stripper
PRCLIN Price of Clinoptilolite
NDTL77 Days per year less than 77°F
AVTL77 Average temperature when
less than 77°F
PRNACL Price of salt
PRMBTU Price for heat
PRLIME Price of lime
FACAOL Fraction of active CaO
Units
mg/1 N
Lbs N/cf clino
fraction of bed
volume
gpm/sf cross
section
gpm/sf cross
section
cf /gal
ft.
ft.
fraction
fraction
fraction
$/cf
No . days
oF
$/lb
$/million BTU
$/ton
Value
0.5
.17
.0026
6.
5.
300.
25.
10.
.75
.60
.85
10.
100.
67.
.01
.5
18.5
in purchased lime
fraction
.9
64
-------
Consequently it should be necessary either to discard some
regenerant liquor or to return it to the process in order
to prevent the buildup of these other ions. The present
Program does not take this into account but assumes that
nothing but NH4+ is exchanged on the clinoptilolite.
Exchanger Number and Size
The size of the exchanger is controlled by a maximum bed
depth, a maximum gpm/sf and a maximum diameter. The number
of exchangers on line and the capability of each is obtained
via the UNITS subroutine. The diameter and height are then
adjusted to the next half-foot dimension. The time to
exhaustion is computed from the bed volume per stage and
the clinoptilolite capacity. The number of exchangers in
regeneration at any time is computed on the basis that
regeneration requires nine hours, four hours circulation
in each tank and one hour for the rinse.
Regeneration System
Each regeneration system consists of three holding tanks for
regenerant solution, a pump for each tank, one drain tank,
and a brine tank and saturator for all the regeneration
systems. It is assumed that each regeneration system will
simultaneously handle two exchangers. The capacity of each
holding tank is four bed volumes. The pumps are sized to
circulate 10 bed volumes per hour at 35 feet head.
The initial charge of NaCl necessary to bring the regenerant
liquors up to 0.1 normal is neglected and the NaCl consumption
is computed on the basis that at each regeneration the regen-
erant liquor residual in the bed (50% voids) enters the
product.
Stripper System
Each exchanger regenerated requires four bed volumes of eluent,
The capability of stripper is generated from that, and the
area of each stripper is determined from a loading GPMSFS,
taken at 5 gpm/sf. Two pumps are required/ each capable of
the stripper capability. The regeneration system pumps are
used for returning the stripped eluent to storage tanks. The
capability of the blower to handle the two strippers is taken
at 300 cf air/gal. On days when the liquor is below 25°C
(77°F), heating will be required to raise the temperature to
77°F. Parameters are inserted for an arbitrary number of days
per year below 77°F and the average temperature on those days.
65
-------
Instead of the Marley 20 foot cross-flow tower type of
stripper used at Tahoe, this design utilizes a stripping
tower similar to a square-type forced-draft aerator because
of the greater ease of access for cleaning. The Tahoe
stripper operates on secondary effluent with Nt^-N
25-30 mg/1 and achieves an 85% removal at 3.2 gpm/sf and
240 cf air per gallon. In the present design it will be
assumed that a square-type aerator tower 10 feet in height
will give the same 85% removal per stage operating on
NH3-N concentrations of 600 mg/1 in the first tower and
90 mg/1 in the second.
66
-------
CLINOPTILOLITE ION EXCHANGE
STAGE PROCESS COSTS
Ion ExchangerSystem Investment
Another Program (IONEX) (Reference 16) had developed costs
of an ion exchange system including vessels, the regenerant
tanks, and the associated valving, piping and controls. The
relation appears in Line 6005 of the Program. This equation
had been developed for commercially available ion exchange
units with bed volume less than about 200 cf and is used in
the Program for which it was developed to the largest size
commercially available exchangers of about 800 cf. For the
Clinoptilolite Program it will be assumed '' t the equation
can be extrapolated to exchangers 25.5 feet^ in diameter and
of 10 foot bed depth.
Tanks Investment
The cost of tanks is taken from data in Reference 17 and the
relation is for example in Line 6020 of the Program. The
factor 1.23 is a factor after Reference 18 to account for
concrete foundations not included in Reference 17.
Pumps Investment
The relation for pump cost is developed from Reference 18
which gives uninstalled costs of carbon steel pumps with
driver and a factor (2.4) to obtain installed costs including
indirect labor. The cost of the piping is assumed to be
adequately covered by the costs for the exchanger system.
The relation appears in Line 6015.
Blowers Investment
Blower horsepower may be computed by Reference 19
HPBLOW = CFMAIR* P (in.H2O)/(6356* efficiency)
in. = inches
The efficiency of blowers is about 60%. Reference 14 computes
a A P of 0.9 inches H^O for a tower 10 feet wide and 27.5
feet high. A A P of 1.5 inches is assumed for the tower to
be used. The CFMAIR is actually to be measured at the suction
pressure and temperature. These data result in the horse-
power equation given in Line 5330 of the Program. The maximum
blower size is taken at 5000 HP.
The relation between blower cost and horsepower is taken from
Reference 20 and the graph there gives results in a relation
shown in Line 6045 of the Program where the factor of 2 enters
because there are two stripping towers.
67
-------
Stripping Tower Investment
The stripping tower to be used is similar to a square-type
forced-draft aerator for which manufacturers prices were
obtained (21). These prices include blowers. The estimated
blower cost was subtracted from the quoted prices and the
installed price adjusted for three additional trays to get
the 10 foot height to be used. A resulting relation between
tower cross-sectional area and cost is that shown in Line 6035
of the Program.
Pipe Investment
It is assumed that piping for the tower is similar to that
for horizontal pressure vessels at an estimated cost of 42%
of the tower cost (18) . Assume that "piping" for the blower
is for the duct work associated and is similar to that for
air coolers, factor 18% (18).
Energy Cost
Electrical energy is needed for regeneration pumps, stripper
pumps, stripper blowers, and exchanger pumps. It is assumed
that the flow through the exchanger is laminar and consequently
pumping power is proportional to square of flow rate. For the
blower the flow is assumed turbulent so that the energy is
proportional to the flow rate cubed. This computation is in
Line 6535 of the Program.
Lime Cost
The lime consumption is taken as that equivalent to NH3
released, in Line 5175 and the cost to purchase this is
computed in Line 6520. If there is excess Ca(OH)2 from
the lime clarification stage the cost is proportionately
reduced in Line 6522.
-------
RESULTS OF INDIVIDUAL
CLINOPTILOLITE ION EXCHANGE RUNS
On the above basis, individual runs on the design and costs
of the clinoptilolite ion exchange stage alone were made
using the parameter values in Table 12. Figure 11 shows the
cost versus QBARE relation and Figure 12 shows the sensitivity
around the base case to certain of the parameters. The base
case at 100 mgd shows a cost of 7.42£/Kgal. The cost is not
decreasing very much beyond 100 mgd because the maximum sizes
of equipment have been reached.
The clinoptilolite ion exchange stage is highly sensitive to
the parameters. It is thought by some that the clinoptilolite
loss fraction might eventually be reduced to l/10th of its
current value, in which case the cost would be reduced by
1.7£/Kgal. Also, the current price of clinoptilolite at 10$/cf,
about 30£/lb seems quite high for a mined and processed mineral.
The initial clinoptilolite comprises about one-quarter of the
investment and the makeup clinoptilolite about one-quarter of
the production costs. The cost is particularly sensitive to
the ammonia nitrogen in the feed, approximately proportional to
it, thus if the NI^-N in the feed were 20 instead of 15 the cost
would be increased by 1.4£/Kgal. The cost is extremely sen-
sitive to the maximum diameter allowed for the exchanger
vessel, reaching a minimum at about 2C/Kgal below the base
case somewhere in the region of 75 feet.
However, the present application of this Program in this
project is restricted to present technology.
69
-------
100]
CLINOPTILOLITE PROCESS
Production Cost & Investment
vs.
QBARE
UBAR =0.5
Influent AN = 15 mg/1
Effluent AN = 0.5 mg/1
Bldg, engr, & operating labor
costs set to zero
10
10
Investment
CLINOPTILOLITE PRODUCTION
COST, C/Kgal, 1969
INVESTMENT
«/gpd
of QDOT
10
QBARE = Expected Production Rate, mgd
100
1000
Figure 11
-------
11
CTl
U)
03
0)
U
2
•H
±> I
a I
o
•H
U
6-
VI
0
U
c
0
u
3
-O
O
rO
JJ
O
EH
5-
Sensitivity of
CLINOPTILOLITE PROCESS
to Various Parameters
QBARE = 100 mgd
Operating Labor = 0
002
3
0
0 25.5 50 75 100 125
DIAMAX - Maximum Exchanger Diameter, ft.
,0010 .0018 .0026 .0034 .0042 .0050 0.0058
FCLPR - Fraction, Clino Lost/Regen.
456 7 8 9 10
GPMSFX - Imposed, GPM/ft2 to exchangers
5 10 15 20 25 30
AN - Ammonia Nitrogen, mg/1 as N
15
"OT
0.4 UBAR O.S CT^ (TTT CT^-
- Parameters Varied in Sensitivity Study
TTTT
TTO
Fig. 12
-------
ACTIVATED CARBON ADSORPTION STAGE DESIGN
On another project the authors, like a number of investigators,
have attempted to find a basis for the rational design of
activated carbon adsorption for waste waters only to conclude
that the behavior of waste water with activated carbon does
not follow the laws and relations which have been developed
for other adsorption processes and upon which scientific
designs can be based. Without actual experimentation the
presently available data do not allow the prediction of
required contact times and carbon loading factor (capacity
for organics) necessary to achieve a given effluent. In view
of that, the present design uses a carbon loading factor and
a contact time as input variables. Until better information
is developed the designer must first determine experimentally
with the liquor to be processed just what carbon loading factor
and contact time is required to achieve his specifications.
The contact time of 50 minutes and the carbon loading factor
of 0.5 IbCOD/lbC (pounds COD per pound carbon) for an effluent
of 8 COD, used in the base case, are thought to be conservative
for most liquors and typical carbons.
The term "contact time" is reluctantly retained in this report
because it has crept into common usage. However, it should be
recognized that this term is a misnomer and does not reveal
the average time of contact of the liquor with the carbon.
"Contact time" is the time required to pass one bed volume of
feed liquor. Since carbon has a void space of about 36-38%,
the actual average time of contact of the liquor with the
carbon is about one-third the "contact time." A rational
unit, preferable because it avoids the indefiniteness of
"contact time" is gpm/cf (gallon per minute per cubic foot)
which is now coming into use in ion exchange technology. The
relation is:
7.48/(gpm/cf) = contact time, minutes
Fifty minute contact time is about .15 gpm/cf.
The two major design and cost studies on activated carbon (as
applied to secondary effluent) are those of References 22,
23, and 24. Condensed cost data from 22 appear in Reference 25,
Most of the performance relationships and cost data used herein
were taken from these two studies.
72
-------
A schematic of a two-stage, two-train adsorption process is
shown on the next page. The term "train" refers to the two
series of adsorbers, ABE and CDF, run in parallel. The term
"stage" refers to the two contactors on line at any one time
in each train, in the sketch, A and B and C and D. The feed
liquor flows through the two stages in series and the two
trains in parallel and issues as product. When lead adsorber,
A, reaches exhaustion it is taken off stream and the second
adsorber B is made the lead adsorber, the spare E becoming
the second adsorber. The sketch is shown at the condition
in which adsorber E containing carbon has just been taken off
the line. Adsorber F is empty. The spent carbon is conveyed
as a slurry to a screen from where it is passed through a
rotary hearth regeneration furnace where it is reactivated
in a steam atmosphere. As it leaves the furnace it is
quenched and transported again in slurry form to adsorber F
which had been empty. The regeneration operation and this
transfer require many days and is timed so that adsorber F
will be full of carbon and adsorber E empty, by the time it
is necessary to place F in service through the exhaustion of
C. Makeup carbon is added to replenish the carbon loss in
the regeneration process. In the complete recycle scheme all
backwash waters, scrubber waters and carbon fines are returned
to the Accelator, but this return is not taken into account in
the present design.
Design of Adsorber Beds
Central to the design after the contact time and carbon loading
factor is the constraint comprising a maximum gpm/sf (gallon
per minute per square foot), GPMSF, through the adsorber.
The total required volume of adsorber beds on line (BEDVOL) is
computed from QDOT and contact time. If the gpm/sf through
the adsorber is greater than about 13, the pressure drop through
the adsorber may be greater than 15 psi which is considered a
limiting level. Bed depth is then computed from gpm/sf, contact
time, and number of stages, and is brought back to the con-
straints of 2.5 and 25 feet if outside these. The UNITS sub-_
routine determines the optimum number of trains and the capacity
of each. The bed volume per stage and the diameter is then
computed.
The number of spare adsorbers, E and F of the sketch, is taken
as two for every 10 trains or fraction thereof.
73
-------
EXAMPLE OF TWO-STAGE CARBON ADSORPTION PROCESS
FEED
OPTIONAL
SURGE
BASIN
95% Of
Backwash
pumi
5% of
5
tic
flG
/
W
A
W
V
X1
A
\>.
\
*
^
>
t
>
^-^
ds(
w
N
S
(
h
,
v '
^
-4
t
arb
S
>
t
}
> J
•^
A
\
>
t
N.
e
f
N.
^
r
:
/
^
^_j
V
ri
X
k
V,
, ,
s
—
f
*• »
<1
f -
t
*
—
D
1^1
>
f
)
1
X.
f
r
J
x.
r
e]
-f
J
(
^
1 -
I
I
^^
V
' .
'-. ._ ,
s
—
'—••,
S
^
f
f
F
ri
v \
r1
(
_j
X,
>
>
)«
£
t
X
'
^
>pe
:ar
5lU
;an
nt
bon
rry
k
^
?rodi
Carbon slurry
water
Induced draft fan
Wet
scrubber
Stationary
screen
ZeolrLte
Natural gas
74
carbon
slurry tank
-------
Furnace Area
The required hearth area of the furnace, sf, was correlated
with Ibs (pounds) C/day regenerated from the data of
Reference 22. The relation is found in Line 8400 of the
Program, obtained from 17 points, o* ratio 1.13.
Energy Consumption
The two basic cost references are not in good agreement on
the energy consumption. Shown in Table 13 is the estimated
consumption for a two-stage plant of three sizes.
TABLE 13
ENERGY CONSUMPTION, KWH/Kgal
(kilowatt hours per kilogallon)
mgd 23, 24 22^
1 1.21 .496
10 1.00 .773
1.36
100 .79 .372
All five points were used in developing the relation:
KWH/Kgal = .939669*QBARE**(-.0775243)
7 points, cr ratio =1.51
Backwash Systems
The number of backwash systems was taken as one-half the
number of spare adsorbers, i.e. one for each 10 trains or
fraction thereof, but this is not used in the final version
of the Program.
75
-------
ACTIVATED CARBON ADSORPTION
STAGE PROCESS COSTS
Cost of Adsorber Vessels
The two main cost references present the cost of adsorber
vessels as well as of other equipment in different ways.
Reference 22 provides a "capital cost" to which is added
40% for contingencies, engineering and profit. In the
present study Reference 22 capital cost figures are
increased by 30% to obtain investment ex engineering and
buildings, 10% of this 40% being considered as engineering.
Reference 23 presents "major materials" which is multiplied
by a factor of the order of 1.7 to obtain plant investment.
In the present study the Reference 23 "major materials" are
multiplied by 1.7 to obtain investment ex engineering and
building. When plotted against total volume in the adsorber
vessel including freeboard the two sets of raw data adjusted
to 1968 define two trend lines having the same log-log slope
but with Reference 23 (12 points, o~ ratio 1.10) being more
than two times the Reference 22 points (24 points, o" ratio
1.29). Both sets of points began at about 2000 cubic feet
but the Reference 22 points extended to 140,000 whereas the
Reference 23 points extended only to 30,000. For that reason
the final chosen equation was obtained by taking the arithmetic
average of the exponents and the coefficients of the two sets
yielding the equation in Line 8505 of the Program:
VADS, K$ = .187254*VOL**.593364
where
VADS = K$ per adsorber vessel
VOL = cubic feet of volume in vessel including
freeboard
Cost of Regeneration System
The cost of the regeneration system in Reference 22 is related
to the square feet of furnace hearth by the exponential relation
in Line 8510 of the Program; number of points 10, range 28 to
1600 sf, o* 1.10. Derived from the Reference 22 40% for contin-
gencies engineering and contractors profit, a factor of 1.3 is
used to generate investment, 10% being considered as engineering.
76
-------
Pumps and Sumps Cost
The two main cost references differ in their estimated costs
for pumps (23) and the corresponding pumps and sumps (22). For
a two stage system Reference 22 costs which are independent of
contact time, increase1 as the 0.3 power of mgd over 10-100 mgd.
The cost at 100 mgd is less than two times that at 10 mgd.
Reference 23 costs are higher than the Reference 22 and have
about a 0.6 factor. Because of the more reasonable slope the
Reference 23 costs are chosen, the relation being:
K$(1968), two stages = 29.3693*QDOT**.710743,
found in Line 8514 in the Program.
With respect to the number of stages the Reference 22 costs at
10 mgd are independent of contact time and increase as number
of stages increase. Reference 23 costs at 50 minute contact
time and 10 mgd decrease as number of stages increases. Unable
to resolve this difference without detailed design work, the
present Program takes pumps and sumps cost as independent of
number of stages. The cost of pumps and sumps is less than 5%
of the total capital cost at any mgd size.
Reference 22 gives separately the cost of the backwash system,
largely independent of the number of stages, and which is
related to the gallons per minute backwash rate by:
K$(1968, investment ex engineering) = 0.444*GPMBW**.493
This is about the same as the Reference 23 investment for
"pumps" at 1 mgd. If it is assumed that the Reference 22
pumps and sumps plus backwash systems correspond with the
Reference 23 pumps then it is found that the Reference 22
costs for the two are about the same as the Reference 23
costs at 1 mgd but are still well below the Reference 23
costs and with an 0.3 factor at the larger sizes. Accordingly,
the Program uses the Reference 23 costs as representatives of
pumps and sumps and backwash systems.
Piping Costs
Both references give piping costs. When adjusted to investment
ex engineering and plotted against gpm flow per train, (adsorber
flow rate) the two sets follow a single trend line (Line 8525
in the Program):
K$ per adsorber, 1968, ex engineering = 2.41222*ADSFR**
.386757
24 points, CT ratio 1.29
where
ADSFR = flow rate, gpm per adsorber train
77
-------
Cost of Electrical System and Instrumentation
Reference 22 gives separately the costs for electrical system
and for instruments. The ratio of the cost for both together
at NSTAGE (number of stages) to the cost at two stages follows
closely:
ratio = .70455 + NSTAGE*.147708
At two stages the costs bear the following relation to mgd:
K$(1968) electrical plus instruments ex engineering =
46.4174*QDOT**.386189
It is assumed that the stage ratio at 10 mgd is also applicable
at other mgd levels. The combined equation is found in Line
8530 of the Program.
Concrete Costs
References 23, 24 give separate costs for concrete, some
of which in some cases are used for a one-day surge basin not
included in the present design. A correlation could not be
established. Concrete costs for foundations, etc., instead
were taken from Reference 18 who gives material costs for
concrete as a percentage of bare module cost (corresponding
to investment installed ex engineering) of 1.2% for centrifugal
pumps and drives, 2.3% for process vessels, and 4.4% for fur-
naces. It is assumed that the installed cost of concrete will
be about four times the material cost leading to the relation
shown in Line 8545 of the Program.
Makeup Carbon Cost
Carbon loss in regeneration is taken at five percent per
regeneration. The price used in the exemplary runs is 26C
per pound. Fuel consumption is taken as 4250 BTU/lb (British
thermal units per pound) carbon regenerated (24). In the
exemplary runs the price of fuel was taken as 25C/mBTU
(million BTU).
78
-------
RESULTS OF INDIVIDUAL ACTIVATED
CARBON ADSORPTION RUNS
On the above basis individual runs on the design and costs
of the activated carbon stage alone were made with parameter
values as in Table 14. Figure 13 shows the cost versus QBARE
relation.
TABLE 14
PARAMETER VALUES USED IN BASE
CASE ACTIVATED CARBON RUNS
Variable
Names
Description
Values
Used
NSTAGE
GPMSF
RHO
PRCAR
PRFUEL
GPMBW
(not used)
COD 8 9
CT
CLF
Number of adsorber vessels in a train
Flow rate through beds
Bulk density of carbon
Price of carbon
Price of fuel
Backwash rate
COD in process effluent
Contact time
Carbon loading factor
Fraction of QDOT under fail-safe
condition
6 gpm/sf
30. Ibs/cf
26. «/lb
25.
-------
100
ACTIVATED CARBON
Investment & Production Cost
(ox engg ,—building,—labor)
UBAR =0.5
CODIN = 82,CODEFF 8 mg/1
CLF = .5 Ib COD/lb carbon
CT = 50 min.
GPMSF = 6 gpm/sf.
10
I
00
o
Investment
Prorh
PRODUCTION COST
C/Kgal
or
UNIT INVESTMENT
-------
11
Effects of CODIN, CT, and CLF, etc
on Total Production Costs for
Activated Carbon Plants
QBARE •-•• 100 mgd
QDOT = 200 mgd
CODF-:FF - 8 mg/l
NSTAGE = 2
Note: Operating Labor Costs
Set to Zero
Costs for provision off
added contact time^* '
by increasing
3 stages per
train
•p
in
o
u
c
o
•H
-u
u
3
-------
It is seen that the cost is highly sensitive to the contact
time. A drop in the contact time from 50 minutes to 30
minutes would lower costs by about l.SC/Kgal. A special
study was made to determine the effect of providing additional
contact time by increasing from two stages to three stages per
train. The results, shown, confirm other more detailed studies
in which two stages have proved to be the optimum.
The cost is reasonably sensitive to the COD in the feed to the
activated carbon stage indicating the desirability of high
removals of COD in the lime clarification stage, if possible.
The sensitivity to the carbon loading factor is not very great,
nor is the sensitivity to the gallons per minute per square
foot flow rate.
The cost is highly sensitive to utilization factor since the
process is highly capital intensive. In the final Program at
100 mgd more than half the production cost comes from the
capital charge and about a quarter from the maintenance and
repair which is proportional to investment.
82
-------
COMPLETE AWT PROCESS RUNS
The Computer Program for the complete AWT process comprising
preliminary treatment, lime clarification, clinoptilolite
ion exchange, activated carbon adsorption chlorination, and
ultimate disposal is given in a following section. Thereafter
follows the definitions of the variables and a description of
the Program. The exemplary runs were made on a composition
typical of San Antonio sewage shown in Table 15.
TABLE 15
TYPICAL SAN ANTONIO SEWAGE COMPOSITION
USED IN EXEMPLARY RUNS
SMATX Value Used in
Subscripts Contaminant Exemplary, mgpl
(1,3) Na 174. (adjusted to achieve
ionic balance)
(1, 4) K+ 12.
(1, 5) Ammonia N 15.
(1, 6) Ca""" 80.
(1, 7) Mg^ 18.
(1, 8) Cl" 82.
(1, 9) F" .3
(1, 10) N02" 2.
(1, 11) N03" 12.
(1, 12) HCO_" 360.
-------
Other data common to all subroutines used in the exemplary
runs are:
Parameter
Name
Description
Value Used
In Exemplary
PLF
TX
XINS
RET
CKWH
CYMSI
RTLAB
PYEX
Plant life
Tax rate
Insurance rate
Insurance rate
Energy price
Current year Marshall &
Stevens
Chemical Process Industries
Equipment Cost Index
Base labor price
Payroll extras factor
20 Years
.01 Annual fraction
of investment
.01 Annual fraction
of investment
.045 Annual fraction
of investment
1.C/Kwh
285. for 1969
3.00$/man hour for
operating labor
.45 Fraction of
payroll
A printout of the base case at 100 mgd, utilization factor 0.5,
is given on the next two pages with 1969, National costs. The
individual process summaries show separately the amortization
and operating costs for each of the three major processes.
These costs do not include buildings, disposal, engineering or
operating labor, which are treated as separate "processes."
Under the heading "OVERALL AWT PROCESS" the term DISP, TPD, in
this case 80.821 tons per day, refers to the quantity of moist
residue for ultimate disposal under conditions of average pro-
duction. The table "CONTAM. IN OUT" refers to the mgpl com-
position of the overall feed and the product. The investment
costs are shown for each process in terms of K$, percent of
the total by process, and unit investment in C/gpd of capability
(QDOT) and of average production (QBARE). Production costs,
operating plus capital charge, are shown by process in K$/yr,
C/Kgal, and percent of the total.
At 100 mgd average production and 200 mgd capability the total
AWT cost is 26.748<:/Kgal of which about 37% arises from the
carbon stage and about 24% from each of the lime and clinoptilo-
lite stages. Overall capital charges amount to 14.728C/Kgal
and operating costs 12.020.
84
-------
10/27/70. 17.22.06.
POGrfAiXi A.WT27
•LIiXiE PHOCESS-
WEAEE 100.000 GDOT 200.000 UB.AR .50000
CAC03 BO.O* i-:G 18.0* P04 25.0* ALK 360.0* TSS10 220.0* COD 500.0
COST SUMMARY* CE^TS/KGAL
A.N.ORT. = 3.766* OPEF. = 2.722* TOTAL = 6.488
•CLINOPTILOLITE PROCESS-
VARIABLE PARAMETERS FOR THIS CASE
QBAHE 100.00 QDOT 200.00 UBAR .50000 AN 50 15.00
COST SU:-aRA;iY* CENTS/KC3AL
AMOHT. = 3.025* OPEFi. = 3.439* TOTAL = 6
ACTIVATED CAEBOw PROCESS
QDOT = sou.ooo MGD* v^rr = 100.000 MGD* UBAR = .5000
CLF = .500* CT = 50.0* ;,0blw = bri-G* UODEF = 8.0
COST SUMMARY, CENTS/KG.AL
At-'OP.T. = 5.827* OPEH. = 3.963* TOTAL = 9.790
85
-------
OVERALL AWT PROCESS
QBARE 100.00 QDOT 200-00 UBAR .50000 DISP.*TPD
CON'TAM. IN
MA 174
K 12
NH4 15
CA 80
MG 16
CL 82
F
N02 2
N03 12
HC03 360
COS 0
S04 52
SI 03 15
P04 25
COD 500
VSS 163
NVSS 57
INVESTMENT
PROCESS
PRELIM.
LIfcE
CLINOF.
CARBON
CHLOR.
BUILDINGS
DISPOSAL
ENGF.
TOTAL
OUT
.00 160.12
.00 12.00
.00 .50
.00 64.02
.00 1.25
.00 99.44
.30 .30
.00 2.00
.00 12.00
.00 69.09
0.
.00 52.00
.00 15.00
.00 .90
.00 8.00
.00 .70
.00 .30
COSTS, K$
PCT. OF CENTS/GPD OF
KS
483.163
14189.357
11395.506
21954.947
478.042
4850.102
0.
2138.145
55489. 261
TOTAL
• 87
25.57
20.54
39.57
.86
8.74
0.
3.85
100.00
ODOT
.242
7.095
5.698
10.977
.239
2.425
0.
1.069
87.745
GJBARE
.483
14.189
11.396
21.955
.478
4.850
0.
a. 138
55.4g9
PRODUCTION COSTS
PROCESS
PRELIM.
LI WE
CLINOP.
CARBON
CHLOR.
BUILDINGS
DISPOSAL
ENGR .
OPH. LABOR
TOTAL
KS/YEAR
74.131
2367.977
2359.267
3573.404
204.292
469.859
44.276
207.135
462.707
9763.052
CENTS/KGAL Pi
• 203
6.488
6.464
9.790
.560
1.287
.121
.567
1.268
26.748
.76
24.25
24.17
36.60
2.09
4.81
.45
2.12
4.74
100.00
COST SUMMARY, CENTS/KGAL
AMORT. = 14.728* OPER. = 12.020, TOTAL = 26.748
86
-------
With the 100 mgd as a base case the Program was used to
generate costs at other average production levels. The
results are shown in Figure 15 and Table 16. The unit
costs decrease in a norirtal manner with increasing average
production. The change is very small beyond 100 mgd
primarily because the maximum sizes of most of the equipment
units have been reached and capability increases are achieved
by replication. At 100 mgd the optimum design calls for 11
Accelator trains, 5 filters, 47 clinoptilolite exchanger
trains, and 47 carbon adsorber trains, all of these close
to the maximum sizes available. Obviously future cost pro-
duction studies should explore the possibilities of relaxing
these upper constraints on equipment sizes. The present
project has not thoroughly explored these constraint relax-
ations.
TABLE 16
EFFECT OF QBARE ON AWT COST
QBARE .1 1
QDOT .2 2
Unit investment 497.0 106.4
C/gpd of capability
Unit production cost, 508.9 112.7
C/Kgal
10 100 177 1000
20 200 350 2000
43.79 27.75 26.38 23.82
43.62 26.75 25.05 22.57
Percent Contribution of
Major Processes to Production Cost
Lime
Clinoptilolite
Carbon
35.4
8.3
19.6
29.3
13.3
22.3
29.7
18.1
28.5
24.2
24.1
36.6
23.2
25.3
37.8
21.2
27.6
39.8
Figure 16 shows the sensitivity of the overall AWT process to
utilization factor and to the fraction of the COD removed in
the Accelators. The process is highly sensitive to utilization
factor. In the 100 mgd size the cost could be as low as 18£/Kgal
if a utilization factor of about 1.0 could be achieved. To
achieve such a utilization factor in a municipal waste treatment
plant, with the flow fluctuations characteristic of such plants,
would require storage of the raw waste for at least days and
possibly weeks or months.
87
-------
COST OF AWT BY THE LIME-CLINOPTILOLITE - CARBON PROCESS
San Antonio Sewage 1968 Utilization Factor,
UBAR = 0.5
FCODAC = .8
100
UNIT INVESTMENT
C/gpd of capability
C/gpd vs. capability
00
00
or
UNIT PRODUCTION
CCST
-------
SENSITIVITY OF AWT Costs to Utilization Factor and to
Fraction of COD removed by Accelators
00
30 —
20 —
26
24
22 -
20 —
18 —
Base case
QBARE = 100
UBAR =0.5
FCODAC = .8
fraction of COD removed in Accelators
FCODAC
FCODAC
1.0 UBAR
Figure 16
-------
One case explores the sensitivity of the AWT cost to the
fraction of COD removed in the Accelators. This is not the
overall percentage removal of COD but measures the fraction
of the total COD influent to the Accelators which does not
appear in the exit liquor from the Accelator. As this
fraction is increased the load on the activated carbon is
decreased. Increasing the FCODAC from .8 to .9 would bring
about a reduction of about l.OC/Kgal from the 26.75C/Kgal
cost of the base case at 100 mgd.
Costs for San Antonio 2000
The next page comprises the printout results of the AWT process
for San Antonio in the year 2000 considered as a single central
plant. The costs are 1969, San Antonio. The QDOT capability
is 350 mgd and the average production 177 mgd. The total
investment, excluding land is estimated at 86.0m$. The total
production costs are estimated at 15.3m$ annually, a unit cost
of 23.8
-------
11/05/70. 16.24.34.
PROGRAM AWTFIN
OVERALL AWT PROCESS
QBARE 177.00 QDOT 350.00 UBAR .50571 DISP.^TPD
143.053
INVESTMENT COSTS, K$
PCT. OF
PROCESS
PRIMARY
LIME
CLINOP.
CARBON
CHLOR.
BUILDINGS
DISPOSAL
ENGR.
K$
625.647
20454.389
18260.176
35338.354
630.553
7530.912
0.
3217.865
TOTAL
.73
23.77
21.22
41 .06
• 73
8.75
0.
3.74
CENTS/GPD OF
TOTAL
86057.895 100-00
QDOT
.179
5.844
5.217
10.097
• 180
2.152
0.
• 919
24.588
QBARE
.353
11.556
10.316
19.965
• 356
4.255
0.
1.818
48.620
PRODUCTION COSTS
PROCESS
PRIMARY
LIME
CLINOP.
CARBON
CHLOR•
BUILDINGS
DISPOSAL
ENGR.
OPR. LABOR
TOTAL
KS/YEAR
112.834
3479.671
3916.268
5829.357
348.967
729.566
71 .532
311.734
551.903
15351.832
CENTS/KGAL PCT.
.175
5.386
6.062
9.023
.540
1.129
• 111
.483
.854
.73
22.67
25.51
37.97
2.27
4.75
.47
2.03
3.60
23.763 100.00
COST SUMiMARY, CENTS/KGAL
AMORT. = 12.905* OPER. = 10.858., TOTAL = 23.763
STOP.
CPU SECONDS
11.763
91
-------
DESCRIPTION OF COMPUTER PROGRAM AWTLCC
Program AWTLCC is a subroutine to be used in a larger RECYCLE
Program for exploring reuse systems. Since AWTLCC itself is
a subroutine the listed Program provides a main Program TAWT
which calls AWTLCC itself. AWTLCC operates by calling the
following six process subroutines:
1. PRLM - preliminary treatment
2. LIME - lime clarification
3. CL.INOP - clinoptilolite ammonia exchange
4. ACTCAR - activated carbon adsorption
5. CHLOR - post chlorination
6. DISP - a rudimentary computation of ultimate disposal
cost
It also used the following non-process subroutines:
COSTC - for computing C/Kgal costs, percent contribution
to costs, etc.
UNITS - for computing optimum number of units, used for
Accelators, filters, and clinoptilolite and
carbon trains
PARAB3 - used in the UNITS scheme
BAJO - used in the UNITS scheme
ENGR - a function subprogram for computing engineering
costs
The Program is constructed to operate on a CDC 6400 computer
via a time sharing terminal. The core space required is 8000
words (60 bit words). Program length is 35,840 characters
and requires about 12 seconds CPU to compile and execute.
There are two types of input data, one contained in data
initialization statements in the Program, the other contained
in a data file named SEW.
The data initialization at Line 25 and 26 sets the printout
regime as follows:
= 0 suppresses effluent and influent
concentration
= 0 suppresses diagnostic printout of
UNITS searches
= 0 suppresses capital cost
= 0 suppresses production cost
= 0 suppresses summary printouts for
individual processes
IPRINT (3) is not used
IPRINT (1)
IPRINT (2)
IPRINT (4)
IPRINT (5)
IPRINT (6)
92
-------
Data common to all cost computations is initialized in
lines 117, 118, and 119, these being plant life, tax rate,
insurance rate, interest rate, energy price, current year
Marshall & Stevens Chemical Process Industry Cost Index (MSI),
factor for maintenance, repair and minor replacement, labor
price, payroll extras factor, current year Building Cost
Index for the Region and National BCI for the year of the
MSI.
Data initialization for the LIME subroutine are contained
in lines 1025-1027, in the exemplary runs in accordance with
the list given in Table 11. Data initialization for the clin-
optilolite subroutine is contained in lines 5025-5031 according
to the listing in Table 12. Data initialization for the acti-
vated carbon subroutine is found in lines 822-823 according to
the listing in Table 14.
The data file SEW is called from Line 1025. It contains the
stream matrix parameters SMATX(!,'), each value on a separate
line. The first two I elements are QDOT' and QBARE, the
remaining 18 elements, the concentration parameters for the
raw sewage as listed in Table 15.
Incidentally although the UNITS optimization was developed to
handle the possible characteristics of an unknown cost function,
out of 40 cases run for this report, 10 for each application,
there were 30 cases in which the optimum was the minimum
number of units, 9 cases in which it was one greater than the
minimum,, one case two greater than the minimum, and no cases
where the optimum occurred further away1 than this. Accordingly
the strategy of the optimization was altered to explore first
these positions near the minumum number of units. As a result
the subroutines PARAB3 and BAJO will be very infrequently
called upon.
The Programs trend the cost to 1969 National by means of the
Marshall and Stevens Chemical Process Industries Equipment
Cost Index (Value: 285). It regionalizes by the ENR Building
Cost Index. The denominator in the conversion, the National
BCI must be set for the year of the MSI base. As it stands
the denominator is set at 1969(802) and the current year BCI
is set at Dallas (San Antonio), 732. Cost elements not adjusted
by these indexes are those involving PRFUEL, PRCAR, PRMBTU,
PRLIME, PRLAB, CKWH,' PRCLIN, PRNACL which should be set
corresponding to the region and year desired.
93
-------
PROGRAM AlvTLCC LKh PuOJECT-162 F :•. AACOG AND Fl'QA 11/6/70
11/07/70- 08.50.39.
10 PROGRAM T AWT < INPUT* OUTPUT* TAPED
13 COMMON/UNITS/IX* YY* QKAX* QMIN*QDOTU*J-*NOv.-GO*KMMAX*KKMIN
14 COKiVjiWi\)CALL/NCALLC* JSJCALLU,MCALLF*NCALLB
15 COMMON/BO/CKWH*CYMSI *VTOT* WTOT*FACKfi* IPhIM*DI TIP * VCLIN*VCAK*Clh
16 COMMON/BLK1/CYBCI
20 DIMENSION VTOTC20)*WTOTC20)* IPRIixH ( 6 ) * SMATX<20* 8)
21 DIMENSION IXC4)*YY(4)
25 DATA (CIPRINTCI>*I=1*6)=
26+ 0*0*0*1*1*0)
30 CALL AWTLCCCSMATX)
35 STOP
40 EiMD
100 SUBROUTINE Av.'TLCC (SMATX)
103 COMMON/UNITS/1X* YY,QKAX*QKIN*G)DOTU*F*NOl'/GO*MMMAX,r'iKMIN
104 COMiVOM/i\:CALL/NCALLG*NCALLU*NCALLP* NCALLB
105 COKKON/BO/CKwK*CYKSI*UTOT*V.'TOT*fACKJ:<*IPHIi\iT,DITIF*VCLIN*VCAH,CIR
106 C OKMON/BLK1/CYBCI
107 DIMENSION SKATX(20*8>*IX<4>*YYC4>
108 HIMENSI ON VTOT ( 20 ) * VJTOT C 20 ) * IPF. INT ( 6 ) * PNAME (10)* CNAME< 20 ) * ARRAY ( 7 )
109 DATA PNAKit/lOHPfiELIM. *10HLIME *10HCLINOP *
110-HOHCAHBON ,10HCHLOR. * 10HBUILDINGS *10KDISPOSAL ,
111+10HENGR. *10HOPR. LABOR,10KTOTAL /
113 DATACCCNAKE(I),I=3,19)='fjHNA ,4HK *4HNH4 ,/ihCA * 4HMG *
11/J+ 4HCL *4HF *4HN02 *4HN03 *4HHC03*4HC03 *4HSO^! *4HSI03*
115+ 4HP04 *4HCOD ,4KVSS ,4HNVSS)
116 DATA IFILT/1/
117 DATA PLF*TX*XIN5*RET*CKVK*CYM£I*FACMK*PYEX*FHLAB/
118+20.*.01*.01*.045*1.*£85.0*.04*.45*3./
119 DATA CYBCI*BCINMS/732.*802./
120 DITIF=TX+XINS+RET/(1.0-(1.0+RET)**C-PLF))
121 RTLAB=PhLAB*
-------
PROGRAM AV.TLCC LKR 'PROJECT-162 FOR AACOG AND FlvQA 11/6/70
11/07/70. 06.50.39.
320 VTOT<10)=VTEKP+VTOT=1.12*RTLAB*<49.*QDOT**.315)*365.24
410 V;TOT(10)=0
415 DO 420 1=1,9 '
420 420 V/TOT C 10 ) =V,'TOT (10) +KTOT ( I )
425 AVDSP=DSPTPD*UBAR
500 PRINT 800
505 PFINT 801,QBARE,QDOT,UBAR,AVDSP
506C PRINTC1) CONTAM IN AND OL'T
510 IF(IPRINT<1))600>600,515
515 515 PRINT 802
520 DO 525 1=3,19
525 525 PRINT 803,CNAKE(I),SKATX(I,1),SKATX(I,6)
599C FRINTC4) INVESTMENT COSTS—
600 600 IFCIPRINT(4))700»700,605
605 605 PRINT 804
610 PRINT 805
615 PRINT 806
620 ARRAY C 5)=VTOT <10)/1OO.
625 ARRAY(6)=GDOT* 10.
630 ARRAY<7)=QBARE*10.
635 DO 645 1=1,8
640 GALL COSTCJ=l,4)
650 CALL COSTC(ARRAY,VTOT(10))
655 PRINT 808,PNAMEC10),,0=1,4)
699C PRINTC5) PRODUCTION COSTS
700 700 IFCIFRINT<5»755,755,705
705 705 PRINT 810
710 PRINT 611
715 ARRAYC5)=1000.
720 ARRAY(6)=3G50.+CEAKt:
725 ARRAY(7)=WTOT(10)/100.
730 DO 740 1=1,9
735 CALL COSTC (ARRAY, VTOTCD)
740 740 PRINT 812,PNAKE(I),CAHRAYCJ),J=H,4)
745 CALL COSTC
801C *SEE LINE 830 FOR FORMAT 801+
802 802 FORMAT C//* CONTA1'.. IN OUT*)
803 803 FOR1^TC1X,A4,F9.2,F8.2)
604 804 FOF^J\T
-------
PROGRAM AV.'TLCC LKh HROJLCT-162 1- OR AACOG AND FV.'GA 11/6/70
11/07/70. 08.50.39.
W5 805 FOkiVATC/£4**FCT. OF CLwTS/GFD OK*)
806 806 FOHKATC* FhOCESS* 10X*?($*5X*TOTAL*4X*QEOT*5X*QEAHE*)
807 807 FOi\KAT< IX, A 10,F 10 .3, F7 .S,2F 10.33
808 808 FORKATC/1X,A10.,F1C.3,F7.£,£:F10.3>
810 810 FOHKAK//* PRODUCTION COSTS*)
811 frll FOH.MA.TC/* PF.OCESS*8X*K$/YEAR CEtfTS/KGAL PCT.*>
812 812 FOHi';AT< IX, A10, 2F 1 0 .3, F7 .2 >
813 813 FOEMAT
820 820 FOHKATC//,* COST SUMMARY, CEivTS/KGAL* >
825 825 FORMAT (/., *AKOhT. =*>F7.3**, OPER. =*^F7.3j
826+*, T01AL =*,F7.3*/>
830 80 1 FOIIMATt/* GBAhk;*Ke.2,3X,*t;DOT*F9.a*3X,*UBAh*F7.5#3X,
835+*DISP.*TFE*,F10.3)
899 END
900 SUBROUTINE PRLh(SKATX)
905 COMKOiM/BO/CKv.'H,Cyf.SI,UTOT,VJTOT,FACKK, I PHI NT, DI TIF , VCLIN^CATi, CIR
909 DIMENSION SKAT;; (20, 6)
910 DIMENSION VTOT ( 20 > * WTOT C 90) , IPPINT (6)
915 DO 920 1=1,20
920 920 SKATX^ON/UI\iITS/IX,YY,OKiAX,QMIN,CDOTU,F,i\0;AGO,W'iMA
1004 COkKON/NCALL/NCALLC,NCALLU,NCALLF,WCALLB
1005 COMiyON/BO/CK'.-:H,CYKSI,VTOT,\v'TOT,FACfiH, IPhIMT,DITIF,VCLIN,VCAR,CIIt
1006 COFJ-iON/BLK 1 /CYBC I
1009 DIMENSION 3MATXC20, S) , IX(4),YY (4)
1010 DIMENSION VTOT(SO) ,feTOT( 20 ) , IPfcINT ( 6 ) , AH HAY ( 7)
1015 REAL WVSS12*NVSS13,NVSS14,NVSS15,NVSS16,NVSS17,NVSS18,NVSS19,
1016+NV5S20,NUSSS7,NVSS28,.MVSS10
1025 DATA XfjGhl2,F0412,TSS12,CAC26,CAC27^TSOL13,TSOLlA,TSOLl5,TSOL16,
1026 + TSOL17,FACAOL,FIMKTD,PRLI^E,FCODAC,FCODF,GPi':SFF/
1027+3., S., 10., 1.5E5,35.,1E5, 100. , 2 .5E5, ^ . 7E4,
1028+6. 5E5, 0.9, 0.9, 18. 50, 0.8, 0.1 8, 4. /
1040 G>DOT = SMATX( 1,2)
1045 QBARE=SMATX(2,2>
1047 UBAR=QBARE/QDOT
1050 CAC10=SKATXC6,2)
1055 XhG10=SKATX(7,2)
1060 ALK10=SMATX( 12,2)
1065 P0410=SMATX(16,£)
1070 COD10=S^^TX(17,2>
1075 VSS10=SK1ATX(18,2)
1080 NVSS10=ShATX<19,2)
97
-------
PROGRAM AWTLCC LKR PROJECT-162 FOR AACOG AND FWQA 11/6/70
11/07/70. 08.50.39.
1082 TSS10=NVSS10+USS10
1085 FVSS=0.2*VSS10/CVSS10+NVSS10>
109O FNVSS=0.2-FVSS
1100 Q10=G1l=Q12=Q25=a27=Q29=Q49=QDOT
1102 XMUL=XSCAOK=Q48=0
1103 PCAC14=PCAC16=PINR19=PCAC20=PVSS20=PNVSS20=PMGH20=PAPA20=0.
1105 DOSCAO=-701.706+18.9183*ALK10-0.151669*ALK10**2+6.73099E-4*
1106+ALK10**3-1.67022E-6*ALK10**4+2.17110E-9*ALK10**5-1.14985E-12*
1107+ALK10**6
1109 ICAOFLG=1
1110 IF(DOSCAO-50)1115,1115,1120
1115 1115 DOSCAO=50
1116 ICAOFLG=P.
1120 1120 ITEH1=1
1125 PCAOKR=DOSCAO*QDOT*8.33*1.321
1130 Q19=DOSCA.O*Q10*7.72E-6
1199C **HEEMTRY FKOM 1865 FOR OVERALL LIME STAGE ITERATIONS**
1200C ***PERFOHKANCE OF ACCELATORS***
ILV.'l 120! CONTINUE
liJlO ;-CAC26 = 6.33-i CDOT*CCAC10 + DOSCAO* 100.09/56.08-2.*ALK10)
1211+-8.33*1214,1214
1213 1213 Q26=0 $ ICAOFLG=3
1214 1214 CONTINUE
1215 Q26=PCAC26/CAC26/8.33
1220 PCAC13=PCAC14+FCAC16+FCAC26+8.33*QDOT*ALK10*2-8.33*Q27*CAC27
1225 PMGH13=ODOT*XMG10*8.33*2.399-Q12*XMGH12*8.33+FMGK20
1230 PAPA13=(QDOT*P0410-Q12*P0412)*8.33*1.763+PAPA20
1232 VSS12=TSS12*FUSS/(FVSS+FNVSS>
1233 NVSS12=TSS12-VSS12
1235 F\/SS13=CQDOT*USS10-012*VSS12>*8.33+FVSS20
1237 PNVSS 13= C QDOT*NUSS 10-Q12*NVSS 1 2 )*8 .33+FINP 19+PVSS20*FiWSS/FVSS
1240 TLBS13=PCAC13+PMGK13+FAPA13+PVSS13+PNVSS13
1245 FCAC03=PCAC13/TLBS13
1250 FfcGOH=FKGH13/TLBS13
1255 FAPAT=PAPA13/TLBS13
1260 FVSS=FVSS13/TLBS13
1265 FNUSS=PNUSS13/TLBS13
1270 Q13=TLBS13/TSOL13/8.33
1275 COD27=(1-FCODAC)*COD10
1280C **** PERFORMANCE OF SECOND HECARBONATOR ****
1285 ALH30=CAC27+30.*50./44.
1300C ***THICKENER FERFOHMANCE***
1310 Q14=Q13*CTSOL15-TSOL13)/CTSOL15-TSOL14)
1312 FCAC14=FCAC03*TSOL14*8.33*014
1315 Q15=Q13-Q14
1400C ***CENTRIFUGE PERFORMANCE***
141O Q16=Q15*CTSOL17-TSOL15)/(TSOL17-TSOL16)
1412 PCAC16=FCAC03*TSOL16*8.33*Q16
1415 Q17=Q15-Q16
1500C ***KILN PERFORMANCE***
1501 PCAC17=PCAC13-PCAC14-PCAC16
98
-------
PROGRAM AWTLCC LKR PROJECT-162 FOR AACOG AND FVOA 11/6/70
11/07/70. 08.50.39.
1503 TLBS17 =TLBS13*PCAC17/PCAC13
1505 PMGK17=TLBS17*FMGOH
1510 PAPA17=TLBS17*FAPAT
1515 PVSS17=TLES17*FVSS
1517 PNVSS17=TLBS17*FNVSS
1580 PCA018=PCAC17*0.5603
1525 PKG018=FMGH17*0.6906
1530 PAPA18=PAPA17
1535 PASH=PNVSS17
1540 TLBS18=FCA018+PKG018+PAPA18+PASh
1600C ***SLAKER PERFORMANCE***
1610 ITER2=1
1612 LFLAG=1
1614C **REENTRY FROM 1655 FOR SLAKER ITERATIONS**
1615 1615 FCAOS=(FACAOL*XMUL+PCAO18 ) / (FACAOL*XMUL+TLBS18)
1630 FCAOHD=FCAOS**10
1622 PINRTS=TLBS18-PCA018+XMUL*(1-FACAOL)
1625 TLBS31=FINRTD*PINRTS/<1-FCAOHD>
1630 PCAOHD=FCAOHD*TLBS31
1632 GO TOC1635*1656>,LFLAG
1635 1635 XMULN=(DOSCAO*QDOT*8.33-(PCA018-PCAOHD*0.7568))/FACAOL
1636 IF(XMJLN) 1637, 1640, 1640
1637 1637 LFLAG=2
1638 X14UL=XMULN=0
1639 GO TO 1615
1640 1640 IFCABSf (XKULN-XMUD/XMULN.-0.01 >166G, 1660* 1645
1645 1645 XMUL=XMULN
1650 ITER2=ITER2+1
1655 GO TO 1615
1656 1656 XSCAOH=PCAO18*1.321-PCAOKD-FCAOHR
1660 1660 CONTINUE
1665 PINR47=(1-FINRTD)*PINRTS
1670 PCAOH47=PCA018*1.321+<1-FACAOL)*XMUL*1.321-PCAOHD
1675 TLBS47=PCAOH47+FINR47
1680 047*=TLBS47*9/8.33E6
1685 Q19=Q47*(1-XSCAOH/PCAOH47>
1690 PNVSS19=PINR47*(1-XSCAOH/PCAOR47)
1691 PINR19=PIWSS19
1695 XMUL=XMULN
1700C ***FILTER PERFORMANCE AND SCRUBBER PERFORMANCE***
1705 Q12=Q11-Q13
1708 Q27=Q12-Q26
1709 Q41=Q30=0
1710 COD28=COD27
1711 P0428=P0412
1712 VSS27=VSS28=TSS12*FVSS/CFVSS-«-FNVSS)
1713 NVSS27=NVSS28=TSS12*FNVSS/(FVSS+FNVSS)
1714 IF(IFILT)1720*1720*1715
1715 1715 P0428=.67*P0412
1716 1716 Q41=Q30=0.03*Q27
1717 COD28=COD27*(1-FCODF)
1718 VSS28=.1*VSS27
99
-------
PP.OGRAM AWTL
PROGRAM AWTLCC LKR PROJECT-162 FOR AACOG AND FV'GA 11/6/70
11/07/70. 08.50.39.
1719 NVSS28=.1*NVSS27
1720 1720 FVSS30=(VSS27-VSS28)*8.33*027
1725 PNVSS30==VSS28
1960 SMATX<19*3)=NVSS28
2000C ACCELATOR SIZING
2005 2005 CONTINUE
2006 Ql"iAX=20• 16 $QMIN=0.7 $QDOTU=Q11 $ F=.65
2008 NCALLU=N5=NCALLC=V/ACCEL=0
2010 2010 CALL UNITSCNTRAIN,QTRAIN,VACCEL,N5)
2011 IFCNOWGO.EQ.l) GO TO 2050
2013C COMPUTATION RE-ENTRY,ACCEL
2014 2014 X=ALOGCQTRAIN) $NCALLC=NCALLC+1
2016 VACCEL=NTRAIN*2*EXP(10.5803+0.356282*X+0.114333*X**2+0.0374586*
2017+X**3-0.00403479*X**4-0.003669 15*X**5)/1000.*CIR*273.1/294.2
2018 VPUKF=NTRAIN*2.4*(7.5771E-4*(20*QTRAIN)**0.740863+0.433922)*
2019+CIR
2020 VACCEL=vACCEL+VPUMPSIF(IPRINT<2))2025,2025,2021
2021 2021 PRINT 2022,NCALLC,NTRAIN,QTRAIN,VACCEL
2022 2022 FORKATC2X,I2,2X*NTRAIN: *I4,2X*QTRAIN: *
2023+F10.3,2X*VACCEL: *F13.3)
2024C ACCELATOR OTFIMIZATION
2025 2025 IF(N5)2032,2032,2026
2026 2026 M5=0 £ YY(2)=VACCEL
2028 2028 CALL PARAB3CNTRAIN,GTRAIN)
100
-------
PROGRAM Av.'TLCC LKR PROJECT-162 FOR AACOG AND FWQA 11/6/70
11/07/70. 08.50.39.
2030 GO TO 2014
2032 2032 IF(NOkGO>2034,2050,2050
2034 2034 IFCNCALLU-7)2010,2036,2040
2036 2036 YYC2)=UACCEL S GO TO 2014
2040 2040 IXC4)=NTRAIN £YY<4D=VACCEL
2042 CALL BAJOCNTRAIN,GTRAIN,J>
2044 IFCNOVJGO>2046.t2014,2050
2046 2046 GO TO C2014,2028,2028,2014,2028>J
2050 2050 CONTINUE
2055 RLTFD=PCA018/2000.
2060 KLPMGD=RLTPD/ODOT
2062 TSPMGD=TLBS18/2000/GDOT
2063 RECOV=PCA018/(QDOT*DOSCAO*8.33>
2065 VRCALC=178.959*CTLBS18/2000)**0.535022*CIR
2070 RCAHEA=Q12*25
2071 VCONC=100./27.*< 40*(1+ SQRT CRCAREA)) + < 2 + SGRTC RCAREA))**2)
2072 VGRIDS=RCAREA*6.
2073 VCARBU=40500*CQ12/50)**0.7
2074 IF(Q12-50.)2075,2077,2077
2075 2075 X=ALOG
2076 VCARBU=1.35*EXFC8.35461+. 178289*X>.0350484*X**2+.0122392*X**3
2077 2077 VRCARB = 2*
-------
PROGRAM AWTLCC LKR PROJECT-162 FOR AACOG AND FWQA 11/6/70
11/07/70. 08.50.39.
2190 IFCNOWGO) 2194,2145*2196
2194 2194 GO TO (2145,2165,2165,2145,2165) J
2196 2196 CONTINUE
2199 2199 VTOTC2)=UACCEL+VRCALC+VRCARB+UFILT
2200C ****OPERATING COSTS***
2201C NOTE: LABOR COSTS ARE SET TO ZERO
2220 tfCCHG=DITIF*VTOT(2)*1000
2225 WMUL=PRLIME=XMUL/2000*365.24*UBAR
2230 WFUEL=365.24*11.2492*(TLBS18/2000*UBAR)**0.767663*CYBCI/754.
2235 COMPHP=1.30789*Q12**0.668189-0.45166
2240 POWACO1.0*011*0.7452/0.75
2242 POWPMP=911*20./3957./0.75*0.7452
2244 POWCOMF=.47757*COMPHP
2245 POWFILT=1.4E5*Q41*IFILT
2246 POWKILN=0.792*TLBS18/2000
2248 WPOU=((POV'ACC+(POWCOMP+POWPMP)*UBAR)*24*365.24
2249++POWFILT*UBAR)*CKWH/l00+365.24*POVKILN
2250 WMRR=1000*FACMR*VTOT<2)
2260 WTOT*2)=WCCHG+WMUL+WFUEL+WPOV/+WOL+WMRR
3000C ***PRINT STATEMENTS***
3001 IF* IPRINT*6»3500,3500,3005
3005 3005 PRINT 4904
3006 GO TO (3009,3007,3008),ICAOFLG
3007 3007 PRINT 4900 $ GO TO 3009
3008 3008 PRINT 4990
3009 3009 CONTINUE
3015 PRINT 4906,QBARE,QDOT,UBAR
3017 PRINT 4907,CAC10,XMG10,P0410,ALK10,TSS10,COD10
3400 3400 PRINT 4950
3405 AMOET=WCCHG/3650/Q3ARE
3410 WOP=(WTOTC2)-WCCKG)/3650/QBARE
3415 WTOTAL=AMORT+WOP
3425 PRINT 4951,AMORT,ttOP,WTOTAL
3500 3500 RETURN
4900 4900 FORMATC/ *CAO DOSE ARBITRARILY SET AT 50 MG/L*/)
4902 4902 FORMAT*15)
4905 4904 FORMAT*////,* LIME PROCESS *>
4907 4905 FORMAT*//, * IMPOSED CONSTANT DESIGN PARAMETERS*)
4908 4906 FORMAT*/* QBAHE*K11,3,2X*QDOT*F12.3,2X,*UBAR*F12.5)
4909 4907 FORMAT**CAC03*F6.1*, MG*F5-1*, P04*F5.1*, ALK*,
4910+F6.1*, TSS10*F6.1*, COD*F6.1)
4956 4950 FORI^AT*//,* COST SUMMARY, CENTS/KGAL*)
4957 4951 FORMAT*/,* AMORT. =*,F7.3,*, OPER. =*,F7.3,*, TOTAL =*,
4990*4990*FORMAT*/* Q26 NEGATIVE. ARBITRARILY SET TO ZERO.*/)
4999 END
5000 SUBROUTINE CLINOP* SMATX,XSCAOK) lin, TM tir/m PTR
5005 COMMON/BO/CKVK,CYMSI,VTOT,ViTOT,FACMR,IPRINT,DITIF,VCLIN,UCAR,CIR
5007 COMMOM/UNITS/IX,YY,Q^-X,QKIN,QDOT,F,NOViGO,MMMAX,MMMIN
5008 COMMONAMCALL/NCALLC,NCALLU,NCALLF,NCALLB
5010 DIMENSION SMATX<20,8)
5015 DIMENSION VTOT*20),ViTOT( 20 ) , IFRINT* 6 ),AnRAY( 7 )
102
-------
PROGRAM AWTLCC LKri PROJECT-162 FOk AACOG AIvD FlvQA 11/6/70
11/07/70. 06.50-39.
5020 DIMENSION IX(4)>YY(4>
5025 DATA AN79*CLLF,F-CLPR,GPMSFX,GPKSFS*C$'APG*DIAfcAX* DEFhAX*
5026+FRCLIN*Fhi\iACL*FRLIK-E*FACAOL*EPUfcF>EELOti,NETL77*AVTL77*PRM3TU
5030 + /0.5*0.17*0.0026.,6.*5.*300.*25.5* 10.,
5031 +10.*0.01* 18.50*0 .9*0.75*0.60* 100*67 «*0.5/
5033 NCALLU=NCALLC=N5=VTOT<3 ) =0
5035 Q50=Q79=QDOT=QDOTU = SI-iATXC1,3)S>¥-1 .
5037 GBAHE=SKATXC2*3>
5040 UBAR=QBARE/QDOT
5045 AN50=ShATXC5*3>
5100C EXCHANGER NUMBER AND SIZE
5105 BEDDLP=DEPMAX
5110 QMAX»GPMSFX*0.00144*0.785398*DIAMAX**2
5112 QhIN=GPMSFX*.00144*.765398
5115 5115 CALL UNITSCNEXCKL,QEXCK,WTOT<3)*N5>
5116 IFCNOV.'GO.EQ.l )GO TO 6775
5119C COMPUTATION ENTRY POINT
5120 5120 AREA=GEXCH/.00144/GPKSFX
5121 NCALLC=NCALLC+1
5125 DIAK=SQRTCAREA/O.785398)
5130 NHFT=2*DIAW+0.99999
5135 DIAM=0.5*NHFT
5140 AREA=0.785398*DIAM**2
5145 BVSTG=BEDDEP*AREA
5150 GPhSFC=QDOT/(NEXCHL*0.00144*AREA)
5155 TLOAD=BEDDEP*CLLl-/CGPKSi1C*(AN50-AN79)*499.8)*lE6
5160 NEXCKR=NEXCHL*9./TLOAD+0.99999
5165 NEXCHT=NEXCHL+NE>;CHR
5170 HPXPMP=QDOT/2/0.00144*35/3957/EFUKP
5175 PCAOH=QDOT*8.33*(AN50-AK795/14.01*74.10/2
5180 PLIME=PCAOH*56.08/74.10/FACAOL
5200C REGENERATION
5205 NRGS=0.5*NEXCHR+0.9
5210 CHTANK=4*BUSTG*7.4806
5S15 CDTANK=1.5*BVSTG*7.4806
5220 GPMREG=1.25*BVSTG
5225 ANXRGD=NEXCKL*24/TLOAD
5227 PSALT=ANXRGD*BUSTG*.18244
5230 HPRPMF=GPMREG*35/3957/EPUMF
5300C STRIPPER
5305 GPMSTR=AfcXRGB*BVSTG/48
5310 TAREA=GPi\STR/GPKSFS
5315 CFKAIR=CFAPG*GPMSTR
5316 BTUAIR=CF!»iAIR*29./392.*1440*NDTL77*0.23e*(77-AVTL77)
5317 BTULIQ=GFMSTR*8•33*1440*NDTL77* C 77-AUTL77)
5320 HPSPKP=GPMSTR*35/3957/EPUMF
5325 NBLOW=1
5330 HPBLOW=0.000236*CF^AIR/EBLOW
5335 IF
-------
PROGRAM AWTLCC LKH PROJECT-162 FOR AACUG AND FwOA 11/6/70
11/07/70. 08.50.39.
6000C CAPITAL COSTS
6005 VEXCH=NEXChT*l.56611*BVSTG**0.504425*CIh
6010 VCLIN=FRCLIN*NEXCHT*BVSTG/1000
6015 VXPUKP=2*2.4*(7.57718E-4*(35*QDOT/2/0.00144)**C.740863+0.433922)
6016+*CIR
6020 VHTANK=3*NRGS*1.23*CO.00506131*CHTANK**0.688148+0.139225)
6021+*CIR*273.1/241.8
6025 VDTANh=NRGS*l.23*CO.00506131*CDTANK**0.688146+0.139225)
6026+*CIR*273.1/241.8
6030 VRPUKP=3*NKGS*2.4*(7.57718E-4*<35*GFMREG)**0.740863+0.433922)
6031+*CIR
6035 VTOWER=2*(0.0890*TAREA+0. 7 19873 ) *CIF:*273. 1/294.2
6040 VSPUKP=2*2«4*<7.57718E-4*(35*GPMSTR)**0.740863+0.433922)
6041+*CIh
6045 VBLOW=NBLOW*2*CO .303308*KFBLO\v**0.698 176 + 0.37941 9)*CIR*G73.1
6046+/108.7
6050 VSPIFE=0.42*VTOWER+0.18*VBLOW
6055 VTOTC3)
6056+VSFUMP+VBLOW+VSPIPE
6500C OPERATING COSTS
6501C ******NOTE:LABOR COSTS SET TO ZERO********
6510 WCCHG=DITIF*VTOT<3)*1000
6515 WKUC=FCLPR*ANXRGD*365.25*BVSTG*PRCLIfo*UBAR
6520 WLIME=PLIKE/2000*365.24*PRLIME*UBAR
6521 IFiSCAOH)/PCAOH
6523 IFCWLIME)6524,6524,6525
6524 6524 WLIME=0
6525 6525 WSALT=PSALT*365.25*PRNACL*UBAR
6530 WFUEL=CBTUAIR+BTULie>*l.E-6*PRMBTU
6535 KPQBAR=(2*HPXPKP+NRGS*3*HFRPhP/UBAh+2*HPSPMP/UBAR+UBAR*HPBLOv)
6536+*UBAR**2
6540 WFOW=0.7452*HPQBAR*24*365.25*CKWH/100
6550 V,'MRR=1000*FACMR*(VTOT(3)-VCLIN)
6551 WTOTO)= WCCHG+WKUC+WLINE+WSALT+VFUEL+WFOV+WMRR
6555 IF (IPRINT(2)) 6705>6705*6556
6556 6556 PRINT 6557,NCALLC,NEXCHL,QEXCH,WTOT<3)
6557 6557 KOKMATC2X,I2,2X*NEXCHL: *I4>2X*QEXCH: *F10.3,2X*WTOT(3): *
6558+F13.3)
67 OOC OPT I MI ZATI ON
6705 6705 IF(N5)6730*6730,6710
6710 6710 N5=0
6715 YYC2)=WTOT(3)
6720 6720 CALL FARAB3
104
-------
PROGRAM AtoTLCC LKIi FROJECT-169 FOB AACOG AND FtQA 11/6/70
11/07/70. 08.50.39.
6765 IFCNOv.iGfj)6770*51 20*6775
6770 6770 GO 10(51 20, 67 SO* 6720* 5 1 20* 6720 > J
6775 6775 CONTINUE
6777 PRINT 6778
6776 6776 FORMAT <* ---------- **//)
6800C **EFFLUENT STREAM***
6805 6805 DO 541O 1=1,20
6810 5410 SMATXCI*4>=SMATX(I*3>
6815 SMATXC3*4)=SMATX<3*3)+PSALT*23.00/58.45/(GDOT*8.33)
6820C **i\OTE+ ONLY NH3 ASSUWfc.D EXCHANGED**
682 5 SMATX (5*4) =AN7 9
6830 SMATyC6,4)=SMATX(6*3)+FCAOH*56.0fc/74.10/CQDOT*g.33>
6835 SKATXC&*4)=SMATX<8*3)+PSALT*35.45/58.45/
7000C ---- PRINT STATEMENT
7005 IF(IPRINT<6)) 7799*7799*7010
7010 7010 PRINT 79O1
7020 PRINT 7903*QBARE,ODOT*UBAK*AN50
7600 7600 PRINT 7970
7605 AfcORT=\vCCHG/3650/QBARE
76 1 0 WOP= ( WTOT C 3 5 -VCCHG ) /3650 /OBARE
7615 WTOTAL=WOF+ArtOET
7620 PRINT 7971*AMOR1*v,'OP*l'.'TOTAL
7799 7799 RETURRi
7800C ---- FORMAT STATEMENTS
7801 7901 FORMAT C ////** ----- CLIK'OFTILOLITE PROCESS ----- *>
7805 7903 FORMAT (/** GBARE*F6 . 2* 3X* *QDOT**t 9 . 2*3X* *UBAR**
7806 + F9 . 5* 3>! * *AN50** F9 . £ )
"7696 7970 FORMAT * IX(4),YY(4)
8020 DIMENSION VTOT ( 20 ) *WTOT( 20 )* I PR I NT (6 ) * ARRAY C 7 )
8022 DATA NSTAGE,GPMSF*RHO*PRCAF,PRFUEL*GPMBk*COD89* CT* CLF/
8023+ 2,6.*30.*26.*25.*8.*8.*50.*.5/
8025 QDOT=QBOTU=SMATX<1*4> $ F= 1 .
8027 QBARE=SMATXC2*4>
8030 UBAR=QBARE/GDOT
8032 COD80=SMATX(17*4)
8035 IF
-------
PROGRAM AWTLCC LKR PROJECT-162 FOK AACOG AND FVQA 11/6/70
11/07/70. 10.34.19.
8105 8105 BEDDEP=GPI*SFC*CT/7.4806/NSTAGE
8110 IFCBEDDEP-25.)8125,8135,8115
8115 8115 GP1*;SFC=25.*7.4806*NSTAGE/CT
8120 BEDDEP=25.
8125 8125 IF**0.5
8300 8300 CONTINUE
8315 8315 NHFT=2*DIAK+0.99999
8320 DIAM=0.5*NHFT
8325 AREA=0.785398*DIAK**2
8327 BEDDEP=BVSTG/AREA
8330 NHFT=2.6*BEDDEP+0.99999
8335 HGT=0.5*NHFT
8337 GPMSFC=QTRAIN/0.00144/AREA
8340 NSPARE=2+2*(NTRAIN/10)
8345 NADSO=NTRAIN*NSTAGE
8350 NADST=NADSO+NSPARE
8360 VOL=AREA*HGT
8400 FAREA=.000640472*CREGDY**1.29912
8405 IFCFAREA-28.26)8410,8415,8415
8410 8410 FAREA=28.26
8415 8415 CARB=(NADSO+NSFARE/2)*BVSTG*HKO
8427 ADSFR=QDOT/NTRAIN/.0014^
8430 CRGDYE=UBAR*CKEGDY
8500C ***********INVESTMENT COSTS***********
8505 VADS=NADST*0.187254*VOL**0.593364*CIR
8510 VFUR=1.3*19.0042*FAREA**0.39326*CIR
8514 VPASBW=29.3693*QDOT**.710743*CIK
8525 VPIP=(NADSO+NSPARE/2)*2.41222*ADSFR**.366757*CIR
8530 VELIN=(46.4174*QDOT**.386189)*(.70455+NSTAGE*.147708)*
853H-CIR
8540 VCAR=PRCAR*CARB/1E5
8545 VCON=.048*VPASBU+.092*VADS+.176*VFUR
8560 VTOT(4)=VADS+VFUR+UFASBW+UPIP+\/ELIN+v;CAK+VCON
8565C **********OPERATING COSTS***************
8567 WCCHG=DITIF*VTOT<4)*1000
8570 WCAR=CHGDYE*PRCAR*3.6524*.05
8571 WFUEL=0.0155125*CRGDYE*PHFUEL
8572 UOL-0
8573 WKRR=1OOO.*FACMR*TOT
-------
PROGRAM AWTLCC LhH FROJLCT-162 FOR AACOG AND FlvOA 11/6/70
11/07/70. fO.34.19.
8582 6582 FORMAT< 2X* I2ȣX*iS)THAIN: *I4*2X*GTRAIK: *F10.3*2X
P5B3+*WTOT<4)• *F12.3)
S584C ACTCAH OPTIMIZATION
8585 8565 IFUM5>6589* 6589, 8566
8586 8586 N5=0 Z YYC2>=WTOTC4>
85P7 6587 CALL PAHAB3 (iMTliAIN*STRAIN)
8568 GO TO 8145
8589 8589 IF(NOUGO) 8590*8595*8595
8590 8590 IF(ivCALLU-7 ) 81 40*859 1 ..8592
8591 8591 YY(2>=WTOTC3> $ GO TO 8145
8592 8592 IX(4)=(v'TRAIA> $YY<4>=WTOTC4) SCALL BA JOCNTHAIW, QTHAIiv, J J
8593 IFCNOwGO) 859^,8145*8595
8594 8594 GO TOC8145*8587*8587,8145*8587>J
8595 8595 CONTINUE
8600 8600 IF /3650 /QBARF.
8875 WTOTAL=wOP+AfcOFiT
6880 PRINT 71>AMOFiT*V?OF»WTOTAL
8899 8899 CONTINUE
8900 1 FOKMAT(///*1X,* ACTIVATED CAl.'BON PROCESS *)
6902 3 FORMAT(/*1X**QDOT =**F11.3>* toGD, GBAEE =**F11.3*
8903+* ^GD* UBAR =**F6.4)
6904 4 FORMAT<1X*CLF =*F5.3*+* CT =*F5.1** CODEF =*F6.4,t+,
8905+4X*CODIN =*F5-1>
8979 50 FORMAT(* FLO\V RATE OVER 13 GPK/SQ.FT. CHECK PFtfcSS. DliOP.*)
8980 70 FOKMAT
8990C*******EFFLUENT STBEAi'i******
8991 DO 8460 J=l*20
8992 8460 SMATXCJ,5>=SWATX=SKATX<16*4)*0.67
8994 S«ATX<17,5>=COD69
8998 RETUKiN)
6999 END
9000 SUBROUTINE CKLOKCSi-iATX )
9005 COMi\ON/BO/CKWH,CYKSIjVTOT*I*'TOI*FACi'-Ji*IF
9006 COMflON/BLKl/CYBCI
9009 DIMENSION SKATX(20,8)
9010 DIMENSION UTOT(20 ),WTOT(20), IPfiINTC6)
9015 DO 90SO 1=1*20
9020 9020 SMATXCI*6) = SFjATX(I*55
9025 QDOT=SKATXC1*5>
9030 QBARE=SMATXC2*5>
9034 VTOTC5)=13.5*QDOT**.658*CIR*273.1/262.9
9040 WCCHG=UTOTC5)*DITIF
9045 WOM=3178.*QBARE**.9O4*CYBC1/718.
9050 WTOT<5)=WCCHG+WOM
107
-------
6 4
PROGRAM AWTLCC LKR PhOJECT-162 FOR AACOG AND FVQA 11/6/70
9055 SMATX (8,6) =SKATX( 8, 5 ) +8 .
9056C ABOVE ASSUMES ALL CL2 CONVERTED TO CL-.
9060 RETURN
9065 END
9200 SUBROUTINE DISP(VTOT,WTOT,DITIF,DSPTPD,UBAR,CIR,CYMSI)
9205 DIMENSION VTOTC20),WTOT(20)
9210 VTOT(7)=0
9215 *TOT(7)=1 .5*365.24*UBAR*DSPTPD*CIR*273.1/CYMSI+VTOT(7)*DITIF*1000
92 SC RETURN
9225 END
10000 SUBROUTINE COSTCCARRAY,VALUE)
10005 DIMENSION ARRAY(7)
10010 ARRAY(1)=VALUE
10015 DO 10020 1=2,4
10020 10080 ARRAY(I)=ARRAY(l)/ARRAY(I+3)
10C£5 RETURN
10030 END
12000 FUNCTION ENGR(VALUE)
12005 ENGR=0.154874*VALUE**(-0.122012)
12010 IF(ENGR.LT.0.05)ENGR=0.05
1210C RETURN
12105 END
13000 SUBROUTINE UNITS(MBACK,QBACK,w,N5)
13001C OPTIMUM NUMBER OF UNITS AND SIZES THEREOF TO PRODUCE QDOTU
130C2C OR F*QDOTU WITH ONE UNIT OUT. FIRST CALL GIVES BOUNDARIES
13003C OF THE DOMAIN OF M AND Q, AND RETURNS FIRST M AND Q TO USE.
13004C SUCCESSIVE CALLS EXPLORE DOMAIN TO LOCATE OPTIMUM.
13005 COKMON/BO/CKWH,CYMSI,VTOT,WTOT,FACKR, IPRINT,DITIF,VCLIN,VCAR, CIR
13006 DIMENSION IFRINTC6)
13014 DIMENSION IX(4),YY(4)
13015 COMMON/UNITS/IX,YY,QMAX,QMIN,QDOTU,F,NOVGO,MMMAX,MMMIN
13016 COMMON/NCALL/NCALLC,NCALLU,NCALLP,NCALLB
13020 M(X)=MAXO((INT(F*QDOTU/X+2.-l.E-7))*INTCQDOTU/X+l.-1 .E-7))
13030 Q(K)= AMAXKF*QDOTU/CK-l),eDOTU/K)
13040 NCALLU=NCALLU+1
13050 GO TO (60,200,220,220,250,300,400)NCALLU
13060 60 MM=K(QiyAX)
13065 NOWGO=-I SV;LAST=I .El?
130.70 IF (MK-2) 80,90,90
13080 80 MM=2
13090 90 QQ=Q(M>i)
13100 IF (QQ-QMIN) 110,120,120
13110 110 CQ=O^IN
13120 120 fJEACK=K-MKIN=MM
13130 QBACK=QQi\AX=QG
13140 140 QQMIN = QMIN
13150 MMMAX = M(OMIN)
13151 IF (IPRINT(2)) 170,170,160
13152 IF(MMMIN-MMMAX>160,154,160
13IS/; 154 NOWGO=0
13158 158 FORMAT(//,* *)
13159 160 PRINT 158
13151
13155 160 IF IPRINT (2) 170, 170, 159
13159 159 PRINT 158
108
-------
PROGRAM AWTLCU LKR PROJECT-162 FOR AACOG AND FWQA 11/6/70
11/07/70. 10.34.19.
13160
13161
13162
13163
13170
13200
13220
13250
13255
13260
13265
13870
13300
13400
13406
13410
13450
13451
13452
13900
13998
13999
14000
14010C
14011C
14014
14015
14016
14017
14019
14020
14040
14045
14046
14047
14048
14049
14050
14051
14052
14059
14060
14062
14063
14064
14070
14080
14090
14100
14110
14120
14126
QfcAX
UMIN*)
PRINT 161
161 FORMAT C* UNITS DOMAIN: * * 4X* *MMIN MMAX
PRINT 163* MMMIN,MMMAX*QQMAX>QQMIN
163 FORMATC20X*I3*I6*F10.3*F9.3)
170 RETURN
200 WLAST=W$fcBACK=MBACK+l$GO TO 900
220 IFCW-WLAST>200*255*255
250 IFCW-wLAST>260* 255*255
255 MBACK=MBACK-1$NOVGO=0$GO TO 900
260 IFCMBACK-KMMAX>270*265*265
265 NO«GO=1 $ GO TO 998
270 YYC1 >=W$IXC1)=MBACK$MBACK=MMMAX$GO TO 900
300 WLAST=W$MBACK=MBACK-1$GO TO 900
400 IFCW-WLAST)450*406*406
406 IFCYYCn-W) 450*450,410
410 MBACK=MMMAXSNOWGO=0$GO TO 900
450 IX ( 3 ) =MBACK$YY< 3 ) =fc$NWAY=MKMAX/ 10+1
IXC2)=MBACK=CNWAY*IXC 1 >+IX(3) )/CNWAY*l)
N5=1SNCALLU=8$GO TO 900
900 QBACK=QCMBACK>
998 RETURN
END
SUBROUTINE FARAB3CMBACK,QBACK>
GIVES PARAB THROUGH 3 POINTS AND X AT MINIMUM* OR A CANDIDATE
X CONSTRAINED
DIMENSION IX(4)*YY(4)
C OMMON /UN I TS / 1 X* YY* QMAX* QMI N* QDOTU* F * NOW GO * MMMAX * MMM I N
COMMON /NCALL/NCALLC*NCALLU*NCALLP*NCALLB
QCK) = AMAXKF*QDOTU/(K-1 )*QDOTU/K)
J=0 S NCALLP=NCALLP+1 ,
C1 = IX(1 )-IX(2) S C2 = IX(2>-IX(3>
C=((YY(1)-YY(2))/C1-(YY(2)-YY(3J)/C2)/(IXC1 )-IX(3»
IF (C) 46*46*50
46 PRINT**APPROXIMATING PARAB HAS A MAXIMUM. NCALLP=**NCALLP
IFCYY(3)-YY(1 ) ) 49*48*48
48 IXLOW=IXC1) S MUSE=MMMIN-1 S GO TO 150
49 IXLOW=IX(3) $ MUSE=MMMAX+1
150 IX(45=CIXLOV'+IXC2) )/2 S IF ( IXC 4) - IXC 2) ) 151*152*151
151 IFC IXC4)-IXLOW> 75*152*75
152 IXC4)=CIXLOW+MUSE)/2 SGO TO 75
50 B=(YYC1)-YYC2))/C1-CIXC1) + IXC2»*C
A = YYC1) -B*IXC1) - C*CIXC1)**8)
XP= -B/C2*C> $ YP = A+B*XP+C*XP**2
PRINT 64* XP*YP
64 FORl«iAT(*PARAB OPT X =**F10.3** WIN Y =**F10.3)
IXC4)= XP
75 IFCIXC4)-KMMAX)73*79*72
IXC4)=MMMAX$OBFLAG=-1 SGO TO 78
IFC IXC 4) -MMMIN 574*79*79
IXC4)=MMMIN $ OBFLAG=1 S GO TO 78
PRINT, *IXC4> RETURNED FROM OUT OF BOUNDS*
S GO TO 4140
72
73
74
78
J=0
109
-------
PROGRAM AWTLCC LKR PROJECT-162 FOR AACOG AND FWQA 11/6/70
11/07/70. 10.34.19.
14130 79 J=0 $ OBFLAG=0
14140 4140 DO 88 1=1,3
14150 IF(IX<4)-IXfI)) 88*85*88
14160 85 J=I $ PRINT 86,NCALLF,I,IX(4)
14170 86 FORMAT(*IX<4> RETURNED FROM PARAB3 AT NCALL *, 13,
14180+ * SAKE AS IX<*,I2,*>; = *,I3>
14190 88 CONTINUE
14200C NEW X(4J FOR DUPLICATES
14210 IFCJ) 194*194,191
14220 191 GO TO <4221,200,4221)J
14221 4221 IFCOBFLAGJ 4222,4223,4222
14222 4222 IX(4)=IX(2)+OBFLAG $ GO TO 90
14223 4223 GO TO <192,200,193 )J
14830 192 IX4DUP=0 $ GO TO 48
14240 193 IX4DUP=0 $ GO TO 49
14250 194 IX4DUF=0 J GO TO 90
14260 200 IF(YYC3)-YYC1» 220,210,210
14270 210 IXLOW=IX<1> $ IXKI=IX(3) $ GO TO 240
14280 220 IXLOW=IXC3> S IXHI=IXC1>
14290 240 IFUX(4)-IX4DUP>250,280,250
14300 250 IFCIX(2)-IXLOU-1>260,270,260
14310 260 IX4DUP = IX(4>$IX<4) = CIXLOW + IX(2»/2$GO TO 90
14320 270 IX4DUP = IX(4)$IX(4) = (IXHI + IX(2»/2SGO TO 90
14330 280 IF(IX(2)-IXHI+1)300,290,300
14340 290 IX<4)=IX<2)-1SGO TO 90
14350 300 IX<4)==IX<2)+1
14360 90 MBACKt=IX(4) S QBACK=Q(MBACK) $ YY(4)«0
14370 RETURN
14380 END
15000 SUBROUTINE BAJOCKBACK,QBACK,J)
150IOC OUT OF 4 POINTS RETURNS 3 LOWEST BRACKETING
1501ic A MIMIMUM; OR IF NONE, 3 LOWEST
1501SC AND CANDIDATE FOR A MINIMUM BRACKETER
15014 DIMENSION IX(4),YY(4)
15015 COMMON/UNITS/IX,YY,QKAX,QMIN,QDOTU,F,NOfcGO,MMMAX,MMMIN
15016 COMMON/NCALL/NCALLC,NCALLU,NCALLP,NCALLB
15019 QCK) = AMAX1(F*QDOTU/(K-1),QDOTU/K)
15020 YMIN=1E17 S J=0 S N=4
15030 NCALLB=NCALLB+1
15035 IXViAS4=IX(4)
15071 DO 95 K=l,3
15072 N=N-1
15075 DO 95 1=1,N
150BO IF(IX-IXCH-1)> 95,95,85
15085 65 XTEMP=IX(I) S YTEMP=YY
-------
Fl AVTLCC LHh PROJECT-162 FOR AACOG AND F'.VQA 11/6/70
11/07/70. 10.34.19.
15136 IF CJ-2) 140*137*140
15137 137 I"KM I N= I X C 1 ) S MiXl/AX = I X ( 3 )
15136 PRINT 139*tfKfcIN*fcl*WA.X
15139 139 KOKKAT ( *ivJ.MX, I i\i = * * 1 4* 2X * *ivji*jix;AX = * * 1 4
15140 140 GO TO (190*320*160*160) J
15150C ---- 160 AKRANGE ---
15160 160 XT£yP=IX(l) S YTENP=YY< 1 >
15170 DO 175 1=1*3
15174 IX(I)=IX( 1+1 )
15175 175 YY(I)=YY(I+1)
151 80 I.X(4)=XTEM?£ YY ( 4 > =YTEMP
15190 190 GO TO ( 1 95* 320* 320* 200 ) I to IN
151 94C ---- 190 NEW XC4) ---
15195
15200
15205
15S06
15207
195
200
205
205
X ) )/2
N £ GO TO
KUSE=MKKAX
IXC4) =(KUSE+ I/ ( IM
S GBACK=Q<
PRINT* +TKREE LO'wEST NOT
15315C --- TEXT FOR X CONVERGENCE ---
15320 320 GO TO (399*330*330*399) J
IF<1-CIXC3)-IX<2))*CIX(2)-IX< 1 ) ) )360* 340* 360
IFCIX(2)-IXWAS4)350*345*350
iW*'GO=lSGO TO 400
NOWGO=0 £ i*iBACh=IXC4)=I>.<2) S GO TO 399
J=5
QBACK=Q (MBACK )
RETURN
15330
15340
15345
15350
15360
15399
15400
15410
330
340
345
350
360
399
400
END
---THE
E N D - - -
111
-------
VARIABLE NAMES
Variable Names For The General Program
AMORT Amortization cost, #/Kgal
ARRAY Array used for cost calculations
BCINMS National BCI for Year of CYMSI
CKWH Cents per kilowatt hour, £/kwh
CCDnn COD in stream nn, mg/1
CYBCI Current year Building Cost Index for Region
CYMSI Current Year Marshall & Stevens Chemical Process Industries
Equipment Cost Index
DITIF Amortization rate, depreciation, interest, taxes, insurance
annual fraction of investment
ENGR Engineering, fraction of subtotal investment
EPUMP Efficiency of pumps
FACMR Maintenance and repair factor, fraction of VTOT/year
IPRINT(n) Print decision variables, n = 1-6
J Dummy variable for UNITS subroutine
NSTAGE No. of stages in series
NTRAIN No. of trains of units operating in parallel
PLF Plant life (years)
PYREX Payroll extras factor
QBARE Expected flow rate into plant, mgd
QDOT Design flow rate into plant, mgd
Qnn Flow rate of stream nn, mgd
RET Interest rate, fraction of VTOT/year
RTLAB $/man hour labor rate including payroll extras
TX Tax rate, fraction of VTOT/year
UBAR Utilization factor, QBARE/QDOT
113
-------
VTOT(n) Investment,, total of process n, K$
WCCHG Capital investment charge, $/year
WMRR Cost of maintenance & repair, $/year
WOP Cost of operation, £/Kgal
WTOT(n) Total production cost, $ /year, for process n
WTOTAL Total production cost, £/Kgal
XINS Insurance rate, fraction of VTOT per year
Variable Names for Lime Process
AFILT Filter area, sq. ft.
ALKnn Alkalinity in stream nn, mg/1 CaCO
CACnn CaCO- in stream nn, mg/1
o
COMPHP Horsepower of compressors, HP
DLIME Required CaO dose, mg CaO/liter
FACAOL Fraction of new lime which is active CaO
FAPAT Fraction of Ca OH(PO4) in solids exit. Accelator #1
J " O
FCACO3 Fraction of CaCO_ in solids exit. Accelator #1
O
FCAOHD Fraction of Ca(OH) in solids disposed of from slaker
£t
FCAOS Fraction of CaO in solids feed to slaker
FCCDAC Fraction of COD entering Accelators removed by Accelators
FCCDF Fraction of filter influent COD removed by filter
FINERT Fraction of inerts in solids exit. Accelator #1
FINRTD Fraction of inerts entering slaker which leave in disposed residue
FMGOH Fraction of Mg(OH) in solids exit. Accelator #1
FNVSS Fraction non-volatile SS in suspended solids
FTSS Fraction of TSS in solids exit. Accelator #1
FVSS Fraction volatile SS in suspended solids
IFILT 'Is there a filter?" 0 = no, 1 = yes
NFILT No. of filter trains required in parallel
114
-------
NVSS Non-volatile suspended solids, mgpl
PAPAnn Pound/day appatiteCa OH(PO4) in stream nn
Oi O
PASH Pound/day ash exit the kiln
PCACnn Pound/day CaCO in stream nn
O
PCAQnn Pound/day CaO in stream nn
PCAOH48 Pound/day Ca(OH) in stream 48
PCAOHD Pound/day Ca(OH) disposed of from slaker (Stream 31)
PCAOHR Pound/day Ca(OH) recycled to supply DLIME
£§
PINRnn Pound/day inerts in stream nn
HNRTS Pound/day inerts entering slaker from kiln and makeup lime
PMGHnn Pound/day Mg(OH) in stream nn
Zf
PMGQnn Pound/day MgO in stream nn
PO4nn Concentration of PO4 in solids in stream nn, mg/1
PRLIME Price of new lime, $/ton
PTSSnn Pound/day total suspended solids in stream nn
QN20 Newly computed recycle stream 20
2
RCAREA Area of recarbonation grids in one unit, ft.
RECOF Recovery fraction of lime (Ib/day CaO exit kiln/(Lb/day CaO in makeup lime)
RLPMGD Recovered lime per mgd, tons CaO recycled/QlO
RLTPD Recovered lime, tons/day, exit the kiln
TLBSnn Total Ib/day solids (dissolved & suspended) in stream nn
TSOLnn mg/1 total solids in stream nn
TSPMGD Tons of solids exit kiln per mgd of Q10
TSSnn mg/1 suspended solids in stream nn
VCARBU K$ for each recarbonation unit
VCONC K$ for recarbonation basin concrete
VFILT K$ for filters
VGRIDS K$ for recarbonation grids
VRCALC K$ for recalculation facility
VRCARB K$ for recarbonation, VCARBU + VCONC + VGRIDS
VSS Volatile suspended solids , mgpl
115
-------
WFUEL $/year for kiln fuel
WMUL $/year for makeup lime
WPOW $/year for electrical power
X Dummy variable for approximation formulas
XMGnn mg/1 Mg in stream nn
XMGHnn mg/1 Mg(OH) in stream nn
XMUL Makeup lime, Ib/day, as delivered (including inerts)
XMULN Newly-calculated value of XMUL
XSCAOH Excess Ca(OH) produced beyond requirement for DLIME, Ib/day
Variable Names For Clinoptilolite Process
ANnn Ammonia nitrogen, stream nn, mg/1 as nitrogen
ANXRGD Average no. of exchangers regenerated per day at QDOT
2
AREA Ion exchanger cross-section area, ft.
AVTL77 Average eluant temperature when less than 77 , F
BEDDEP Depth of resin in ion exchange beds, ft.
BTUAIR BTU/year required to heat stripping air to 77° F (25°C)
BTULIQ BTU/year required to heat eluant to 77 F
3
BVSTG Bed volume per stage, ft.
CDTANK Capacity of one drain tank, gal.
3
CFAPG Cu. ft. air per gallon of eluant, ft. /gal.
3
CFMAIR Cu. ft. per minute of air to stripper, ft. /min.
CHTANK Capacity of one holding tank, gal.
3
CLLF Clinoptilolite loading factor, Ib. NH -N/ft. resin
3
DEPMAX Maximum depth allowable for clino. bed, ft.
DIAM Diameter of clino. exchanger, ft.
DIAMAX Maximum allowable diameter of ion exchange vessel, ft.
116
-------
EBLOW Efficiency of blower
FACAOL Fraction of active CaO in lime used for regeneration
FCLPR Fraction of clino. lost per regeneration
GPMREG Regeneration flow rate, gpm
2
GPMSFC Calculated gpm/ft. in exchangers
2
GPMSFS gpm/ft. liquid loading to stripping tower
2
GPMSFX Imposed maximum gpm/ft. in exchangers
GPMSTR Flow rate of eluant to stripper, gpm
HPBLOW Horsepower of one blower, HP
HPQBAR Total horsepower required for pumps and blowers at production of
QBARE, HP
HPRPMP Horsepower of one regeneration pump, HP
HPSPMP Horsepower of one stripper pump, HP -
HPXPMP Horsepower of one exchanger pump (2 required for plant), HP
I Subscript dummy variable
NBLOW No. of blowers in stripper system
NDTL77 No. of days per year eluant temperature is less than 77 F (25 C)
NEXCHL No. of exchangers on line (at QDOT)
NEXCHR No. of exchangers being regenerated (at QDOT)
NEXCHT Total no. of exchangers installed
NHFT No. of half-feet in exchanger diameter
NRGS No. of regeneration systems
PCAOlf Pounds/day of Ca(OH) used for regeneration (excluding inerts;
3
PRCLIN Price of clinoptilolite, $/ft.
PLIME Pounds/day of lime used for regeneration (including inerts)
PRMBTU Price of one million BTU, $/mBTU
PRNACL Price of Nad, $/lb.
QEXCH Design production of one exchanger, mgd
QMAX Maximum production allowable for one exchanger, mgd
QSMALL Design production for plants small enough to use only one exchanger (not used)
117
-------
TAREA Stripping tower cross-section area, ft.
TLOAD Time to exhaust one exchanger at QEXCH, and QDOT, hours
VBLOW Investment of blowers, K$
VCLIN Investment of initial clinoptilolite, K$
VDTANK Investment of drain tanks, K$
VEXCH Investment of exchangers and associated pipes, valves and brine tanks, K$
VHTANK Investment of holding tanks, K$
VRPUMP Investment of regeneration pumps, K$
VSPIPE Investment of stripper system piping, K$
VSPUMP Investment of stripper system pumps, K$
VTOWER Investment of 2 stripping towers, K$
VXPUMP Investment of 2 exchanger pumps, K$
WLIME Cost of regenerant lime, $/year
WMUC Cost of makeup clinoptilolite, $/year
WPOW Cost of electrical power, $/year
WSALT Cost of salt for regeneration, $/year
Variable Names For Activated Carbon Process
ADSFR Adsorber flow rate (gpm) at QDOT
2
AREA Surface area of one adsorber, ft.
BEDDEP Depth of carbon in one adsorber, ft.
BED VOL Cu. ft. carbon required at QDOT
BVSTG Bed volume/stage, cu.ft.
BWF'R Backwash flow rate, gpm
GARB Total pounds carbon in equipment
CLF Carbon loading factor, Ib. COD/lb. carbon
CODEF ppm, COD in effluent
CODIN ppm, COD in influent
118
-------
CREGDY Pound carbon/day regenerated at QDOT
CRGDYE Pound/day carbon regenerated at QBARE
CT Contact time (minutes)
DIAM Diameter of one adsorber, ft.
DOSAGE Pound carbon/million gal. water
2
FAREA Furnace area for regeneration of carbon, ft.
2
GPMBW gpm/ft. flow rate for backwash water
GPMSF gpm/ft.2
2
GPMSFC Calculated gpm/ft. flow (as opposed to GPMSF which is imposed)
at flow rate of QDOT
NADSO No. of adsorbers operating (total, all trains)
NADST Total no. of adsorbers, including spares
NSPARE No. of spare adsorbers
PRCAR Price of carbon, £Ab.
PRFUEL Price of fuel £/mBTU
3
RHO Density of carbon, Ib/ft.
3
VOL Total volume of one adsorber, ft.
119
-------
CHAPTER 3
THE LOGISTICS OF MUNICIPAL RECYCLE
ILLUSTRATED BY THE SAN ANTONIO SUPPLY IN THE YEAR 2000
THE FLOW PATTERN IN MUNICIPAL RECYCLE
Outline of_ the Recycle Pattern
Figure 17 is a schematic flow diagram showing the flow pattern
in municipal water use and recycle. The portions boxed in heavy
solid lines are the inputs and outputs from the municipal system,
The portion boxed in heavy dashed lines is the advanced waste
treatment appendage to the conventional system which permits
recycle. The unblocked portions of the chart are the con-
ventional system.
The inputs to the system are the source water and the contam-
inant increment which occurs on one pass through the
municipality, the latter including both the organic, the
inorganic and the organism additions to the water through
use. The water passes through the distribution and use system
picking up the contaminant increment and is collected in the
waste collection system. From the distribution and use block
there occurs a loss of water used in irrigation, lawn watering,
street washing, etc., here termed "lawn loss." From the waste
collection system there occurs a loss from seepage out of the
pipes, termed "pipe loss." Actually there may and does occur
infiltration into the waste collection system from ground water
but this is not taken into account in the present study so that
the pipe loss is actually the net of infiltration and pipe loss.
That portion of the waste which enters the collection system
and does not appear in pipe loss is delivered to the conven-
tional waste treatment plant where most of the organics and
the organisms are removed or rendered harmless and only a
negligible loss of the water itself occurs. The effluent from
the waste treatment plant is discharged to a receiving water
body or water course. Thus, the output from the municipal
system is the lawn loss, pipe loss, and the disposal quantity.
All of the input must appear as output with the exception of
the organics which are oxidized into harmless gases in the
conventional waste treatment. Thus the water content of the
input must equal the water content of the output, and the
contaminants both in tbe water source and in the municipal
increment must equal the contaminants in lawn loss plus pipe
loss plus discharge. These contaminants are primarily the
minerals, which pass through the treatment system unchanged.
121
-------
PROCESS ELEMENT SEQUENCE IN MUNICIPAL RECYCLE
to
[Si
H
»•—•
-J
Water Sources
V
Water Treatment
•MBW-
Return
<—
<-—
>
v. Water
— >'
Was
V
__ A WT <
v !
Explicit 1
deminerali2ation|
1
, 1
Blend
distribution and use
\
te Collectioi
/
\
Contaminant i
Increment |
i <-
1 V
1 Existing or conventional
1 Waste treatment
1
~>
Lawn Loss
Infiltration
— -?>
Pipe
Loss
V
Discharge and disposal
. x
s
Output I
-------
In municipal recycle three elements are appended which recycle
•some or all of the water to the distribution system. (1) One
of these elements obviously is the return or "conveyance back,"
namely, the conveyance system used to return the water from the
collection point to the use .point. (2) The advanced waste
treatment process has the purpose of further reducing the
organic and organism content beyond that achieved in conven-
tional waste treatment. (3) The AWT process may also achieve a
demineralization incidental to the other processes going on
in AWT. If this incidental demineralization is not adequate
then there must be further appended an explicit demineralization
process to remove the inorganic ions which have been added by
use. Waste streams may be developed from the AWT process or the
explicit demineralization process which would require disposal.
In a simple municipal process in which the water is taken from
a stream upstream and the treated effluent is discharged down-
stream, the ultimate in pollution control is to so treat the
waste that its level of contamination is no greater than that
at the intake, thus leaving the stream to which the effluent
is discharged no more contaminated than it was before use. It
is obvious that if this ideal situation were reached the use
would equally well be served by returning, the purified dis-
charge to the intake such that there would be no discharge to
the stream. This recycle, completely closed with respect to
water, is the ultimate goal of advanced waste treatment for
reuse. In this scheme the output modes for the contamination
which occurs in use are through the la*m loss, through the pipe
loss, and by ultimate disposal of the wastes from the treatment
processes as dry solids or gases.
The present study accepts the lawn 'loss and the pipe loss and
attempts to develop process schemes by which the quantity of
discharge may be appreciably reduced.
If it were reduced to zero, the quantity of water to be taken
from the source, that is the quantity of makeup water, would
be an amount to equal the lawn loss and the pipe loss. Actually
it may be found that to carry the recycle to this extreme, that
is to cut the discharge to zero, would be more expensive than
to stop at some intermediate point. For example, the reduction
of the discharge to solids and the disposal of the solids may
prove much more expensive than the reduction of the discharge
to a concentrated solution and the disposal of that, containing
some water. The objective of the project is to provide means
"for determining in any specific case just where this o£ imum
quantity of discharge lies, that is which quantity of discharge
would produce the cheapest overall system and still meet the
water and discharge quality constraints. And this does not
exclude the possibility that that optimum point may be in some
particular cases exactly where it is now, namely, to discharge
all and recycle none.
123
-------
Qualitatively the factors which move this optimum point towards
zero discharge may be identified, a priori. They are:
1. A high cost of the water source.
2. A low cost of the recycle processes (the heavy
dashed box on the chart).
3. A low cost for disposal (as distinct from
discharge).
4. Extreme quality restrictions on discharge.
The Central Role o_f Loss Ratio
The input-output water balance for the system described is:
M = L + P + D
where
M = makeup water quantity, in ratio to water used
L = lawn loss quantity, in ratio to water used
P = pipe loss quantity, in ratio to water used
D = discharge or disposal quantity, in ratio to water used
A characteristic quantity termed "loss ratio" determines these
input-output relations:
Loss ratio = (L + P)
The total loss, lawn plus pipe,is the difference between the
water distributed and the sewage collected at the treatment
plant. The contaminants enter with the makeup water and with
the municipal increment. They leave in the lawn loss,, the pipe
loss and the discharge or disposal. In a conventional once -
through municipal system all the water used in each pass appears
as losses or discharge and disposal and thus all the contam-
inants entering in each pass leave the system in these three
streams.
If now some of the discharge is returned for use, without
demineralization, then the inorganic contaminants will build
up in the recycling water until their concentration in the
waste collected is high enough so that the water lost and dis-
charged can now carry out the input contaminant. Thereafter
the system will operate at this steady state concentration. If
some of these concentrations of the inorganic ions are accept-
able for use it is necessary either to increase the discharge
quantity and thus the makeup quantity, or to take inorganic ions
out of the water by demineralization and dispose of them pre-
sumably separately from discharging the water.
124
-------
However, the same function, output of inorganic contaminants,
is accomplished by the lawn loss and pipe loss. Therefore,
to maintain the same steady state condition with the same
input, a municipality with a high loss ratio will require
less discharge than one with a low loss ratio. Stated in
another way, this means that if the discharge is reduced to
zero by recycling all of it without demineralization then
the steady state contaminant level to which the system will
build depends upon the loss ratio, the higher the loss ratio
the lower the steady state concentrations.
Figure 18 shows a practical example of this being based on
the actual input concentration and municipal increment assign-
able to some of the ions for San Antonio. The graph shows the
concentration of these ions in the blend, that'is in the water
being used, when there is no discharge and no demineralization
(and when the sewage treatment process does not alter the
inorganic ion concentrations). Since the discharge is zero
the abscissa represents the loss ratio.
It is seen that as the loss ratio is cut back the steady state
concentrations in the blend increase but they do not increase
very much until the loss ratios.get below about 0.4. (The
Texas Water Plan calls for a loss ratio of 0.523 for San
Antonio for the year 2020.) If there were no demineralization
and no discharge the total dissolved ions in the blend being
used would rise to about 600 mgpl, the HCO3 to about 310, the
Ca to about 84, the SiC>2 to about 31. This would be the con-
dition if the treatment consisted of conventional biological
treatment. If the treatment comprised advanced waste treat-
ment with lime the inorganic ions would be affected. The
HCO3, Ca, Mg and probably Si02 would be reduced. Depending
upon the extent, this reduction could lower the total dis-
solved ions to a level even below the level in the makeup
Edwards water, even without explicit demineralization.
It is possible that with a loss ratio of 0.523 the blend would
be suitable for use, in the inorganic ions, without any de-
mineralization 'other than that occurring in the lime treatment.
On the other hand, for a loss ratio of 0.2 it is likely that
demineralization would be required since the ions not removable
by the lime treatment are built up to a high concentration.
For example, Na and Cl are at the 300 mgpl level. Even removing
the Ca and HC03 completely would still leave a TDI of about 800
mgpl which is not acceptable.
125
-------
1000
Blend Con-
Concentration
mgpl
100
(Lawn watering
+ Pipe losses
+ Discharge)/
City Use
SAN ANTONIO RECYCLE MINERAL QUALlTi' AS
INFLUENCED BY LAWN WATERING AND SEWER
LOSSES
No discharge, no demineralization; make-up,
Edwards Water
Present, no
recycle
0
.9 1.0 Figure 18
126
-------
Actual Seasonal Loss Ratios at San Antonio
The 0.523 is a projected loss ratio for the year 2020.
Actually the loss ratio can l?e expected to vary from month
to month. A study was made using actual San Antonio data on
total water withdrawn by all users in the sewageshed and total
sewage collected by all known treatment plants, as described
in the next chapter. The data for the years 1961 to 1965
are shown in Figure 19. It is seen that there is regular
variation throughout the year with a high loss ratio maximum
around 0.6 in the summer and a low loss ratio around 0.25
in the winter. The year-to-year pattern is rather consistent.
The five year average loss ratio is about 0.38 meaning that
on the average about 38% of the water withdrawn does not
appear as sewage collected and therefore is not available for
recycle.
Comparison of Figure 18 and Figure 19 leads to the rather
unusual situation that on recycle the San Antonio water blend
will be better in the summertime than it is in the winter,
quite the reverse of typical conventional supply which with
respect to mineral quality is usually better in the winter
than in the summer. Demineralization would be required in
the winter months and might not be required in the summer
months.
The Seasonal Variation of the Municipal Increment
The municipal (concentration) increment as available from
surveys and as used herein is the concentration difference
between the sewage treatment plant effluent and the water
being used, the difference being taken as the concentration
increment attendant upon a single municipal pass. The data
have been obtained by analyzing the waters in spot samples
or as averages. The question to be explored is: 'does this
municipal concentration increment have a seasonal variation
corresponding with the seasonal variation of loss ratio?
To explore this let the symbols be:
B = mgd of water distributed, called BLEND
b = concentration of BLEND in some ion, ppm (parts per
million)
L = mgd lawn loss, mgd
1 = concentration, ppm
Similarly:
P and p = pipe loss
S and s = sewage delivered
I and i = municipal increment
127
-------
WATER-SEWAGE LOSS RATIOS, SAN ANTONIO
Total water withdrawn, all users in sewage shed
Total sewage collected
5
3
0
1961
LOSS RATIO
Water-Sewage
Water
Five Year Average
\ I
\ I
V
MONTH NUMBER
10
11 12
Figure 19
128
-------
The symbol i represents the municipal concentration increment,
ppm, determined as described. I, the liquid volume associated,
is zero. The daily quantity of ion entering with the BLEND
is Bb, of which the units are ppmmg (parts per million times
million gallons) a weight of ion equal to 8.34 Ibs.
A material balance will show that on the day the i was
measured:
B - L = S + P
Total input of icn = Bb + (B - L)i, QI = (B - L)i = (S + P)j,
where QI is the symbol for the municipal Quantity Increment,
ppmmg. This assumes that the pipe loss is of the same con-
centration as the sewage delivered at the analysis point. (In
conventional sewage treatment there is practically zero change
in the inorganic ion concentrations other than sometimes phos-
phate and possibly ammonium-nitrate.) The question is: how
do the concentration increment i and the quantity increment
(B - L)i change with season, that is as B changes and L
changes?
Figure 20 shows, as gpcd (gallon per capita per day) the
quantities B and S for San Antonio developed from the data
used in Figure 19. It is seen that the sewage flow is
practically constant throughout the year. The lawn and pipe
losses in the wintertime are about 35 gpcd. If it is assumed
that the lawn loss component is practically zero in the winter-
time then the pipe loss is 35 gpcd. (See Figure 28 beyond.)
(P + L) winter =35 gpcd
L winter = 0
P winter =35
Formally this water here provisionally assigned as pipe loss
actually includes some irreducible minimum of water use which
does not return the water to the collection system and which
is not seasonal, such as street flushing, fire fighting, etc.
However, this pipe loss cannot be greater than about 35 gpcd.
In the peak months something of the order of 160 gpcd is used
for lawn watering and similar purposes which do not return
the water to the sewage collection system. San Antonio has
very few combined sewers, and while it is true that a heavy
rain will be reflected in an increased sewage flow, the
monthly sewage flow does not follow the monthly rainfall
pattern, thus let it be assumed that there is little actual
infiltration...presumably because the sewers are well above
the water table. That being the case, if the sewage flow is
constant there is no reason for the pipe loss to vary with the
season. Accordingly,
129
-------
MONTHLY AVERAGE PER CAPITA WATER WITHDRAWAL AND SEWAGE
DELIVERED - SAN ANTONIO 1961-65
360
63.--1
300
61
MONTHLY AVERAGES
as gpcd
200
120
100
I
MOiNTH NUMBERS
11 12
Figure 20
130
-------
P = constant at 35 gpcd
Now since P is constant and S is constant
S + p = constant = B — L = approximately 155 gpcd
and incidentally P = approximately .295P = approximately
.23 (S + P) .
If (S + P) is constant then the quantity increment QI = (S + P)i
must vary directly as i the concentration increment.
We do not know either how the quantity increment varies or how
the concentration increment varies. However, it does seem
logical that quantity increment, namely, the amount of ions put
into the sanitary sewer system by the day-to-day activities of
a municipality, should be constant and little affected by
season, rainfall, etc. The well known and often used figure
of .17 Ibs BOD per capita per day, for example, implicitly
bows to this concept. And if the quantity increment is con-
stant then the concentration increment must be constant because
(S + P) as just deduced is constant. However, this is only a
deduction containing many assumptions and the seasonality of
the municipal concentration and quantity increments must be
checked in a number of communities before the general recycle
problem can be attacked with confidence. For the present study
it will be assumed that both the concentration increment and
the quantity increment per capita are invariant.
Under recycle and reuse at steady state conditions all of this
quantity increment must in some way be removed from the system.
However, some of this removal is accomplished by the pipe loss
and the lawn loss. In the conventional system the remainder
is removed by discharge. In the recycle and reuse system this
amount must be removed by the AWT process including, if neces-
sary, explicit demineralization. The quantity involved termed
net quantity increment is:
NQI = S(b + i) - Bb, ppmmgd
In the recycle and reuse scheme the blend is no longer the
source water as it is with the conventional scheme but comprises
the blend of the source water with the return recycled water.
The NQI equation shows that if the blend is to be maintained at
the same concentration throughout the year, then S (b + i\ being
constant the net quantity increment decreases as the blend
quantity increases. This means that the load on the explicit
demineralization unit will be less in the summertime when B is
high and greater in the wintertime when B is low. An illustra-
tion of this is given in the demineralization section of
Chapter 6.
131
-------
The Geographical Variation or. Loss Ratio
The seasonal variation of loss ratio has been presented for
San Antonio. However, it is highly likely that this seasonal
pattern will vary with geography and climate as well as socio-
economic factors which vary from city to city. We have already
demonstrated that the seasonal pattern of loss ratio is a
highly important factor in. the engineering and the economics
of recycle. Before general recycle computations nationwide
can be made it will be necessary to develop such data for
numerous cities.
The Detailed Recycle Scheme
Figure 21 is a more detailed flow diagram for municipal recycle
with special regard to mineral quality. The capital letter
designations are quantities in mgd per mgd of water supplied
to the distribution system, i.e. of blend. (Note the difference
from the symbols as used on the immediately preceding pages.)
The lower case letters are the concentrations of an individual
ion in each stream.
The SOURCE includes the present source and its conveyance line
and the existing conventional water treatment, if any. In
addition, a supplemental source is provided. These are
combined as makeup water which is the input water.
In the USE step the makeup water is blended with the return
water to make the blend, the water supplied to the distribution
system. A small amount of blend is used in the various sub-
sequent processes. Shown on the chart is a use for regen-
eration water for ion exchange demineralization. The lawn loss
output comes out of the blend. The remainder is the water to
which the municipal increment is applied to produce sewage.
In COLLECTION the sewage is transported through the collection
system with an attendant pipe loss, the net of infiltration
and true pipe loss.
For CONVENTIONAL systems the sewage is delivered to an existing
conventional treatment plant which produces "secondary effluent"
which is discharged.
In the AWT recycle scheme tfie sewage is delivered to the AWT
process where under the scheme of this study it is treated
with lime to reduce the organics, Ca, HCO3, Mg, and P04;
then treated to remove the NH3 and finally passed through
activated carbon treatment to remove the last traces of
organics. It is also provided that the effluent from con-
ventional secondary treatment may be treated in a tertiary
stream, the same processes for tertiary treatment being
required as are here shown for direct AWT. The discharge
from AWT is discussed later.
132
-------
FLOW DIAGRAM MUNICIPAL RECYCLE
(with special regard to mineral quality )
Present source i
Make up
water
M. m
e.g. regeneration
water
R,b
Increment ;
0.0. i i
Inliltration not
taken into account
delivered
1-L-K-P. h
AWT Processes
I Existing conventional treatment
CONVENTIONAL
Effluent from conventional
C, b-M
I - L - R -P - C+T. 2
Discharge from AWT
T+M-L=R-P-C. z
Main plant discharge
M-L-R-P. z-
Carbon efluiem
HI-MINERALIZATION
Demin. feed
X. z
Dc'min. by-pass
Demineralizatioi
M. L, etc = quantities, e.g. mgd, of each stream.
m, b, etc. = concentrations, e.g. mcl, of jth ion in each stream.
f = leakage in demincralization
133
Figure 21
-------
If DEMINERALIZATION is required the carbon effluent is treated
by some demineralization process here shown as an ion exchange
process, producing a demineralized product. The product from
several of the possible demineralization processes would be
more pure than required for recycle and therefore, it is
possible to by-pass some of the carbon effluent around the
demineralization process.
The by-pass and the demineralized product are mixed as the
RETURN liquor which is conveyed back to the distribution system,
The ultimate goal of recycle is to achieve zero discharge. If
in the flow chart shown the AWT and demineralization processes
could accomplish the necessary purification then the conven-
tional treatment line could be eliminated and the only disposal
would be of the demineralization waste water. However, some
demineralization processes are not capable of removing all
ionic contaminants. Those that are removed are output from
the system in the regeneration water but those that are not
removed must be output by a purge of the recycling liquor
before the concentration of the particular ion reaches unaccept-
able levels. In an AWT system not incorporating a conventional
waste treatment plant it would be inefficient to discharge the
carbon effluent since probably for discharge or disposal it is
not necessary to remove the last traces of organics. Accord-
ingly, it is likely that the discharge from the AWT process
would be made prior to the carbon treatment. If, however, the
system contains a conventional treatment plant, for which the
capital costs are already sunk, it would probably be more
economic to operate the existing conventional plant in order
to achieve an effluent suitable for discharge which could
serve as the purge for the ions suffering excessive build up.
There are numerous possibilities not shown on the chart. For
example, it might occur that the effluent from the conventional
treatment plant was discharged to a stream while the regen-
eration water from demineralization and other process waters
from the AWT processes, if any, are handled by ultimate
disposal in some manner.
Also, of course, ultimately this project should seek to
provide for situations in which the split of the delivered
sewage between the existing and the AWT processes is not a
matter of choice but is forced by some physical configuration.
For example, it might be uneconomic to convey into an AWT
sewage shed the flow now going to an existing conventional
treatment plant.
134
-------
PRELIMINARY LOGISTICS OF THE
SAN ANTONIO SUPPLY IN THE YEAR 2000
This section develops data for San Antonio and uses these to
determine the flow pattern in the year 2000 for two extreme
situations: (1) conventional supply, treatment and discharge
with no demineralization, and (2) complete recycle and reuse
via advanced waste treatment with no discharge (and no de-
mineralization) . The costs of these two extremes are
developed in Chapter 6.
Area Population Projection
The preliminary 1970 Census places the population of the San
Antonio District as 684,322 on April I, 1970. The district
includes the City of San Antonio and its surrounding com-
munities and military bases, thus approximating the population
served by the water and sewage system corresponding to the
1961 to 1965 water and sewage data described in the previous
chapter.
The actual 1960 to 1970 growth for the San Antonio District
falls considerably short of the growth projected by the Texas
Water Plan for the major cities of the county. However, it is
assumed in this study that in the year 2000 the system under
consideration will be serving the entire population of Bexar
County. The Texas Water Plan projections for the 1960 to 1970
growth of the County were closer to the actual experienced
growth. The experienced growth to the 1970 Census population
of 830,661 was about 20% from the 1960 population and a
projection at 20% per decade yields 1.44 million population
for the County in the year 2000 which is satisfactorily close
to the Texas Water Plan projection for that year of 1.42
million. The 2000 projection of the Texas Water Plan for the
major cities of the County is 1.33.
Historical Water Withdrawal and Sewage Delivered
To obtain the seasonal gpcd figures mentioned in the previous
chapter and used in this chapter the following procedure was
used. The U.S. Geological Survey San Antonio office collets
annual data on the pumpage from the aquifer for Bexar County
and other surrounding counties broken down in detail by the
actual withdrawal agency of which the major one, of course, is
the San Antonio City Water Board. The breakdown includes the
some 25 other independent public supplies, the military bases,
the City parks and zoo, industry broken down by individual
135
-------
establishments, private commercial use, air conditioning,
etc., private schools, country clubs, etc., flowing wells,
springs, domestic stock and country estates, and irrigation.
For the years 1956 to 1965 this tabulation is available by
months. From this total pumpage from the aquifer, which
represents the sole water supply, there are subtracted the
flowing wells, the springs, the irrigation, and the domestic
and stock; the remainder taken to represent the water with-
drawn which is potentially contributory to the sewer system.
Annual and monthly figures for sewage delivered are available
for the Rilling Road and Leon Creek sewage treatment plants.
For the seasonal studies of the previous chapter the annual
sewage delivered at the smaller sewage treatment plants in
the area was estimated by the procedure described beyond under
long term sewage trend. The total sewage delivered each month
was estimated by applying to the sum of the Rilling Road and
Leon Creek plant flows the annual factors, ranging from 1.041
to 1.050, determined in the long term study.
The flow of the San Antonio River at the Elmendorf Street
gauge is available from October 1962. The flow for previous
months was estimated by a correlation between the Elmendorf
flow and the Falls City station flow for the period October
1962 to September 1966. The relation is:
Elmendorf = .934997*Falls City**(1.0146}
The correlation coefficient is .98220; the a ratio 1.103.
This relation is not used in the present project because the
project does not go so far as to consider the possibility of
using the future Elmendorf net flow for water supply.
The population at each month in the 1961 to 1965 period was
computed by assuming a uniform logarithmic increase between
the San Antonio District populations of 618,944 in 1960 and
684,322 in 1970, a monthly increase of 0.083713%.
The monthly data so developed have been used in the previous
section (Figures 19 and 20) illustrating the pattern and are
used beyond in setting the logistics of the 2000 supply
(Figure 28).
In an effort to obtain some correlation which might allow
predictions, some manipulations of the five year data were
made. "Reducing" the loss ratios by dividing each by the
average of the monthly loss ratios for the year does not
effect much of a compression of the band such as observed
in Figure 19.
136
-------
While there is a downward trend as expected, the monthly loss
ratios are not correlated simply with the monthly rainfall...
in other words, very little of the variance in the 60 monthly
loss ratios is removed by plotting against monthly rainfall.
The 60 loss ratios do show a correlation in the expected
direction with average monthly temperature, increasing as
the average monthly temperature increases from about 45 to
87° F (degrees Fahrenheit). An appropriate form for the
relation is a second degree polynominal. However, it is
noted, as would be expected, that with approximately equal
monthly temperatures the months having high rainfalls tend
to have lower loss ratios. Accordingly, a better predictive
equation is obtained from a multivariate regression yielding
a relation:
Z = .703396 - .0192044T + .00214679T2 - .0267191R
where
Z = calculated monthly loss ratio
T = average monthly temperature, °F
R = monthly rainfall, inches (airport)
N = 60, standard deviation = .0573, correlation
coefficient = .852
About 72% of the original variance is removed by this multi-
variate correlation. Analysis of the residuals shows that
compared to the true loss ratio the quantity Z tends to be low
in the spring and high in the fall, varying sinusoidally with
month throughout the year. Observationally, this means that
for months having equal temperatures and equal rainfalls a
calculated loss ratio tends to be lower than the observed if
it is a spring month and higher if it is a fall month, and to
be about equal if in the winter or at midsummer. This residual
variance would be reduced if a sine term is included in the
multiple regression.
Of course, it is a misnomer to call this a predictive equation
since in order to use it it is necessary to know a future
average monthly temperature and a future monthly rainfall.
However, the analysis was made in order to be prepared for
comparisons with similar data in other cities. When such a
study is made it will draw on the more sophisticated studies
of municipal use relations which are being made by other
investigators (26,27,28,29). For example, Reference 26 pro-
vides a correlation for lawn irrigation which includes such
variables as average irrigable area per dwelling unit, mean
monthly temperature, monthly percent of daylight hours, effec-
tive rainfall, and an empirical monthly crop coefficient for
grass.
137
-------
The USGS San Antonio office has been recording, estimating,
and summarizing the water withdrawal by category in Bexar
and surrounding counties, published information going back
to 1934. The summary sheets for these publications were
located and they were brought up to 1968 from recent annual
reports. The uses categorized are:
Municipal (:subtotal of 4)
Industry ( subtotal of 2)
Country clubs, private
schools, etc.
Domestic and stock, estates,
misc.
Irrigation
Salado Creek and
other flowing wells
San Antonio and San
Pedro Springs
The four categories in the left column are those considered
to measure the future water use of Bexar County. The City of
Schertz is included in the municipal supply but its population
is not in Bexar County. Over the years 1956 to 1968, the with-
drawal for Schertz was about .25% of the total withdrawal. The
County withdrawals were corrected for Schertz by subtracting
Schertz in the years available and applying the ratio .9975 in
other years.
To obtain the corresponding population of Bexar County leading
toward a gpcd figure the Census population at each decade was
interpolated between at a constant annual percentage increase.
These annual increases incidentally are: 1930-1940 1.46%;
1940-1950 4.00%; 1950-1960 3.22%; 1960-1970 1.91%.
The resulting gpcd figures are shown in Figure 22. It is seen
that there is no strong long term trend toward either an in-
crease or decrease in the gpcd use. It might be considered that
the trend was rising during 1935 to 1956, but with the breaking
of the drought in 1957 the gpcd usage fell to the level it
bore 20 years prior, and in the subsequent 10 years has hovered
around 200 mgd. The 35 year average is 214.1 gpcd. Explora-
tion for the significance of the trend taken as a Cartesian
regression line indicated a slope of about 0.42 gpcd/year, and
not significantly different from a slope of zero, at about a
50% level of significance.
Even if the trend should be significant the regression shows
a gpcd of 207 in 1934, 220 in 1964, and 235 in 2000. Table 19
beyond, indicates that the annual average gpcd corresponding
to the seasonal pattern base taken for design in 2000 is 236
gpcd.
This 35 year trend, incidentally, while small, is even larger
than the long term trend for large U.S. cities in general. The
hundred-year average trend for some three-score large U.S.
cities is about .30 gpcd/year. That per capita municipal use
is rapidly increasing is a popular misconception - if applied
to even moderate sized cities.
138
-------
BEXAR COUNTY ANNUAL WATER USE
ex IRRIGATION
OS
VO
300
250
200
150
100
50
0
40
Municipal, Industrial
Country Clubs, Schools
Domestic, Stock, Estates,
ex Schert z
gpcd, ex irrigation
mgd, ex irrigation
50
YEAR
60
68
Figure 22
-------
Not so easily resolved is the long term gpcd trend for sewage.
In 1969, there were about 14 sewer collection systems and
small treatment plants in addition to the two City plants,
Rilling and Leon Creek. Also, there are eight incorporated
communities for which Census data are available and one over-
all "District" which includes these, the City of San Antonio,
and some contiguous urbanized territory, all within the
sewage shed. The Kelly Air Force Base collection system
and treatment plant has been operating since before 1940.
The other small treatment plants and collection systems came
into being starting about 1956 and most of them were installed
by 1963. Some of the separate Census entities have been
served by the City sewer system since 1940, others have
varying inception dates of sewer service, up to as late as
1966.
Data on starting dates for service to the various communities
were obtained from San Antonio City Hall (Finance Department)
together with some data on connections. Data on current pro-
duction rates of the various treatment plants were obtained
from TWQB (Texas Water Quality Board), except the historical
record for Rilling, Leon Creek and Kelly AFB (Air Force Base) .
Data on the starting date of these treatment plants were
obtained from the operating agencies, Water Control and
Improvement Districts, etc. together with information on the
generalized growth pattern of each. Sewage flows between the
starting date and 1969 were then estimated in accordance with
the growth pattern. For Rilling and Leon Creek the complete
records were available. For Kelly AFB records were available
back to 1955. Data on the Kelly performance 1940 to 1955 were
obtained from the retired foreman of the plant. Populations
in the intercensal years were interpolated from the Census
data for the various civil divisions on a constant logarith-
mic increase characteristic of each decade. For other than
census divisions population figures were estimated from
connections. When all known surrounding communities are
added to the San Antonio census population, a figure is obtained
which is less than that given for the "San Antonio District."
The difference was assigned to a population entity termed
"missing from District" and the population for intercensal
years interpolated as for an actual census division. The
"missing from District" includes unincorporated areas not in
the City limits but closely associated with it. The "San
Antonio District" is a term generated by the local census
operation, has a population somewhat less than the San Antonio
urbanized area, and the population "missing from District" is
about 1% of the total population served.
140
-------
The total flow generated by the 17 treatment plants was
divided by the total population served in the 13 population
entities to generate the gpcd sewage flow. The resulting
data are shown in Figure 23, the gpcd water use, the gpcd
sewage collected and the difference, that is the loss between
water and sewage. While the water use shows no trend except
for the peak in the drought period ending 1956, the sewage
data reveal an upward trend.
It is believed that part of the long term upward trend in
sewage arises from a deficiency in the data, namely, that
in the period before 1956 not all of the City population was
served by sewers and, therefore, the denominator being too
large the gpcd becomes too small. This hypothesis is con-
firmed by an analysis of the loss ratio (water minus sewage
in ratio to water) as a function of annual rainfall shown on
Figure 24. There is not much discernible trend with rainfall.
If there is a trend it would appear to be in the proper
direction that is toward lower loss ratios'at higher rainfall.
However, the important point is that the data fall in two
groups, 1957 to 1968, and prior to 1957. The latter loss
ratios fall in a group significantly higher than the former.
This is the condition that would result if the sewage gpcd's
were too low in the period prior to 1957. To resolve this
question would require a detailed study of the population in
the various tracts served by sewers in each year as the City
expanded its sewer system. :
Figure 25 shows the time trend of the loss ratio from which
it is apparent even ignoring the pre-1957 data that the loss
ratio is in a downward trend, resulting from an increasing
sewage flow in the face of a constant water use. The in-
creasing sewage flow might have been attributed to an erron-
eous inclusion of the numerous small systems which have sprung
up since about 1957. However, even if this had been the case
the effect of these on total sewage flow has been indeed quite
minor. The ratio of total sewage flow to the flow from the
City's two treatment plants was about 1.04 around 1957 and
has gradually risen to about 1.07 and this 3% difference is
not nearly enough to bring the loss ratio from around .4 to
around . 2 .
Without further study this project cannot arrive at a pre-
diction for the future sewage flow and loss ratio. The design
arrived at in the subsequent section "Logistics of the 2000
Supply" takes the upper envelopes of the 1961 to 1965 period
as the total use and makeup quantities and assumes no change
from 1961-65.
141
-------
LONG TERM WATER AND SEWAGE RELATIONS
SAN ANTONIO
NJ
Period of the
Seasonal Study
LOSS,(W-S)
GPCD,
ANNUAL
Sewage probably too low and
loss too high in this range
because population served
probably too high
I
1940
1950
1960
YEAR
1965
Figure 23
-------
ANNUAL RAINFALL vs. LOSS RATIO
RELATIONS, SAN ANTONIO
(Showing two distinct time groupings)
CO
.5
.4
.3
.2
LOSS
— RATIO
12
15
ratio considered higher
an true value
I
20
25 30 35
ANNUAL RAINFALL, Inches
40
45
48
Figure 24
-------
LONG TERM ANNUAL LOSS RATIO
SAN ANTONIO
Period of the
seasonal study
ANNUAL
LOSS
- RATIO
Ratio probably too high in this
range because of population
served by sewers too high
40
YEAR
Figure 25
-------
In San Antonio the maximum day in each year occurs in the
maximum month or in an adjacent month. Data are conveniently
available 1953 to 1969. To eliminate the variability of
population served and total annual water use there was computed
the ratio of the maximum day to the maximum month, both ex-
pressed in mgd terms. Plotted, these ratios showed a cyclic
pattern with no time trend. The 17 ratios are log-normally
distributed with a mean of 1.21 and a a ratio of 1.06 (68%
of the points lie within about 6% of the mean). The popu-
lation 90 percentile is 1.30. This means that there is only
a 10% chance that the ratio in any single year might be greater
than 1.30. The population 10% is 1.12.
It is concluded that once the maximum month has been established
in the planning, the maximum day can be established as 30%
higher.
Similarly the annual reports and other data from the City of
San Antonio sewage treatment plants, available 1950 to 1968
can be used to assess the maximum day/maximum month ratio for
sewage collected. Plotted, the data show no time trend. The
20 points are log-normally distributed with a mean of 1.46 and
a o* ratio of 1.20, signifying that about 68% of the points lie
within about 20% of the mean. The population 90% for the ratio
is 1.85. It is concluded that there is only a 10% chance that
this ratio in any year may be above 1.85. Accordingly, this
factor can be used to compute the maximum design day from the
maximum month. The sewage maximum day is higher and has a
greater variance than the water maximum day, presumably because
of the great influence of rainfall on sewage flow.
In a more thorough study than is possible here it would be
desired to study not only the maximum day for water and for
sewage, but also the maximum day for the difference between
them and, indeed, the whole statistics of this net difference.
For in a reuse scheme, unless storage is resorted to, only that
much waste water can be recovered on each day as the community
uses water on that day. Any daily excess of sewage collected
over water used must be discharged, and any daily excess of
water used over sewage collected must be made up by makeup
water. This aspect must be left for further detailed investi-
gation .
145
-------
Existing Facilities
The peak day capability of the three existing or planned San
Antonio conventional sewage treatment plants, the Rilling,
the Leon Creek and the Salado are Rilling 80, Leon Creek 12,
and Salado (under construction) 24, for a combined peak day
capability of 116 mgd. These figures are based on detention
times in the aeration basins sufficient to provide an effluent
meeting the present specifications (30). The average pro-
duction from these plants as operated at present is greater
than they will be able to supply under the seasonal use
pattern later to be developed for this study.
The City Water Board's production facilities (31) comprise
eight primary stations having 34 wells and 26 smaller secondary
stations having 26 wells, the installed capability of the well
pumps being respectively 285.3 and 94.3 mgd. Dropping out the
largest well at each primary station would reduce the capability
by 81.8 mgd. Thus the ratio of installed capability to firm
capability at the eight primary stations is 1.4, firm capability
being defined as the highest daily production that could be
obtained with the largest unit out of service. In the compu-
tations it will be assumed that this ratio of installed cap-
ability to firm capability will apply to the entire ground water
facility system in 2000.
In addition to the City Water Board production the other
municipal supplies produce an additional 20% or more. The
average of the ratios of the total water production from all
municipal supplies to that from the City lAkter Board over the
four years 1965 to 1968 was 1.229. Assuming that the installed
capability follows the same ratio as the production, the
estimated installed capability of all existing municipal wells
would be 466 mgd, and by application of the 1.4 factor the
estimated firm capability of all existing municipal wells 333
mgd.
The Municipal Increment for San Antonio
Table 17 shows the assignment of the San Antonio municipal
increment either from the San Antonio data or where this was
not available taken from the average for western cities. In
the former case, the western cities average is shown also for
comparision with the apparent San Antonio increment. The
average increment for the western cities comes from one to 22
cities from Reference 32.
146
-------
TABLE 17
MUNICIPAL INCREMENT, SAN ANTONIO AND WESTERN CITIES AVERAGE
Apparent Increment S . A . Western Cities Average
MAJOR IONS
Na 74
K 11
NH 18
Ca 15 13
Mg 0 7
Fe(2) .4 .2
Cl 67 92
HCO, 81
NO, 7
NO, 2
F "* •
SO 29
CO4 -1
17
PO , total (raw)20 28
PO , ortho 25
MISCELLANEOUS
Total alkalinity 66
Conductance
Residue 352
Temperature 0
pH .0 -.4
Hardness 43 58
MINOR ELEMENTS
Al .8
Ba < .4
B .3
Mn < .1
Sr
Cr < .2
Pb < .8
Mo < .2
Co < .2
Ni < .2
Cu .3
Sn < .2
Zn < .5
Ag < .2
Li
ORGANIC
BOD 23
COD (96)
147
-------
The study took the analytical data for the water supply and
the sewage treatment effluent for a number of cities, up to
as many as 33 for some contaminants, and developed the range
and averages for the concentration increment for eastern
cities, western cities, and for both together. The San
Antonio data came from analytical data on the composite
water 1/1/64 (31) , and from the average of six well stations
2/8/62 (33) . The minor elements analysis came from spectro-
graphic analysis from one of the stations. The waste water
analyses which were for sewage treatment plant effluent except
as noted came from analysis of the effluent of the West plant
in the Rilling Road complex made on composites between 5/28
and 6/9/68, the minor elements again by spectrographic analysis,
Also included were some spot analyses of raw sewage on the
north side of the sewageshed 1968 (34). Also used were some
data from the Rilling complex July and August 1966 showing
the hardness and incidentally demonstrating that hardness
does not change on passage through the treatment plant. The
figure for total phosphate is one of those obtained from the
raw sewage and this is somewhat important since San Antonio
plants effect a reduction in phosphate. Some of the other
analyses also were from raw sewage but this is not important
since there is no change in those ions on passage through the
sewage treatment.
This Table should not be taken as anything more than a very
sketchy approximation to the real situation in San Antonio.
In the first place as has previously been indicated, we are
not sure that the municipal concentration increment does not
fluctuate according to season, but the analyses are based on
spot samples or short time composites. Secondly, as Reference
32 shows, the ranges from city to city for each contaminant
are considerable. About an eightfold range is typical for
the western cities.
It is also clear that these increments, coming from averages
and spot analyses, will not be in ionic balance. This is not
of particular importance except for explicit demineralization
processes. When used for such purposes the water composition
must be adjusted for ionic balance.
If these data are used for the design of a recycle system, it
should be recognized that the result would be only illustrative
How far such a design would be from reality in equipment sizes,
in performance, and in dollars is completely unknown and a
real design would have to await much more extensive data for
the individual city.
148
-------
The General Pattern of Compositional Changes on Recycle
Figure 18 has shown how the composition of the blend water for
San Antonio would depend upon the loss ratio with no discharge
and no demineralization if there were no compositional changes
in the waste treatment process. The monthly loss ratios having
been developed as in Table 19 beyond it is possible to show how
the average monthly composition of the blend would change from
month to month.
Since the Program AWTLCC (Chapter 2) provides the composition
of the AWT effluent for any given feed composition it would be
possible also to incorporate the effects of the compositional
changes which occur in the AWT process. The equilibrium com-
position of the blend must be obtained by iteration, as in a
recycling flow sheet. When the complete RECYCLE program is
established this will be done. Meanwhile the compositional
effects of the AWT process were approximated from AWT runs
already made and this surrogate AWT process was used in a
small program (WGMONSA) to compute the blend composition for
San Antonio.
The values used in the composition study are shown in Table 18,
the first column being the composition of the source water, i.e,
of the present San Antonio supply, the second column being the
municipal increment as previously established; and the third
and fourth column pertaining to the AWT process itself. In the
third column is given the effluent concentration for those con-
taminants for which that value is set by the assumptions, for
example, 0.5 mgpl for NH-j-N. In the fourth column are given
the changes in concentration for those contaminants not so set.
These are the changes actually observed and generated by the
AWT program operating on San Antonio waste as described in
Chapter 2.
Figure 26 shows the concentrations of some of the contaminants
as a function of the loss ratio. Comparison with Figure 18
reveals the significant differences attendant upon the com-
positional changes in AWT. Na and Cl are higher because
NaCl is used as a reagent in AWT and contaminates the product
water. Ca is unaffected by loss ratio because it happens
that the increment taken is almost exactly the same as the
fixed concentration taken for the concentration in the AWT
effluent and accordingly, the loss of Ca is almost precisely
equal to the increment of Ca at a blend level close to the
source water level. The behavior of Mg and HCO^ is quite
different from that in Figure 18 because large removals of
these contaminants occur in AWT. Indeed, contrary to the
general trend the lower the loss ratio the lower the concen-
tration of these two contaminants in the blend. The overall
effect of the AWT compositional changes is to lower the TDI
over that of Figure 18.
149
-------
TABLE 18
VALUES USED IN COMPOSITION STUDY WQMONSA
Na
K
NH,
4
Ca
Mg
Cl
F
N02
N03
HC03
C°3
S°4
Si03
P°4
COD
VSS
NVSS
TDI
Source
Water
7.8
1.0
0.
64.
17.
15.
0.3
0.
4.5
241.
0.
23.
15.2
0.
0.
0.
0.
48.2(2)
Municipal
Increment
74.
11.
18.
15.
0.
67.
0.
2.
7.
81.
0.
29.
17.
20.
492.
162.3
56.7
341
AWT
Cone . *
0.5
64.02
1.25
69.09
0.9
8.
0.7
0.3
Effluent
A Cone.
-6.12
0.
-17.44
0.
0.
0.
0.
0.
o.
(1)
(1) Summed by the Program
(2) Unlisted ions to bring total ions to 437
* Concentration
150
-------
BLEND COMPOSITION AT VARIOUS LOSS RATIOS
SAN ANTONIO
1000
100
AVERAGE
MONTH LY
CONCENTRATION
mgpl
0
.4 .6
LOSS RATIO
1.0
151
Figure 26
-------
Figure 27 shows the average monthly compositions of the blend
water for each month resulting from the design loss ratios of
Table 19. It is seen that the TDI is lowest in July and
August and highest in December and January, and that this
also is true of all the other contaminants except Mg and HCO^
for which the reverse is true.
The major revelation, however, is that under these conditions
the mineral composition of the blend is too high in any month
of the year. The TDI of the undemineralized blend barely
falls below 500 mgpl in July and August and in December reaches
more than 1,000. Since the compositional changes of AWT have
only been approximated in the program which generated this
figure there is a slight possibility that the blend resulting
from AWT runs will be somewhat different, and possibly better.
But this is not very likely.
Accordingly, it appears likely that either demineralization
or discharge or both will be required in order to generate a
blend of sufficiently low TDI to be generally acceptable, in
San Antonio. Conditions would be better if the city had a
higher loss ratio or had a lower municipal increment. San
Antonio already has an extremely high loss ratio, up to more
than 60% in July and August and it is doubtful if there will
be found in the nation many other cities having loss ratios
this high. As for the municipal increment, the increments
used for San Antonio either are or were deliberately selected
close to the average for western cities. Accordingly,
probably about half of the western cities have lower municipal
increments and, eastern municipal increments being generally
lower, more than half of the eastern cities have lower incre-
ments. A separate study is being suggested of monthly loss
ratios and municipal increments in other cities but it seems
clear that there are cities, San Antonio being one of them,
in which either demineralization or discharge will be required
to maintain blend quality at an acceptable level.
152
-------
100..
^!DL_ ....KM
_„_
JJJANIJ KSJJX _•;] j:JSP CO\n x>-iLIlQjs
SAN ANTONIO kf.<)
d._Y.bii I
-------
Logistics of the 2000 Supply
Figure 28 shows the monthly average per capita excess of water
withdrawal over sewage delivered for the 1961 to 1965 period
generated from the aforementioned data. These curves represent
the gpcd of makeup water that would have to be supplied if the
entire sewage flow were reused. The five-year average is 89
gpcd. For the design basis for this study the upper envelope
points of these curves were taken except for May where the
1961 point is very high, in which case the second highest was
used. The monthly figures for the envelope are shown in
Column 2, Table 19. Multiplication of these figures by 1.44
million yields the monthly average design mgd of makeup for
the year 2000 shown in Column 3, of Table 19, and as the
correspondingly labeled curve in Figure 29. This annual
average design is 114 gpcd, higher than the 89 gpcd five-year
average because the upper envelope was used in the design.
A similar procedure using the upper envelope was applied to
the water withdrawal curves from Figure 20 of the last section
and represented as total intake in Column 4, Table 19. These
gpcd rates applied to the 1970 population give the 1970 intake
shown in Column 5f and applied to the 2000 population give the
2000 intake in Column 6, represented in the corresponding
labeled curve, Figure 29.
There are two constraints on the use of ground water, an
annual constraint of 215 thousand acre feet per year imposed
by the Texas Water Plan as a safe annual yield which ;will
maintain the flow of the springs, and a peak monthly with-
drawal constraint. The experienced peak monthly withdrawal
in the 1961 to 1965 period was about 9,400 mg (million gallons)
for the total pumpage, corresponding with about 7,000 mg with-
drawal contributory to the sewers, that is after the deductions
for irrigation, flowing wells, etc. The maximum limitation of
10,000 to 11,000 mg per month for July and August consecutively
has been offered as a constraint (35) . In average mgd units
these constraints are 192 mgd annual, and 339 mgd for the
July-August peak months.
Column 6 indicates that ground water within these constraints
cannot supply the year 2000 demand since this calls for an
annual average of 341 mgd and a peak month of 515, both of
which exceed the constraints.
154
-------
240
220
200
180
160_
140_
120.
100_J
80
60
40 _
20
0
MONTHLY AVERAGE PER CAPITA EXCESS OF WATER WITHDRAWAL
OVER SEWAGE DELIVERED
SAN ANTONIO 1961-1965
MONTHLY AVERAGE
Water-Sewage
gpcd
1
557
MONTH NUMBER
10
11 12
Figure 28
155
-------
500
400
350
300
250
t
mgd
©
©
®
(3)
X
-y^t
surfa
_.•
..-•"
s
MONTHLY
WATER PATTERNS
2000 A.D.
San Antonio
jf
S
/
7
<^
Basin
7
'
*j^>
Ground Water,
conventional
Average 192--
x
:e
..--"
7^
i
f^js
y ••'
/
_•
/
/
/
•^Q-<
/
/
/
/:
i :
i •
.'
.
I
/
'
/
/
l
i
t
f
/
/_ \
^^
^•^
\ Ground water , reu
' Average 164 • —
\
\
\
\
\
•
\
\
•A
\\
\\
\\
\\
\\
\\
\\
\\
*
:
;
"-•^^
^^-^
se J^"
_ — "^
Conventional
vs.
Reuse
V
\
\
\
\
'•• \
•. \
'. \
t 1
*
'
CJ Column in table
- 00 Ground Water, reuse
© Total intake
(j) Ground water conventional
® Surface water conventional
Month
Numb €
r
stal ii
Ttake
0
, 3'' 1 Avprqf p
V
\
\
\
\\V-
^i.
>
\
•.
'•
%
A How-
month
groun
\
\
able •'
ly pea
d wate
/
k
r
Allowable
annual ^
6
round water
CD
— Surface water
conventional
149 Average
X
* ^
* .
(a)
*-'
d)
.
Figure 29
3 39
ii
123456789 10 11 12
Paee 156
200
150
100
-------
TABLE 19
SEASONAL LOGISTICS OF WATER SUPPLY
BEXAR COUNTY, 2000, CONVENTIONAL vs. REUSE
Makeup
4 5
Total Intake
7 8
Conventional 2000
Month
1
2
3
4
5
6
7
8
9
10
11
12
Avg.
gpcd
47
61
72
113
121
165
233
231
137
94
52
38
114
mgd
2000
68
88
104
164
174
238
336
332
197
136
75
55
164
gpcd
170
180
192
228
250
290
350
358
268
218
174
160
236
mgd
1970
117
124
132
157
172
200
' 241
247
185
150
120
110
163
mgd Proposed
2000 ground water
pattern, mgd
245
259
276
328
360
418
504
515
386
314
251
230
341
125
129'
136
172
195
253
339
339
218
149
126
122
192
Required
surface
water
6-7
120
132
140
156
165
165
165
176
168
165
125
108
149
Sewage
treatment
or AWT
6-3
177
171
172
164
186
180
168
183
189
178
176
175
177
157
-------
But Column 3 shows that ground water used alone as makeup
in a recycle scheme would be adequate in the year 2000. The
peak month is 336 mgd compared to the constraint of 339 and
the annual is 164 compared to the constraint of 192. There-
fore, if all the waste water could be recycled the entire
makeup to the year 2000 could be supplied by the allowable
ground water withdrawal.
Projection of the population to years beyond 2000 indicates
that the peak month constraint is just barely met since it
would be violated about the year 2005. The annual constraint
would be exceeded about the year 2017. Starting in 2001 there
would have to be provided some surface storage for ground water
pumped in the winter and spring months and stored to avoid
exceeding the peak allowable in July and August. The storage
period and quantity would have to become larger and larger as
the population grew. Beginning in 2017 no amount of storage
would suffice and it would be necessary to supplement the
ground water supply.
However, of course, these projections depend upon the constancy
of the gpcd water use and sewage delivered. Under a reuse
scheme any steps taken to reduce the gpcd water intake and
increase the gpcd sewage collected would be favorable toward
postponing these critical dates.
In any event if the sewage collected would be completely
recovered the allowable ground water withdrawal would meet
the requirement in the target year 2000. However, anything
even slightly less than 100% recovery of the sewage collected
would violate the peak month constraint in 2000.
As may be seen the thrust of this project is to utilize ground
water to the fullest before drawing upon surface water supplies,
The reason is economic. Ground water will cost less than con-
veyance of water from Cuero and Cibolo reservoirs or from the
Colorado River. Obviously, economics demand that the cheaper
source be used to its limit before resorting to the more
expensive source. The cost disparity is even greater than the
mere conveyance costs suggest since surface water would require
treatment at a cost of additional cents per kilogallon.
But this consideration of using the cheaper source also applies
to the competitive scheme of conventional supply. Even if
water were to be imported this conventional system would make
use of the cheaper ground water up to the allowable limit in
order to reduce the overall cost of the supply. In order to
fairly take this into account in comparing the economics of
reuse versus conventional importation, it is necessary to
158
-------
determine the monthly pattern of ground water use in 2000
which will (a) produce an annual amount equal to the annual
constraint, and (b) avoid exceeding the monthly constraint in
any month, and (c) minimize, the cost of the conveyance and
treatment of the imported water. The last goal involves
maximizing the utilization factor of the pipeline, and also
of the water treatment plant. Utilization factor is the ratio
of the average production to the design capability. The <=/Kgal
cost increases as utilization factor decreases. Therefore,
maximizing the utilization factor minimizes the cost.
Column 8 shows the requirement for import water if the ground
water is pumped so as to just attain the monthly and annual
constraints. Since the intake in the maximum month is fixed
and the ground water contribution also fixed this means that
the import water requirement in that maximum intake month is
also fixed, at 176 mgd. The remaining months of the ground
water withdrawal in Column 7 have been adjusted so as to meet
the constraints and to have no month's requirement for import
water greater than the maximum 193 mgd. Column 8 is the
difference between the ground water and the total intake.
These two curves, (7 and 8), also are shown on Figure 29.
Table 20 summarizes the quantities involved in the conventional
and reuse schemes in 2000. Under the reuse scheme the average
withdrawal of ground water would be 164 mgd and the peak day
435 mgd for utilization factor of .359. The lawn and pipe
losses would be 164 and the amount returned to San Antonio
River zero.
Under the conventional import scheme the average withdrawal
of ground water would be the limit, 192 mgd, and the peak
monthly also the limit, 339 mgd. The average surface water
withdrawal would be 149 mgd. The total withdrawal would be
341 mgd with the peak day 670 mgd for an overall utilization
factor of the system of .51. The lawn and pipe losses would
be 164 mgd and the quantity returned to the San Antonio River
177 mgd.
In the conventional import scheme the load for the peak day
can be thrown toward the ground water or toward the surface
water and in practice this would be done in the direction and
to the extent that produced the minimum overall cost. If all
of the burden of the peak day were thrown on the imported
surface water the peak day for ground water would be the same
as the peak monthly 339 and the remainder of the overall 670
mgd peak day load would be placed upon the surface water,
331 mgd. In that case the utilization factors for the ground
water would be .567 and for the surface water .429. In the
other direction the entire burden for the overall peak day
159
-------
TABLE 20
LOGISTICS OF THE NEW SUPPLY - 2000
Note: Peak day is the 90% level--i.e. expected to be exceeded in only 10% of
the years.
Ground water withdrawal, average
Peak monthly
Peak day
Utilization factor
Surface water withdrawal^ average
Peak monthly
Peak day
Utilization factor
Total withdrawal
Peak monthly
Peak day
Overall utilization factor
Water treatment, average
Peak monthly
Peak day
Utilization factor
Lawn and pipe losses
Sewage treatment or AWT, average
Peak monthly
Peak day
Utilization factor
Demineralization, average (rough)
Peak monthly
Peak day
Utilization factor
Disposal to the Gulf, average
Peak monthly
Peak day
Utilization factor
Discharge to San Antonio River
Storage required
AWT Reuse
Scheme
mgd
164
336
435
.377
none
withdrawn
164
336
435
.377
not
used
164
(AWT)
177
189
350
.506
117
261
486
. 241
7
_ _. .
14
.5
none discharged
yes
Conventional
Import Scheme
mgd
192
339
339/440/494
.567/.436/.389
149
176
331/230/176
.429/.649/.84S
341
515
670
.510
149
176
331/230/176
.429/.649/,845
164
(Conventional STP)
177
189
350
.506
i not
used
not
used
177
no
160
-------
could be placed upon the ground water which would give it a
peak day of 494 and a utilization factor of .389, leaving
176 mgd as the peak day for, the surface water for a utilization
factor of .845. The present project does not as yet go so far
as to determine the proper allocation between the two sources.
Instead, the 1.3 factor which relates peak month to peak day
for the demand characteristics (at the 90 percentile level) is
applied to both the ground water and the surface water resulting
in a peak day for ground water of 440 and for surface water of
230 mgd, with corresponding utilization factors of .436 and .649,
The category "demineralization and by-pass" refers to the
explicit demineralization portion of the recycle scheme. It
has been shown that some demineralization will be required even
in the summer months where the blend is of the better quality
if a TDI much less than 500 mgpl is to be achieved in the blend.
The extent of demineralization need only be such as to produce
in the blend concentrations of the various contaminants which
just pass the blend requirements. Obviously, to reach say 400
mgpl of TDI in July and August from 500 mgpl will require a
lesser degree of demineralization than in December and January
from about 1,000 mgpl. Some demineralization processes, for
example, reverse osmosis, can be operated to achieve various
degrees of demineralization in the effluent. If such a process
is used it would be continuously adjusted to achieve the degree
of demineralization required day-by-day to meet the blend
constraint. Other types of demineralization, for example, ion
exchange, more or less completely demineralize the water and
cannot efficiently be modified day-to-day to do otherwise. In
such cases in order to avoid the economic inefficiency of over-
demineralizing the quantity demineralized would be varied by
by-passing some of the AWT effluent around the explicit demin-
eralization stage. The determination of the exact amount of
by-pass which with a given discharge is allowable in order to
just meet the blend constraints is the purpose of the RECYCLE
program, not yet completed. For a rough approximation to their
quantities, see the demineralization section in Chapter 6 of
this series. Table 20 merely indicates that the requirements
for explicit demineralization cannot be greater than the figure
given, but it may be less.
Table 21 provides some details on the treatment plant require-
ments. The combined peak day capability of the three existing
or planned San Antonio conventional sewage treatment plants,
the Rilling, the Leon Creek, and the Salado will be 116 mgd.
The corresponding average flow handled by these plants under
the seasonal pattern described will be 59 mgd. Therefore, the
new capability required in 2000 will be 234 mgd which will
handle an average flow of 118 mgd. This may be compared with
the requirement for the AWT plant, from Table 20, of 350 mgd
capability, and average flow of 177 mgd.
161
-------
TABLE 21
LOGISTICS OF THE NEW SUPPLY - 2000
Sewage and Water Treatment Plant Alternatives
mgd
Existing or U.C . STP (3 plants)
Peak day capability
Average
New capability required
Peak day
Peak monthly
Average
Utilization factor
AWT
Reuse
Scheme
not
used
(AWT)
350
189
177
.506
Conventional
Import
Scheme
116
59
(STP)
234
-
118
.506
Discharged to San Antonio River
none
177
Demineralization and by-pass*
Peak day
Peak monthly
Average
Utilization factor
350
189
177
.506
not
needed
* Demineralization required cannot be greater than this
Water treatment
Peak day
Peak monthly
Average
Utilization factor
not
needed
331/230/176
176
149
.429/.649/.S45
162
-------
Table 22 provides some details on the ground water facility
requirements, using the utilization factors characteristic of
the AWT reuse scheme and the conventional import scheme from
Table 20. With these utilization factors the average pro-
duction from the existing ground water facilities would be
125 mgd for the AWT reuse scheme and 145 for the conventional
import scheme/ since the latter would operate at a higher
utilization factor. The new facility required would be 102
mgd and 107 mgd, respectively, of firm capability an average
production for the new facility of 39 and 47 mgd, respectively.
TABLE 22
LOGISTICS OF THE NEW SUPPLY - 2000
Ground Water Facility Alternatives
mgd
AWT Conventional
Reuse Import
Scheme Scheme
Existing GW facilities
Peak day firm capability 333 333
Average 125 145
New facility required
Peak day, firm 102 107
Average 39 47
Utilization factor .377 .436
Table 23 shows the conveyance alternatives. The AWT reuse
scheme in the single AWT plant embodiment would require the
conveyance back from the AWT plant illustratively at the Rilling
site to the water distribution system illustratively taken as
the Hildebrand tank in the north part of the City. This pipe-
line conveyance system would have a peak day of 350 mgd and an
average of 177.
For the conventional import scheme, several alternative sources
are available. One of these is the Cuero Reservoir supplemented
by the Cibolo Reservoir. Since Cuero is in the Guadalupe Basin
the Texas Water Plan calls for a "reimbursement" of the Guada-
lupe Basin by the San Antonio Basin by a transfer from Goliad
Reservoir to the Guadalupe in the neighborhood of Victoria.
This conveyance is part of the Cuero-Cibolo system and part of
the cost. Another alternative is the conveyance of water from
the Colorado at Austin much of which would be by canal. A
163
-------
preliminary engineering study has been performed on these two
alternatives (36). A still more recent alternative involves
an Applewhite Reservoir on the Medina River which is now being
studied by the City Water Board. The Applewhite Reservoir is
not in the published versions of the Texas Water Plan. Other
alternatives, some of which have received engineering study,
include Canyon Reservoir, and Cloptin Crossing Reservoir.
TABLE 23
LOGISTICS OF THE NEW SUPPLY - 2000
Conveyance Alternatives
mgd
AWT Conventional
Reuse Import
Scheme Scheme
Cuero and Cibolo to Hildebrand
Peak day 0 230
Average 0 149
Goliad to Victoria
Peak day 0 141.4
Average 0 113.1
Rilling to Hildebrand
Peak day 350 0
Average 177 0
When engineers come to a final decision on one of these or
other alternative conventional supplies they will take into
account the engineering, economic, and political considerations
which govern such choices. The politics and the emotions
surrounding alternative Texas water schemes can become heated.
Our selection of the Cuero-Cibolo alternative for the conven-
tional supply in this study should be taken as purely illustrative
and not a recommendation for that scheme as against any other.
Our proposal for this recognized that alternative conventional
water supply schemes are multifarious and this project could
not hope to be instrumental in selecting the best of them for
any particular city. Such a selection even for a single city
would in general require more funds than allotted to our entire
project, the main purpose of which was to obtain a methodology
for a comparison between reuse and the best alternative conven-
tional supply. San Antonio was selected as the practical
situation on which to explore the methodology and when a final
decision is reached by other parties as to the best of the
conventional supply alternatives the corresponding logistical
and cost data can be plugged into the model for a comparison.
164
-------
Among the available alternatives the Cuero-Cibolo scheme was
chosen for this illustrative comparison for the following
reasons. The quantities, the reservoir yields, the distances,
and the reimbursement requirements were clearly set forth in
the Texas Water Plan (37). For the Lake Austin and some of
the other alternative schemes the reimbursement feature under
the Texas Water Plan was not clear. The Lake Austin convey-
ance system would be largely by canal with seepage losses
and costs thereof unknown. We believe that the future water
supply of cities will be conveyed mostly by pipeline, only to
a small extent by canal, and our project did not yet cover
canal conveyance models. The Applewhite scheme had not been
sufficiently formulated to use in our illustrative model.
165
-------
THE NATURE OF THE RECYCLE PROBLEM
One of the major alternatives to have been studied in this
project is recycle and reuse, to as high a degree as possible,
or at least to as high a degree as economic. It was therefore
necessary to come to quantitative grips with the problem that
is usually swept under the rug in discussions on reuse. Most
discussions of reuse are content implicity or explicity to
consider a single reuse, relying on some treatment process to
produce a water for reuse that is acceptable, and in some
components not too much worse than the original starting water.
This is fine for a single reuse, but if one is considering
indefinite reuse, these components in which the return is only
a little worse than the original water must build up in the
recycling water and ultimately become intolerable. The standard
reply to this extension of the problem in turn is that the worst
liquor, in our case the secondary effluent, will be purged from
the system to maintain the return at an acceptable level in all
components. This is where the problem is usually left. We
believe no one has ever worked out a quantitative balance for
a real recycle and reuse process and particularly determined
whether or not the purged quantity might not be so great as to
leave little for reuse.
The problem turns out to be very complicated mathematically and
in the real application is further complicated by the high
seasonality of the loss ratio. At one stage of the study it
appeared that the problem was one to which no solution existed.
The project now has demonstrated that a solution does exist
but we have not had enough time to complete the solution (this
was but one of 22 tasks under the project). This chapter out-
lines the nature of the problem, the boundaries and constraints
in it, and the method by which the solution is to be achieved.
The problem is stated as follows:
In a system of recycle for reuse having water for
use as a blend of makeup water and recycled water,
having a contaminant increment attendant upon use,
having losses in use not returned to the treatment
plant (lawn losses), having losses in transit of
waste not returned to the treatment plant (pipe
losses), having conventional sewage treatment,
having advanced waste treatment with some attendant
demineralization, having explicit demineralization
and allowing some by-pass thereof and having
discharge and disposal.
166
-------
GIVEN: makeup water quality in N contaminants, criteria
(maxima) for water quality in use in N contaminants,
municipal concentration increment in N contaminants,
quantity of pipe losses, quantity of lawn losses,
any set of treatment and advanced treatment processes,
any set of explicit demineralization processes,
criteria for water quality of the discharge in N
contaminants, and any set of disposal processes.
DETERMINE: for any given quantity of effluent discharged what
quantity of the recycle must be demineralized (and
what quantity by-passed) in order to maintain in the
blend water and the discharge a steady state concen-
tration meeting the criteria in N contaminants.
A Glimpse at the^ Solution
The work toward the solution of this problem has revealed the
following. The answers to the problem comprise the quantities
to be demineralized and the quantities to be discharged. For
any given set of conditions, water use quantity, losses,
municipal concentration increment, characteristics of treatment
and demineralization processes, etc., there may be or there may
not be a feasible solution. If there is, there are a set of
discharge quantities that will allow a solution which just fails
to violate the criteria. This set of values is continuous and
has upper and lower limits which are non-trivial, i.e. do not
merely comprise zero percent and 100 percent of the sewage
quantity. For any of the allowed discharge quantities there
are two demineralization quantities which will satisfy the
criteria. Thus, the solution comprises a set of "pairs" of
discharge and conjugate demineralization quantities which are
inter-determined...that is if one is chosen the other two are
fixed. The set of demineralization quantities is also con-
tinuous and bounded non-trivially. The pairing relation is
reciprocal, that is if the demineralization quantity is chosen
its paired discharge quantities are fixed/ and likewise, if a
discharge quantity is chosen its paired demineralization
quantities are fixed. Finally, each pair of solution quantities
has an associated cost for the entire system and in general one
of these pairs shows a cost lower than all others, i.e. is an
optimum.
167
-------
Linear Programming Model
With this much information those familiar with the field will
assign this as a linear programming problem. In the simplified
form it is indeed a linear programming problem. In the real-
world form it is not a linear programming problem, but we shall
discuss it first in the simplified form to lay the groundwork
for the real problem. Figure 30 is a flow chart of a simplified
municipal recycle scheme showing quantities of water and con-
centrations and quantities (ppmmgd) (mgpl times mgd) of a
particular ion. A subscript j is to be considered as applying
to concentrations and ion quantities as well as leakage. The
flow diagram has been simplified by assuming that no mineral
change or mineral addition occurs in the conventional STP or
AWT processes. Also, no water occurs as waste or backwash
from the demineralization process; the mathematics is simplified
to make the leakage independent of the feed concentration and
to place no constraints on the discharge quantity or concen-
tration.
It is seen that the overall input is the makeup water M and
the municipal increment. The output is the lawn loss L, the
pipe loss P and the discharge M-L-P. The amount demineralized
is X.
There are two overall material balance relations on water
quantity and ion quantity. One of these comes from the over-
all input-output balance. The other comes from the two
possible computations for the ion quantity in the return,
one from a backward computation and one from a forward com-
putation. Both of these relations yield the identical equation
which is the basic material balance equation of the problem:
Mm + Bi = M + X (1 - f) (b + i) (1)
To make it easier to discuss the problem we shall replace this
equation with another obtained by dividing through by B, retain-
ing now the symbol M to represent M/B and X to represent X/B.
The equation then becomes:
Mm + i = M + X (1 - f) (b + i) (2)
In solving this to meet a given BLEND quality, p^ and ignoring
any discharge specification the constraints are:
0< (X+M) < l
0 < M < 1
0 < X < 1
0
-------
Makeup watei
M, m, Mm
SIMPLIFIED FLOW DIAGRAM
FOR ILLUSTRATING
MATHEMATICS OF RECYCLE
BLEND
B,b,Bb
RETURN
B-M, - , Bb-Mm
also
Lawn
L,b,Lb }
/
Increment
0,i,
' i (B-L)
SEWAGE
B-L. b+i
(B-LJ (b+i)
P, b+i
Sewage Delivered
B-M.W),
(B-M) (b+i)
By-pass
B-M-X, bfi
Demin. feed
X(bH)
X, f(b+i)
Xf (b+i)
mgd
^^—ppm
Mm).
<— ion quantity
ppmmgd
f, leakage = Product concentration
feed concentration
Conventional STP
M-L-P, bfi
(M-L-F) (b+i)
1 —
product
i)
\
/
Demin. waste
0, - ,
X(l-f) (b+1)
j Discharge 1+ ^'^ (b+i)
'(M-L-P + XJl-f )) (b+ i)
169
Figure 30
-------
Basic Linear X, M Relation
Tlie linear relation is between X and M:
M(m-b-i) + i
* ~ (1-f) (b+ i) (3)
x/r X(l-f) (Mi) - i
M ~ (iii-b-1) (4)
For each ion there is a value of X which will just meet the p constraint on that
ion in the blend . Call this S.
^(m.-p.-9-f-i. (5)
J
(1-f.) (p. +i.)
3 J J
At any given M value the highest of the S.Ts is of course the X that must be used.
* J
Call this S . The blend concentration of that ion will just meet the p constraint
for that ion, and the concentrations of all the other ions will be less than their
P constraints, according to:
, Mm. + i, . _ /£.
b = ^c * - (6>
where k is the subscript representing any of the other ions.
A schematic linear programming diagram for X and M is shown in Figure 31.
170
-------
M = L + P
Feasible region for X. and
M at given P; , irtj,
L, and f. independent of
composition and concen-
tration
X line
Figure 31
171
-------
In this diagram on the simplified model with three ions it is
seen that X is a linear function of M for each ion. The con-
straints that X and M must lie between zero and one are shown
on the axes; the constraint X + M = 1 is the hypotenuse. The
XI line represents the relation for ion number 1. Having been
drawn it divides the diagram into two regions; that above the
line being allowed, that below the line being disallowed
because X, M pairs in that region would not achieve the Pj
criterion. The X2 line when drawn similarly excludes an area
below it, in the sense of at lower X's. Similarly the X3 line.
The X3 excluded region, however, has no effect because it is
over-ridden by the line representing the constraint that
M = (L + P) .
The Feasible Region
Thus the onlv region in the entire field which is feasible for
meeting the p constraints on the various ions is the unshaded
polygon. Any other combination of X and M would fail to meet
the blend criterion for at least one of the ions.
Note that it is not necessary that there be any feasible region
at all. For example, if L + P were quite high, higher for
example than the intersection of the XI line with the hypotenuse,
then there would be no feasible region, which means that there
is no possible combination of X and M which would meet the
blend constraints. In other words, no real system could hold
the blend concentration which is specified.
Non-Linearities
So far this has appeared to be a straight linear programming
problem in which the next step would be to impose a linear cost
surface on the diagram, and to recognize that the lowest cost
feasible solution must occur at one of the apexes. However, the
real situation is not amenable to simple linear programming
because the system departs from linearity in at least three ways.
First, the cost surface is not planar, and therefore the con-
tours representing that surface on the linear programming
diagram are not linear. In general this means that the mini-
mum cost solution is not necessarily an apex and could even be
in the interior of the feasible region. The real situation is
not as recalcitrant as this, however, and it will shortly be
shown that the minimum cost solution lies on one of the feasible
region boundaries.
172
-------
Second, the linearity of the X, M lines depends on the
assumption that the leakage, f, is independent of the com-
position and concentration of the demineralization feed. In
most demineralization processes this is not strictly true (with
some it is strongly untrue) and this non-constancy of f makes
the X, M lines non-linear.
Third, in some demineralization processes the concentration of
some ions in the product must depend upon the combination of
concentrations pf its companion; product ions. For example,
in ion exchange the concentration of Na in the product is
determined not by a leakage unique to itself but by the
leakages of the other ions since its leakage only occurs
in maintaining ionic balance. This not only makes the line
for Na non-linear but it also makes non-linear the line for
total dissolved ions which can also be one of the /3 constraints.
Ions Not Amenable to Demineralization
Some demineralization processes fail to remove any of certain
ions, e.g. SiCK in weak base anion exchange. For such ions
(1 - f) becomes zero, and the overall material balance
equation becomes:
Mm + i = M (b + i) (7)
For each ion there is a. value of M which will just meet the (3 constraint on that
ion in the blend. Call this M.:
M = -i./(m - P- -U (8)
J J J J «J
For any set of ions there is one .ion which will give the highest
of the M-; values. Call this value M*, which is the lower bound
on M to meet the blend criteria for all ions for which (1 - f)
approaches zero. This boundary is a vertical line on the linear
programming diagram, which may be above or below M = L + P.
The highest of the several lower bounds, M* or M = L + P is
the controlling bound for M and sets a minimum M. Of course,
it is understood that these vertical boundaries might be
completely over-ridden by the .exclusion fields of some of the
other ions and so not come into consideration. (However, the
existence of these vertical lines determines a strategy in the
computer solution.)
173
-------
The Cost Surface
While the cost surface is not planar and thus it is not possible
to place the optimum at an apex, nevertheless it is possible to
make some general statements about the cost surface which can give
a further clue as to the solution. The following chart shows how
the costs of the various cost components change as M increases at
constant X and as X increases at constant M.
Increase in Cost No Change Decrease in Cost
As M increases at constant X:
New water Demineralization AWT
Sewage treatment Disposal (of demin. Conveyance back to
Disposal waste) point of use
(of discharge)
As X increases at constant M:
Demineralization New water None
Disposal Sewage treatment
(of regeneration waste) AWT
Disposal (of discharge)
Conveyance back
This shows that as X decreases at constant M no cost component
increases, some remain unchanged, but two components decrease.
Therefore, at constant M as X decreases, total cost decreases. It
follows that no point in the interior of the feasible region can be
economic over a point vertically beneath it at a lower X value. The
minimum cost will be found on one of the boundaries of the feasible
polygon but not on a boundary which has any other boundary vertically
beneath it. To illustrate, the minimum cost will be found somewhere
along the line segments A, B, C in Figure 31. The mathematical
description of the situation is:
(dCOST/dX) > 0
Unfortunately, the relation for the change of cost with M at constant
x cannot yet be so simplified. We do not yet know the nature of the
surface in the M direction since it contains both positive and
negative partial derivatives.
174
-------
No Demineralization
The relations in certain reduced forms of the recycle pattern are
of interest. With no demineralization X = 0 and for all ions the
relation is as for a single ion with f = 1. The makeup necessary
to maintain a given blend constraint for a given ion is as in
equation (8). The Mj's are the intersections of the X vs. M lines
with the X axis. The highest of the Mj's, called M*, is the
controlling M and the concentration of all the other ions is given
by:
b. = m. - i. +i. /M* (9)
No Discharge
With no discharge M is fixed at M* = L + P, and the X necessary to
achieve the blend constraint for each ion is given by:
x. =
J (1-fKP -i)
J J
The highest of the Xj's, called!*, is the controlling one, i.e.
the demineralization required to maintain a given blend constraint,
and the concentration of each of the other ions is given by equation
(6).
No Discharge and No Demineralization
With no discharge and no demineralization X = 0 and M = L + P. This
is the only value that M can have, called M* , and^ the blend concen-
tration of each ion will be given by equation (9)'. This is the
relation on which Figure 18 is computed. If any of the bj's is
greater than the corresponding & the blend constraint cannot be met
for that ion and the constraint must be relaxed or else discharge or
demineralization must be allowed.
This can be re-expressed :
(b-m) -K
This states that the increase of the blend concentration over the
makeup water concentration in any ion depends only on the municipal
concentration increment and on the loss ratio.
175
-------
The Real Recycle Pattern
The real recycle pattern is more complicated than this simplified
form in at least the following ways:
There are some 50 likely contaminant ions that should be
explored for buildup. Possibly any ion may prove to be the
controlling one.
The leakages are functions of the concentration arid composition
of the demineralization feed.
Some water may be used in the backwash/ regeneration and
reject from the demineralization process.
Disposal of the demineralization waste separately from the
discharge may be indicated.
The AWT process causes changes in the ionic concentrations,
indeed some ions are actually added to the water in the
AWT process comprising an additional input.
Where there is no existing conventional plant it may be
desirable to discharge at an intermediate stage in the AWT
process, e.g. prior to the carbon stage, or this may be
so even if there is an existing conventional plant.
The presentation has been in terms of a fixed usager B,
lawn loss L, pipe loss P, and sewage delivered (B - L - P).
Actually B'and L are subject to seasonal variations. The
physical system must be capable of handling the worst
conditions and the costs to be optimized are the summations
over the year under the fluctuating seasonal conditions.
Possibly the municipal increment also has a seasonal
variation.
The nature of the seasonal variations does not allow the
picking of a particular time instant, i.e. a particular month,
as containing the extremes for design. The highest requirement
for hydraulic flow occurs in July and August but the highest
requirement for demineralization occurs in December-January.
Thus, for example, the makeup water system must be sized on
the July-August flow but the demineralization equipment must
be sized on the December-January flow.
176
-------
Status o_f the Computer Program RECYCLE
The reader who has waded through the increasing degrees of complexity
in the foregoing exposition may now understand why so many discussions
on recycle stop at "let's treat the sewage to make drinking water out
of it and put it back into the mains."
Under this project we have developed several computer programs for
solving the recycle problem in increasing degrees of complexity.
None of the ones that are completed are close enough to reality to
be worth reporting. The one that is close enough (and still a long
way from design reality) is not complete, and cannot be completed
within the time schedule of the project.
That program handles the following degree of complexity. About 20
of the major ions plus COD are considered. The quantities are fixed
and are not seasonally varied. The physical system comprises a
single makeup source, a single AWT plant, a single conventional sewage
treatment plant, and a single explicit demineralization plant. No
constraints are placed on the discharge (in dissolved ions). The
program only determines the feasible boundary in the X, M field and
does not as yet find the optimum pair for minimum cost. During the
development the program uses surrogate AWT and activated sludge sub-
routines for simulating AWT performance and conventional STP per-
formance but the full programs are available in Chapter 2 and
Chapter 5. The program uses a surrogate subroutine for approxi-
mating the performance of a demineralization process. Programs are
available for reverse osmosis ion exchange and electrodialysis, but
none of these are in shape for immediate insertion into the recycle
program.
As for the current program the mathematics have been worked out and
reguire checking. The computer program itself has been roughly flow-
charted but not written.
Despite the incompleteness of our work on the recycle problem, we
believe that the most important accomplishment of the project has
been the demonstration of the nature and the complexity of the
recycle problem.
177
-------
CHAPTER 4
COMPUTER PROGRAM FOR DESIGN AND COSTING
OF CONVEYING WATER BY PIPELINE
This Chapter covers the development of a computer program for
the design and costing of pipelines and the conveyance of
water or other fluids through them. Water conveyance is one
of the important elements bearing upon the economic competition
between conventional water supply and wate treatment typically
by importing water from remote sites, and renovation or reuse
by advanced waste treatment.
WHAT THIS PROGRAM DOES
This program takes the specified characteristics of the
conveyance situation, designs a pipeline which will minimize
the cost of conveyance in that situation, and returns the
design data and the cost breakdown. The special details
are as follows.
The line is designed in segments (up to three) as may be
specified. As the program now stands each segment may have
its individual mileage, beginning elevation, ending eleva-
tion, terrain factor (a factor concerning the cost of line
maintenance) and construction factor (a factor concerning
a construction cost). The program optimizes each segment
and returns the design characteristics of each segment
and the cost of the entire line.
The input quantities conveyed, obviously the same for each
segment, are: QMAX, the mgd on the maximum day in the design
period; QBARE, the expected average mgd over the entire
design period; and two quantities not yet used in the program,
the actual average day as distinguished from the expected
average day, and the minimum day. The maximum and average
gph and gpm rates are taken to be 1/24 and 1/1440 times
the mgd rates.
The program optimizes each segment for the conveyance of
QBARE in a facility which has the firm capability of
conveying QMAX.
Within each segment having multiple pump stations the program
designs with equal-sized stations, a design which if it
could be achieved in actual practice would minimize the
cost over unequal-sized stations. The program generates
a firming factor which computes installed pump station
horsepower from firm pump horsepower.
179
-------
Given the state and the region in which the bulk of the
line lies and the future year for which the estimate is
desired the program then generates the necessary cost indexes
for the future year and for the region. If the
-------
WHAT THE PROGRAM DOES NOT DO
Some of the items mentioned in this section are discussed
more fully in the text., All of them possibly might better
appear in a discussion and recommendation section following
the text. However, they are placed here in order that
the reader may peruse the rest of the text with the fore-
knowledge of what the program does and does not do.
The program uses the annual cost method rather than.the
present value method partly because it is simpler and partly
because many of the programs with which this one will be
tied are couched in the annual cost and C/Kgal terms rather
than the present value terms.
The program does not compute the costs for each year's
production as it occurs, presumably under some growth
pattern. Instead, it computes a cost as if each day's
flow were the average flow over the entire period, namely
QBARE. Other studies of the authors have shown that this
produces a cost which is lower than the true cost, but
not much lower.
The program does not adjust for inflation during the project
life. It computes all costs in "current year dollars"...i.e.
if IYEAR is set to 1980 all costs will be in 1980 dollars.
The program does not stage the construction of facilities.
It assumes that all facilities are constructed in a given
year, the "current year," and of such a size as to meet
the requirements in the target year, in this project 2000 A.D.
Pipe sizes and pump station horsepower are treated as con-
tinuous functions, not discrete functions as they are in
actuality.
It is assumed that in any segment the hydraulic gradient
created by equal size pump stations will at every point
be higher than ground elevation. This might not actually
be the case if the profile is not of constant slope through-
out the segment. The most obvious of such violations is
accounted for by the stipulation that no segment may have
an intermediate high point which is higher than both the
beginning and the ending elevation. This serves to break
the pipeline into segments which are less likely to have
the hydraulic gradient intersecting the ground elevation.
Even in this case, difficulties are encountered when the
segments consist of a relatively short lift segment followed
by a segment with a small negative slope. This is more
fully discussed at the proper point in the text.
181
-------
The program does not take into account the higher pressure
class of pipe which would be required at the bottom of
a U-shaped profile, nor does it assign a pressure class
other than 100 psi for gravity lines.
182
-------
SYSTEM DESIGN
Pump Station Design Computations
The philosophy of the design model is to achieve the necessary
total horsepower for pumping by using equally-sized pump
stations. Previous work (38-44) has shown that for pump station
costs, as well as most other investment costs, the lowest cost
is achieved by using equal-sized units. Any departure from
this so as to have unequal-sized units results in a higher cost.
Real pipelines, of course, cannot achieve exactly equal-sized
pump stations and to the extent this is not achieved real costs
will be somewhat higher than those computed by the model.
Under this philosophy the pump station computations are as
follows:
FRHDOT = 318.4346*FDOT*QDOT**2/DIAM**5
TDHDOT = FRHDOT+SLOPE
NUMSTA = (TDHDOT*PMILE/HDLIM+.99999)
PUMILE = PMILE/NUMSTA
HDSTA = TDHDOT*PUMILE
HPSTAF = 0.175615*QDOT*DENS*HDSTA/EFF
HPSTAI = HPSTAF*FIRM
where:
FRHDOT = Friction head at design capability, feet of fluid
per mile
FOOT = Moody friction factor at design capability
QDOT = Design capability, mgd
DIAM = Inside pipe diameter, feet
TDHDOT = Total dynamic head, feet of fluid per mile
SLOPE = Uniform slope of pipeline, feet/mile, (elevation
difference/pipeline miles)
NUMSTA = Number of equal-sized pump stations (truncated to
"the least integer not less than")
PMILE = Pipeline length, miles
HDLIM = Head limitation on pump station, feet of fluid
PUMILE = Interstation distance, miles
HDSTA = Head per, station, feet of fluid
HPSTAF = Firm horsepower per station
DENS = Fluid density, gm/ml
EFF = Wire to water efficiency, fraction
HPSTAI = Installed horsepower per station
FIRM = Firming factor, ratio of installed horsepower
to firm horsepower
183
-------
The Moody friction factor, or rather the Moody correlation of
the Darcy friction factor (45), is computed by a subroutine
MOODY developed for this study, which generates the MOODY
friction factor, over most of the range according to the
Colebrook and White formula, and also generates the Reynolds
number and the flow type, whether turbulent, transition, critical,
or laminar. Since the Colebrook and White formula is non-explicit
for the friction factor the subroutine uses an iterative proce-
dure for solving the equation. The other parameters required
in addition to the flow rate and the diameter are viscosity,
VIS, and absolute roughness, EPS. The absolute roughness used
in the exemplary computations is 0.0003 feet, corresponding to
new or fairly new smooth concrete average workmanship or hot
asphalt dipped or centrifugally applied concrete lined steel
pipe, continuous interior butt welded (46).
The equations for density and viscosity given in the program
listing cover the range from 4 to 36 degrees C (centigrade).
The density equation exactly reproduces a five-place density
tabulation with standard error of estimate being about 3x10~^.
The viscosity equation has a standard error of estimate of
about .0057 in millistokes, corresponding to about 0.05%. It
is converted to viscosity in feet ^/second.
The wire-to-water efficiency used in the exemplary computation
is 0.75.
The firming factor is taken as 2.0 at QDOT 1.0 or less, and
1.25 at QDOT 10.0 or more, Cartesian linearly interpolated
between.
Types of Conveyance Situations
Depending upon the pipeline slope and the variations in required
daily flow there exist four distinguishable types of conveyance
situations. With the range of daily flows from QMIN the minimum
to QMAX the maximum, whether or not pumping is required on a
given day, that is at a given flow, depends upon the relative
magnitudes of the friction head at that flow and the pipeline
slope. The total head loss, feet/mile, is the algebraic sum
of the friction head loss and the slope, both in feet/mile.
Consider the changing situations "as a high positive slope is
continually decreased. At any slope above zero pumping will be
required on every day, that is at QMIN as well as QMAX, and the
situation is termed "pumped." As the slope continues to decrease
through zero and becomes slightly negative there is no change,
in that pumping is still required on each day but the pumped
flow is assisted by the gravity gradient and this situation is
termed "pumped, gravity assisted" or "assisted pumped." Mathe-
matically this is in no way different from the pumped situation.
However, as the slope continues to become more negative it
eventually reaches a condition at which the absolute value of
184
-------
the slope is greater than the absolute value of the friction
head on the minimum day. Thus, on such days the sum of the
friction head and the slope becomes negative and the energy
requirement is zero. The conveyance situation under this
circumstance is that gravity alone is adequate to convey the
required flow on some days but not on all days and the
situation is termed "gravity boosted." Finally, as static
head continues to become more negative it reaches some magni-
tude such that absolute value of the static head is greater
than the friction head even on the maximum day, i.e., on the
design day. Beyond this pumping is not required on any day
and indeed pump stations are not required. The energy con-
sumption is zero and the situation is called "gravity."
Table 24 shows some of the characteristics of these types of
conveyance situations as defined by the indicated relationships
between slope and friction head, where:
FMIN = Friction factor for the minimum flow
FDOT = Friction factor for the design flow
QMIN = Minimum daily flow, mgd
QDOT = Design or maximum daily flow, mgd
DIAM = Inside pipe diameter, feet
For the gravity situation optimization is not required since
the lowest cost is obtained at the pipe diameter which will
make the friction on the maximum day just equal to the negative
of the slope. Thus:
DIAMG = ( (318.4346*FDOT*QDOT**2/(-SLOPE))**.2
where:
DIAMG = Diameter of the smallest line that will just
suffice on the maximum day
For the other three situations optimization is required.
It will be found that for the pumped and pumped gravity assisted
conveyance types the optimum diameter of the pipeline is practi-
cally independent of the slope. The very small dependency that
does occur results from the somewhat erratic effect of pump
station horsepower on pump station OMR as used in the program.
In the gravity, boosted conveyance type if the slope lies in
the range between the QMIN and the QDOT term (i.e. if the slope
is not simply the negative of the QMIN term), then there will
be days in which the friction head is less than the negative
of the slope, and on those days the TDH becomes negative. Since
the energy term in the energy summation is proportional to TDH
these days would appear in the summation as negative energy days,
Of course, the correct energy consumption in such days is zero
and such TDH values must be returned to zero.
185
-------
TABLE 24
TYPES OF CONVEYANCE SITUATIONS
Range of Slope
Type of
Conveyance
Optimization
Required
Is Optimum Pump
Diam . Stations
Dependent
on Slope?
Cost
Equation
+ 00 tO 0
°oto , QM
-318. 4346 FMIN-^
Ui-n
31S AlAf* TTKATM ^"^
J18.4346FMIN
t0 QE
-318. 4346 FDOT -^r;
-318. 4346 FOOT -^7
to - co
Pumped
^2 Pumped,
—5 Gravity
assisted
N2
M Gravity
)OT_ boosted
AM Gravity
Yes
Yes
Yes
No
No Yes
No Yes
Yes Yes
Yes No
As written
As written
Modified*
Modified*
* Modification consists of replacing negative total dynamic heads with zero in the
summation term.
5280
318.4346 =
.6463229 * 2g (
.6463229 = conversion cfs to mgd
g = 32.17398, standard gravity
186
-------
The relations shown In Table 24 are explicit and give sharp
demarcations between conveyance types for a given diameter.
However, the problem is to determine which coneyance type
is the cheapest at the optimum diameter. For the trivial
transition between pumped and pump assisted the decision is
clear since it merely involves whether the slope is positive
or negative. However, for the other two transitions between
gravity and gravity boosted and between gravity boosted and
pumped assisted, under certain conditions even high precision
computer optimization breaks down in making the decision. Under
these conditions the present program utilizes a small area of
tolerance in making an arbitrary decision and does not arrive
at a mathematically precise transition. However, the area of
tolerance is so small as to be inconsequential in the practical
application.
Q Variable and Q Constant
The discussion up to this point has involved the real situation
in which the daily flows are varying. However, this would in-
volve an integration over the varying flows in computing the
energy cost. To avoid this degree of complexity the present
program substitutes for the real situation a simplified
situation in which the flow on each day is held constant at
the average flow value, QBARE, where QBARE is the average flow
expected over the project period. However, the pipeline is
designed so that the system can achieve the design flow, QMAX.
Since the energy cost is proportional to the cube of the flow,
the cost for the simplified model with Q constant will always
be less than the cost computed with the real model with Q
variable. However, the authors (40) have tested this approxi-
mation by comparing the costs for a Q constant model against
those for a Q variable model in which the variability is among
the highest occurring in real water conveyance systems. It
was found that the Q constant model produces costs which are
for most slopes within 5% of the extreme Q variable model. The
discrepancy reaches as high as 10% at slopes in the vicinity of
zero. As slopes decrease in the direction of the gravity line
the discrepancy decreases until it vanishes for the gravity
line, since the energy term drops out. Likewise, as slopes
increase to high positive values most of the energy cost becomes
that for overcoming the static head and the discrepancy again
approaches zero. Most water conveyance variabilities do not
approach the extreme used in the comparison and the simplified
model accordingly provides a satisfactory approximation for the
intended purposes. (It is intended later to incorporate the Q
variable model in the program.)
187
-------
In terms of the simplified model then the conveyance types
have the following strict meanings. For the gravity situation
there are no pump stations and no energy is expended on any
day. For the pumped and pumped assisted situations there are
pump stations and energy is expended on each day, the energy
being that required for a flow of QBARE. For the gravity
boosted situation pump stations are required (in order to have
the capability of meeting the maximum day) but no -energy is
expended on any day. The pump station capital charge and the
pump station OMR costs are incurred.
A common occurrence in pipeline profiles comprises a rather
short segment to an intervening high point followed by a longer
segment of negative slope. If it should turn out that the
negative slope section is optimum as a gravity line while the
positive slope section is, of course, the pumped type then the
program accepts that situation. However, if the negative slope
segment should turn out to be a pumped assisted or boosted line
then the program provides a small pump station for the positive
slope segment and one or more pump stations for the negative
slope segment. In that case it would in general prove cheaper
to consider the line as a single segment in which the hydraulic
gradient at the beginning is great enough so as to exceed the
elevation of the intervening high point at that high point.
This would make the pump stations of equal size in the program
design and more nearly of equal size in the real design.
The present program does not explore that alternative since
it is one of a number of problem situations relating to the
proper gradients for and segmenting of pipelines for real
terrain profiles which hopefully may be tackled in a more
detailed future version of this program. Meanwhile for such
profiles as described above the costs generated will be in error
by being slightly too high.
Optimization Strategy
The strategy used in selecting the cheapest pipeline is as
follows. If the slope is steeper than -50 feet/mile it is
judged that there is practically no chance that a boosted line
would be economic and only the gravity line is computed. If
the slope is greater than zero feet/mile only a pumped line will
suffice and only a pumped conveyance type is computed. If the
slope is negative between 0 and -50 feet/mile the gravity line
is first computed and then the line with pump stations is com-
puted, a process involving optimization and which may result in
a boosted or an assisted conveyance type. If the optimization
search does not find a cost less than 110% of the gravity cost
in eight iterations it is concluded that the optimization is
homing in on a value which cannot be as low as the gravity line.
188
-------
Accordingly, the search is terminated at eight iterations
and the gravity line selected. If the optimization search
producer any cost less than 110% of the gravity cost there
is judged to be a possibility that the ultimate optimum will
be less than the gravity cost and accordingly the optimization
search is continued to the convergence. If the so-located
optimum shows a cost less than the gravity line it is selected.
189
-------
COST RELATIONS
Pipeline Investment
A concurrent study (47) correlates the investment costs trended
to 1968 for some 825 pipelines and presents equations and cost
index factors by which the investment in a pipeline can be
estimated for any diameter and any of the 21 regions. It is
shown that there are large regional differences in costs of
pipelines which must be taken into account in estimating the
cost.of conveyance. The equations reduce to the three relations
in Statements 300-304 of the Program which also include the
regional and temporal cost adjustments and the special regional
pipeline cost adjustment.
The erratic for the correlation is approximately 1.3. A corre-
lation is also given for the cost of offshore pipelines, but
this is not included in the program since very few water con-
veyance systems will be submarine.
The parameter CONSFAC (construction factor) is provided to permit
an engineering judgment on the deviation of estimated costs for
a particular installation from the median costs given by the
equations. Thus, for example, setting CONSFAC at 1.3 will pro-
duce a cost which is on the average exceeded by only 16% of the
pipelines in the basic data. The user is cautioned against
using the construction factor intuitively as a regionalization
factor. The cost index system already takes into account the
fact that pipelines in the Boston region cost 2.5-3 times as
much as those in the Denver region. The CONSFAC is to be used
to adjust for costs which are atypical within a given region.
The user is also cautioned against over emphasizing the right-
of-way costs in setting a CONSFAC. As shown in the reference,
right-of-way costs in general are but a small portion of total
pipeline investment. If right-of-way costs were increased ten-
fold over the average the cost of pipelines would only be
doubled over the average.
Only a small fraction of the mileage represented by the 825
pipelines in the basic data occurs in urbanized territory, so
that overall the construction factor should be greater than
1.0 for water lines in urbanized territory. But in urbanized
regions as, for example, the Boston region, a greater fraction
of the pipelines in the basic data occur in urbanized territory
as compared with an open country area such as Denver or Atlanta.
Accordingly, the construction factor for urbanized territory
for the Boston region should probably be lower than would be
the construction factor for urbanized territory in the Atlanta
region. The authors can give no firm guidelines for construction
factors in urbanized territory. However, a construction factor
of 2.0 represents a cost which is exceeded by only about 8% of
the pipelines in a given region.
190
-------
The reference shows that down to two or three miles there is
no effect of length on unit investment, a constancy that must
break down, of course, at very short distances.
Pump Station Investment
Earlier studies (40) developed a correlation of pump station
investment as a function of installed wire horsepower. This
relation with appropriate cost index adjustments is found as
Statement 452 in the Program.
It is recognized that major factors influencing pump station
investment are not only horsepower but also TDH (total dynamic
head) and firming factor. Since the above relation was
developed other authors have developed correlations which
take some of these into account. However, some of these are
not supported by actual data in the publications; others have
used firm capability rather than installed capability in the
correlation; and it was felt that'the subject really required
rather an intensive review using actual investment as installed.
This was judged too big a task for the present purposes, parti-
cularly since pump station investment is generally a rather
small fraction of total investment in a conveyance system and
makes a relatively small contribution to the cost.
Pump station price is. trended by the USER (United States Bureau
of Reclamation) Pumping Plant Building and Equipment Cost Index
and regionalized by the BCI (Engineering News Record Building
Cost Index).
OMR on Pipeline
The correlation used for OMR (operation, maintenance, repair
and minor replacement) on pipeline is from earlier work (40)
admittedly based on rather poor data. However, the contribution
of OMR on pipeline to total cost is quite small and a greater
degree of accuracy is probably not warranted. The relations
are given in Statements 314-335 in the Program. A terrain
factor (TERFAC) of 1.0 represents good terrain conditions for
maintenance in relatively open country and ready access. Sug-
gested terrain factors for other conditions are:
Medium marsh 2.0
Bad swamp 3-6
Mountainous (5)
An appropriate terrain factor for urbanized territory is difficult
to assign. A provisional suggestion is 1.2 for that mileage in
urbanized territory.
191
-------
The OMR costs are average over the pipe lifetime. Appropriate
factors, not used in the program, for other ages are suggested
as:
New lines 0.2
10 years old 0.7
Pump Station Operation and Maintenance Costs
The Bureau of Reclamation (48) studied 174 pumping :plants
ranging in size from 5 to 15,000 horsepower and concluded from
the data that annual operation and maintenance costs can best
be estimated by considering operation costs and maintenance
costs separately.
Multiple correlation against a number of possible parameters
indicated that the factor having the most influence on operation
costs are attendance (whether unattended, semi-attended, or
attended), station capability, design TDH, and length of the
operating season. The last of these refers to the operation
of particularly irrigation pumping plants during only the
irrigation season. For maintenance costs the significant
parameters were station capability (mgd), station horsepower
and annual water pumped.
The maintenance and operation covered are for the pumps, motors,
accessory electric equipment, miscellaneous equipment, and the
plant structure. The costs do not include the operation and
maintenance costs for the intake channel or the G&A (general
and administration) expenses. More details on definitions and
coverage will be found in the reference.
The nomographs and the equations presented in the reference
were translated into a portion of the computer program, State-
ments 453-504 in the Program. The basic computations allow for
selection of any degree of attendance with any horsepower. The
present program assumes an unattended plant if installed station
HP (horsepower); is 450 or less, a semi-attended plant if it is
5,000 or less and an attended plant if it is greater than 5,000.
If the installed station horsepower is over 15,000 a different
relation is used as explained in the reference and incorporated
in the program. The program, of course, takes the season as
52 weeks per year, i.e. continuous operation. The labor portion
of the O&M (operation and maintenance) costs is trended with
the Labor Cost Index; the non-labor costs are trended with the
Maintenance Cost Index.
192
-------
Energy Costs
A Function Subroutine CKWH generates (if a C/Kwh electric rate
is not prescribed with GVCKWH), the average electric rate for
the state, the year, and the Kwh/yr consumption. The base state
averages for industrial service 200,000 Kwh/mo and 1,000 Kv/h
demand are the January 1, 1969 state averages from Reference 49.
The adjustment for future year is by the Energy Cost Index from
COSTN described beyond. The adjustment for consumption level
measured by Kwh/yr was obtained by averaging the relation between
cumulative C/Kwh and Kwh/yr obtained from a variety of electric
utilities (50). This study was facilitated by a program. SLECTR,
which may be of interest to some readers. It gives the cumu-
lative C/Kwh effective unit price for any or any series of
Kwh/mo consumptions at any load factor, from input data
consisting of the block limits and block rates for Kwh and
for Kw demand taken from the typical electric utility rate
schedules as found for example in the National Electric Rate
Book.
Incidentally, this ELECTR study revealed that it is necessary to
use caution in interpreting rate schedules. Just because a rate
schedule contains a lowest energy charge of .3 in the highe&t
consumption block does not mean that the cumulative energy price
will become asymptotic to three mils as energy consumption
becomes very large. The rate schedules in general contain fine
print that cause the asymptote rate to be considerably higher
than the lowest rate in the schedule. It is not at all unusual
that a rate schedule containing a three mil block actually does
not allow the cumulative rate to become less than seven to
eight mils.
The Cost Index System
COSTN is a Function Subprogram providing regional and temporal
cost indexes which are used to adjust historical data to soire
common year and to adjust regional data to a national average.
This subprogram will ultimately incorporate all cost indexes
which are useful in chemical and water and waste process costing.
The regionalized indexes in COSTN which are used in the present
program are the 21 region BCI, and a pipeline adjustment
factor (47); and from the non-regionalized indexes there are used
the composite pipeline cost index, the electric energy cost index,
the pumping plant cost index, the maintenance cost index, and
the average hourly earnings in manufacturing.
193
-------
The Function COSTN returns an index projected for a future
year. All the indexes have been subjected to a time trend
analysis. In general, it was found that the cost indexes over
the past 20 years can be remarkably well represented by a
Cartesian linear relation from which there are found two types
of anomalies. One group of indexes has a hiatus in the period
about 1960 to 1965 in which the index does not increase very
much, and thereafter increases at about the same slope as
before the hiatus. In these cases the projection has been
made by dropping the hiatus years, in effect shifting the
prior years upward in time by the amount of the hiatus such
that the new set of points define a line with the same slope
as before and after the hiatus.
In the other type of anomaly the reverse has occurred. Begin-
ning somewhere in the period 1965 to 1968 many indexes have
taken a sudden upward turn. This is true of all the BCI,
some to an extreme degree. Since there are not enough his-
torical years to establish a new slope or level or both if
such is to be, it is very risky to make a projection. The
present projections use this device: the projection has the
level of the actual 1969 value at the year 1969 and has the
slope of the regression line 1948 to 1968.
The energy cost index returned by COSTN is the projection of
the national average cost for industrial electric service
200,000 Kwh/mo and 1,000 Kw demand (49).
194
-------
RESULTS OF EXEMPLARY COMPUTATIONS
Following this section are the program listing, the variable
names and instructions for running the program.
Conveyance Cost in Horizontal Lines
The characteristics of conveyance in a horizontal pipeline at
a utilization factor of 0.5 under average U.S. conditions are
shown in Table 25. The contribution of each of the five cost
elements is shown in Figure 32.
TABLE 25
CHARACTERISTICS OF OPTIMIZED CONVEYANCE
IN HORIZONTAL PIPELINES
(1968, National, UBARE = 0.5)
Average conveyance rate, mgd
.1 1 10 100 1000
Optimum pipe diameter,
inches
Pump-station spacing,
miles
Investment C/mile/gpd
capability
Line
Total
Conveyance cost,
C/Kgal/mile
5
12
8
9
.5
.5
.4
11
8
2
2
.3
.3
.7
32
25
0.90
0.96
86
33
0.32
0.36
220
50
0
0
.12
.14
3.9
1.2
0.41
0.16
0.070
195
-------
Contribution of Cost Elements to Conveyance Costs
Horizontal Lines, 1968, National
100
80
60
40
20
_OMR Pump
-OMR Line
_ Energy
_ Pump^
I Capital
/Charge
_ Line J
. }
1 10 100
Average Production, QBARE, mgd
inno
Figure 32
Parameters which enter into the conveyance cost are water
temperature, pipe roughness (in computations herein taken as
0.0003 ft), pump-station efficiency (taken as constant at
0.75), firming factor (amount of emergency standby pump-
station capability), over-all pipeline slope, and utilization
factor (ratio of average conveyance rate to design capability)
The sensitivity of conveyance cost to most of these parameters
is quite small, a 100% change in parameter value bringing
about only a few percent change in conveyance cost. The para-
meters to which conveyance cost is sensitive to a degree
greater than this are pipeline investment, slope, utilization
factor, and energy price.
As a determinant of conveyance cost, the price of electric
energy is much less important than is generally believed, for
the cost for energy is not an important factor in the total
conveyance cost, except when the pipeline has a high positive
slope, and then it becomes dominant only at high conveyance
rates.
The effect of utilization factor on optimized costs in a
horizontal line is also not very great. Over the range
reasonable in municipal and industrial practice, the optimized
cost change from UBARE = 0.5, with utilization factors between
0.4 and 0.7, would be less than + 9%. These differentials
should not be confused with the differential between the con-
veyance cost in a line optimized at one utilization factor
and then operated at a different utilization factor.
196
-------
Conveyance Cost in Inclined Lines
The deviation from horizontality in a pipeline is an important
parameter affecting conveyance costs. Actual pipelines follow
the profiles of the land, but for approximate cost computation
it can be shown that every pipeline can be expressed as a
two-section line, of which the one-section line is a special
case. The upstream section is that from the beginning to the
highest intermediate point higher than the beginning. The
downstream section is from the highest intermediate point to
the terminus. The model pipeline thus has two sections, the
first having a positive gradient, the second a negative
gradient, and either one of the two sections may be missing.
Conveyance costs must be computed separately for each section,
although for lines several hundred miles in length, the cost
will rarely differ by more than 25% from that for a hori-
zontal line. The program also allows segmenting of the line
in other ways, so long as in any segment no intermediate high
point is higher than both ends.
For a line having a positive gradie'nt, or a negative gradient
of small magnitude (such that it falls within the pumped or
pumped gravity-assisted regions of Figure 34) , the additional
conveyance cost over that for a horizontal line is closely
proportional to the slope (at constant energy price). The
proportionality constant is the cost of raising a million
gallons one foot, and this is termed the cost of static lift.
Figure 33 shows the cost of static lift at various average
conveyance rates and utilization factors, and at constant
energy price of l.BC/Kwh. Above one mgd the cost of static
lift is practically constant, and the cost in cents of
raising 1,000 gallons 1,000 feet is about six times the
£/Kwh energy price.
Note, however, that the present program lowers the energy price
as annual energy consumption increases. Under these circum-
stances the cost of static lift reflects the changing price of
energy, such that the cost of static lift decreases both as
the slope increases and as the flow increases, since both of
these result in higher energy consumptions and thus lower
energy prices. As an example, Table 26 shows the cost of
raising 1,000 gallons 1,000 feet at various QBARE. Quantita-
tively, the effect of slope at constant flow on the cost of
static lift is small and somewhat erratic because of the
optimization. The effect of flow on the cost of static lift
in the ranges used is major and while the differences are
reduced by optimization, they are far too large to be reversed.
197
-------
10 r
COST OF STATIC LIFT REGARDLESS OF FLOW DISTRIBUTION
Pumped and Pumped Gravity Assisted Lines,
Constant Energy Cost of 1.5
-------
TABLE 26
EFFECT OF CONVEYANCE RATE ON COST OF STATIC LIFT
(With energy price varying)
(National, 1968, UBARE = 0.5)
QBARE Cost
mgd C/mg ft
.1 43.6
1 15.2
10 10.9
100 8.0
1,000 6.5
This static lift proportionality is not maintained in regions
of negative slope labeled gravity and gravity boosted on
Figure 34. This figure shows the effect of slope on conveyance
cost at a series of average conveyance rates, computed at a
utilization factor of 0.5. Figure 35 shows the contribution
of the five cost elements at a high-positive and a high-negative
value of slope. It is important to note that at high positive
slopes energy cost displaces the fixed charges on the pipeline
as average conveyance rate is increased. This means that at
high slopes and at high capabilities it is the price of energy
rather than the pipeline investment which is a major contri-
butor to conveyance cost.
Effect of_ Pipeline Length
The cost of conveyance is directly proportional to the conveyance
distance except at quite small distances, at which it is some-
what greater than the per-mile costs illustrated herein mostly
because of the high unit price of the small pump stations. The
distance at which this effect begins is about three miles at
two mgd capability and one mile for 20 mgd and higher. The
program itself correctly computes the costs even at these short
distances, but it does so on the basis that the unit investment
for line is unchanged at short distances. This has only been
demonstrated down to two to three miles. At some unknown
distance, less than this, this must no longer be true.
Conveyance Costs in the Future
The COSTN subroutine allows the projection of the cost of con-
veyance at future dates. Table 27 shows the costs predicted by
this program for the conveyance of 100 mgd at UBARE = 0.5 in
the next three decades. For a plant built in 2,000 both the
unit investment and the conveyance cost will be almost double
that for a 1970 plant in current year dollars.
199
-------
.1
Average Conveyance,
mgd.
Conveyance
Cost,
£/Kgal mi
(1968, National)
(Field boundaries)
are approximate)
10
100
1000
]OST OF CONVEYING
WATER BY PIPELINE
UBARE - 0. 5
Effect of QBARE and SLOPE
-50 -40 -30 -20
-10 0 10 20
Line Slope, ft/mi
40 50
Figure 34
200
-------
CO
o
o
CO
o
o
CONTRIBUTIONS OF COST ELEMENTS TO CONVEYANCE COST
Inclined Lines
100
80
60
40
20
0
0.
OMR line
OMR_pump.
energy
CR line
//, = +50 ft/mile
kO<
10
(a)
100 1000
OMR line
CR pump+OMR pump
+ energy =0
=-50 ft/mile
201-
0
1000
Average production, 0, mgd
(b)
201
Figure 35
-------
TABLE 27
COSTS OF CONVEYANCE IN FUTURE YEARS
Horizontal Line, 100 mgd, UBARE = 0.5, National
1970 1980 1990 2000
Investment, C/gpd mile .328 .427 .526 .625
Conveyance, C/Kgal mile .147 .188 .229 .270
Comparison With an Actual Engineering Estimate
When the PIPELIN program was used for the San Antonio compu-
tations a comparison was made between the PIPELIN costs and
the costs developed by a preliminary engineering study of a
conveyance system, the Cuero-Cibolo-Hildebrand link of the
Texas Water Plan, which had been made by a consulting engine-
ering firm (51). The comparison is discussed in more detail
in Chapter 6 of the report but it is briefly mentioned here
since it bears on the accuracy which may be expected out of
PIPELIN.
The engineering study laid out an actual route and designed
a pipeline and pump station system under a set of ground rules
laid down by their client. PIPELIN was run with the same basic
data. The corresponding 1969 costs of the two studies were as
follows:
QBARE Conveyance Costs
C/Kgal
Engineering Study PIPELIN
100,000 afy 7.15 6.72
200,000 afy 5.72 5.70
300,000 afy 5.05 5.45
This information is presented here simply to illustrate the
confidence that can be placed in PIPELIN costs as reproducing
costs from rather detailed preliminary engineering studies
which include field studies, map routing, topographic and
geologic profiles, and item-by-item preliminary cost esti-
mating. The firm demonstration of confidence, of course, would
require many such comparisons in various regions of the country,
202
-------
INSTRUCTIONS FOR RUNNING PROGRAM PIPELIN
This program is set up to run in Fortran IV on a CDC 6400
computer. There are about 900 cards. The compile time is
about 11 CPU seconds and six peripheral seconds, the execute
time about three and three resp. The program will compile
with a core memory of 60,000 60-bit words. It will execute
from a compiled program or binary deck with 22,000 core words.
The main set of data cards are those numbered 1 to 11 on
Page 203. The data deck structure is shown on Page 204. The
program can be manipulated to iterate successive cases with
new values of one of the data cards by inserting additional
data cards 12, 13, etc. and using the proper values for LOOP.
The values read from the data cards are shown on the next
page. Card 7 comprises actually a set of cards, one for each
segment. (If the number of segments is greater than one then
an instruction for a short printout will abort the run.) If
it is desired to run just a single case without looping the
LOOP card should be punched as one. The program also will not
loop, i.e. will compute only one case, if LOOP is set at 2, 3,
4, , 6, or 10.
Witu LOOP at 5, 7, 8, 9, or 11 the program will return after
the first case and read the next card in the deck, 12, 13, etc.
With each auxiliary card or card set so read it will compute
one case and continue to return until the deck is exhausted,
increasing the case number by one on each iteration. It is
not possible in a single run to loop on more than one type of
data, i.e. on more than one data card number. For example,
a series of cases can be run in which the QMAX is varied
(Card 8) and in a separate run a series of cases can be run
in which DEGC is varied (Card 11). But it is not possible to
iterate varying both QMAX and DEGC at the same time.
When looping is used the data on the new card is printed out
just prior to the new case number. (This print is not suppressed
with IPRINT(6).)
203
-------
Data Card Format
No. No.
1 Run no. 101
2 LOOP 101
1, 2, 3, 4, 6, 10.. .will not loop
5, 7, 8, 9, 11...loops on data in corresponding
data card number by reading cards 12, 13, etc.
3 This data card number not used
4 IPRINT(6) 85
zero suppresses; 1 prints:
(1) parameters imposed
(2) parameters of the optimized design
(3) cost breakdown
(4) short printout
(5) search iterations out of OPTIM
(6) data cards 1 to 11
5 STATE, NMSTAT, NREG, IYEAR 95
6 NUMSEG 101
7 One card for each segment
7a
7b PLMILE0), ELEVB(I), ELEVE(l), TERFACfl), CNSFAC(I) 111
etc.
8 QMAX, QBARE, QBARA, QMIN 111
9 PLLF, PULF, BASCL, RET 111
10 TX, XINS, PRLAB, PYEX 111
11 GVCKWH, DEGC, HDLIM, EFF, EPS 111
Data for subsequent cases in loop chosen
7/8/9
6/7/8/9
204
-------
IBM
FORTRAN Coding Form
X28-7327-6 U/M 025
Printed in U.S.A.
*A standard cord form. IBM elecfro 8B3157, ii ai/o;lol>le for p^nthimj jlo^m
-------
SAMPLE PRINTOUT
-- — RUN NO, . 6
111011
_.3.!i*zaa2
3P.7OCG
.23P.pn.Q3
7C . n!?PG
?« ,5.aDJI_..J>JIO_r.lfl.DJl.
1*»9»OQOP 1 *j9§ D03C
?5. 0°0 P 100. Or!00
.niOi ...2».QQQC
?i,50oc ^no.onoc
.0... l.CO.QMO
1.0500 1.05GOTO
.7=00 .OC0700
-.: CASr MQ.t. . 1. --r- .
Jtl.AR_QF F5TI_MATF _ . lcf-9_
STATF TEXAS
REGION NUMBER P
MILEAGE 7fc.?000
_IF_RRfi IN1 .FACTOR 1, OJ3 00
CONSTPUrTION FACTQ? l.COQO
SLOPE, FT/MILE d^fj
SFGMFMT ?
1.05HO
DFSIGM CAPABILITY . .
EXPECTED AVtRAGF PRODUCTION'
ACTUAL AVERAGE PRODUCTION
MAXTMUH DAILY PRODUCTION
MTNIMU.^ .nAILY .P.P.OnyCTIPM
PIPELIME LIFE
... PUMPSTATJON LIFF _
INTEREST RATE
TNSUPAHCr PATF
TAX PATF
LIMITING HEAD PUHa STATIONS
TEMPEPATUPF OF WATER
LABOR PPICE .
GIVEN ENFPGY PRICE
...P.AYPOLI FVTRAS FACTOR
l«»9j,.Q.n.O.O
jJLtOO.
• Pino
7.00
MLLIHN GALLCNS/OAY
»-'ILLTON GAtLCNS/OAY
Kit LION GALL
BILLION GAL LI
BILLION GALlCNS/HAY
PFACTION/YEAF
. Cc FLUTO
GDC C C P
1 •' • '-. ^ l» ,
FPflTTTON
206
-------
PARAMETER? OF THE OPTTMI7EO OFSTr.N
~ SEGMENT 1 SEGMENT ?
CONVEYANCE TYPE PUMPEQ. PUMPED
_CLPTTMUM...PJPF._DiaMIJLR_I_»Jli. T_NC_HF_S.. _ .. F8,<4<+98 fi3.t*k73
J1AX.IHU? P R.ES5U.8.E.JdLASf ?S1 15.0, Q 0 Q 0 150.0000
JJl)M?FP OF PUMP STATIONS STATIONS 2 ? . .
JIACU P_UMP...S.TAT.TO_N £J.i__C!£_-fJL..uT.n_ 221.SPSU 286.8592
JJHEPSIAT.t(lN..iISIANEE _.._ MILES 17.1 COO 12.7667
DESIGN IN^TALLFQ HP HP/STATION l/t ° G..21 f» 5.9 8 19426.1518
!£S.I5JLJ?£Y!10L.QS_NyjlS£P_ „ 5.6E95E + 06 6.21it^E + 06
.FLOW. TYP.E .^DESIGN. .._ TPANSITION TRANSITION
FLOW TYPE. AVEPAG^ .TRANSIT ION TFANS IT ITN
.5.PJJ.M.UM VF..LQ.CITY_j...n.£5_I£M_ FI.«/S£C.j . 8,3^98 9.3706
VELCC I.T.Y_a A.VER.A G.E FT *_/£££_, 5.^028 6.0705
DESIGN FRICTION HEAP FT./HTLE _.. 9t.318.5. . 11,2^8
AVERAGE FRICTION HEAP ^T./MTLE _ ... ^.5855 4 . 8?6?
ENEFGY_ PPI.CE_ .'. ._0.£NT_S/KJWH^ . ,9615 .9560
207
-------
. pFP.CrNT.
^ TOTAL
TCTAL
PFP MILF
.SL»—.H
ion.
CflPPPILITY
7 Q , /4 i, Q
CENTS/GFD
FRODUCTICN
37.936
7.522
.627
OHP ON PTPF.LTME
_QHF_Jl N_ P ILM.PJSIA T. IO_N S_
iTCTAL OPF5ATIMG
CAPITAL CHA^GF ON PTPFLTK'F
CAPITAL CHG OM
TOTAL PPCHUCTION COSTS
PFP MILF
ANNUAL
: DOLLARS
I^P.Uf;
3^8. it a.
UMIT
CrNTF/KGAL PFFCENT OF TCTAL
1970.?(,
.12
n c;
3.537
«*.fi?3
f. <='?G
1.201
?.59^
?7.22?
^tt.Big
51.B90
13.1455
208
-------
PROGRAM PIPELINdNPUT, OUTPUT,TAPE5= INPUT)
100
110 C
120 C
130 EXTERNAL FUNC2
1*»0 INTEGER STATE,FMT
150 COMMON/eOOOl/VIS,EP?
160 DIMENSION PLMILE <*»> ,ELE VBU) ,ELE VE (*•) .TERFAC (U , ARRAY (7) ,
170 +GPOC(22),GPDP(16),CNSFAC<«O
180 DIMENSION SEGMNT <<», 25) , IPRINT(6) ,PC (22) , NMSTAT ( 2)
190 DIMENSION C NTY PE <<») ,FL TYPE CO
?00 DIMENSION HHCK«»)
21C DATA FLTYPE/10H LAMINAR,10H CRITICAL,10HTRANSITION,10HTURBULENT
220 + /
230 DATA CNTYPF/10H GRAVITY,10H BOOSTED,10HASSIS PUMP,10H PUMPE
2»»0 +D/
250 ARRAY(3)=1.
?60 30 READ (5,101) NRUN
270 PRINT *»0,NRUN
280 kQ FORMAT(1H1,///,10X,* RUN NO. *,I2,» *)
290 READ (5,101) LOOP
300 L = LOOP-i»
310 NSET=1
320 READ (5,85)IPRINT
330 85 FORMAT(5X,611)
3<*0 READ (5,95) STATE,NhSTAT,NREG, IYEAR
350 95 FORMAT (5X,12,2A7,12,IM
360 CYMNI = COSTN(20,0,IYEAR)
370 READ (5,101) NUMSEG
380 101 FORMAT(5X,12)
390 IF(IPRIMTCf)*NUMSEG-l)102,102,2110
-------
650 117 CONTlMUf
660 VTHF.S = 0
670 C.....LOOF RttK'T^Y
680 150 KTIrtIS =
690 IF (IPRI>
700 157 PRINT ITS, NSrT
710 1?8 FO??MAT
720 11«» ;?TLAB=Pr-LA^* (1 .+°v?X) *COSTM (11,P ,IY?"AP)/2
730 IF fNTI^T-l) 12.3,12.1,75
75 (.0 TO rf
108C •"!=
1090 TF
1100 Ifl8 DOLO=U
1110 &0 TO 1*3
1120 100 T=0»l?.
1130 "SFGINTtN'SFr-,!) =HHTY?r (1)
1150 S
1160 ^-G^NKt'SFG,?) -FLTY^n (TTYP; )
1170 AVFAC=1.
1180 GO TO 3f:f:
1190 210 ^EGMNT(^SF^,,1V) -VL!^
210
-------
1200
1210
1220
1230
1240
1250
1?60
1270
1280
1290
1300
1310
1320
1330
13AO
1350
1360
1370
1780
1390
1400
1410
1420
1430
1440
1450
1460
1470
1 /, ft n
m o u
1490
1500
1510
1520
1530
1540
1550
1560
1570
1580
1590
1600
1610
1620
1630
L \J v* U
1640
1650
16*0
1670
1680
1690
1700
1710
1720
1730
2f)7
SFGMMT ( WSE- G , 17 ) =WCOKPL
S F G MNT { NSF G , 1 S ) - WL Of 'P
NT OT = WCf HF L +KL (V1R
SEG '•INT ( K'St G, ?? > =WTOT
IF ( SLOP f +F? . >70a,7r>P,?n9
KK = i*
C APPROXIMATES FIRST PSTM'TPP fiKin iTS-'TT'v. PIJMPFn ITMr
2 09
210
220
22?
224
226
230
2^2
223
235
2~1HI=2 ,*niAM
IF (KK- 1 ) E:Ti* ,23^,234
niAHMI=n
IF (D-ni &MLO) 22?, 2 3^,? 35
01 AivtLO=f!/^.
01 A^ = DIAMl 0
t^K^^iLLi^e" jjj-n
NCALL=G
KKK-a
— — P P M T P v f ^ P o P T T fi i one
^ - ; ' \ ! 1 1 \ / - ' I 1 J . L • / - J
STATION BOOK n?Pr=ci
R OF PI)? P STATION'S
CALL Hrorv(QnoT,DIf'i:'l,c?rY, FifOCO, JT Y"r )
crpM!~>QT— "^1? 1+ T -4 ^^pKOC^^^OOT ^*^/01 A "H ^ * '"^
TC"fp^>LJQf"1-|.C|pP^J"^f-tt^ 7f, t^ ">7^
NUMSTA=1
HOSTA = i? .
GO TO 377
KK = 2
*»U''STA= «FHMD'">T + SL;JPr) *PMILr/HOI I?- -f . -T993 'J9)
HD5TA-( FPHDOT^SLOP: } *n>''ILF /V'U'-'ST A
'"' U M I L L-- F M I L F /N U M jT A
REY11QT=F FY
TTDOT=ITYFF
........' HJ'JSTT-JG rrr" IMVrc.TMr"NT FC° Qf<" S^L'f r HIA^'^
IF ( TP^'Y.I. T .?) l°*f .S = ?
PRCL^X= TPF rs*Su .
r--!pCLS=IhFTT>-l
AVFAC = ri .
00 ^uri ^i = 1 ,NFCL S
PRCL=5Q .+?T.*J
AVFAC= AVFAC+r>-3FAJ(OTAM,pf?CL,rASf.-LI
AVFAC= AVFAC/MPCLS
T-fjI AM*1?
211
-------
175H
1770
178C
1790
1600
1*10
1820
1830
1650
1P60
1S70
1890
1900
1<510
1°20
1930
1950
1Q60
1970
1980
1990
?noo
2010
2020
2030
2050
2060
2070
2080
2000
2100
2110
2120
2170
2150
2160
?170
2180
219C
22CO
2210
2220
2230
2250
2260
2270
228C
2290
C LIN't INViSTi^NT EK lit? f 51
7GG rONTINDF
CPMI = COI>;SF*AV-AC
GO TO (701,301, 702, ?01,301, 3-01, oGl,3.!37., 301, 332, 303, 301, 3J1 ,~02,
1^0 2, 3 01, 7 C?, 30 1,30 1,3-; 2, 301) ,NRf G
301 CPHT =
r^O TO
3'~2 r:DMl =
GO TO
31 1 IF(T-51.) ^
315 T = ALOG(T)
r-n TO 37?
330
•n-
335 ^
VLT'4E=Cf
WCCH°L=VL
GO TO (fO
fOM-ii
HPSTAF=C .
/COSTN(1,?1,TYFA!?)
'^^) ,KK
C U^!IT TNV-ISThFMT ANQ 01'^ f. X PNFHGY
C....FUMP ST4TTCN PRICE TO "F 9~VJSF.D WTTH Q-TOH FARAMtTF1? AND
C... .U^'TTS SUPPC'JTT^r RE°1ACING FIRM .....
TO
{.1016&f.7»
C V A^JR W PUMP
C I'??" MALFfTr'CM, CONTIMUOUS OPFT710N PU^P
TF ( HP iT A I- 1 5 0 0 3 . ) i» '-J Y- , 5 " 2 , 5 0 ?
C ...... U N A T T!T i pr 0 FL A N TS
or.a.)i*7n,^70,tf75
C ...... Sc 1I-ATTFMOFD
GC TO
PLANTS
C. ..... PU -!P •t
212
-------
2 ? 0 Q H •*"' 0 *:; — i ^f rJ *° T (_ 4 ° + CY;
2310 !F(-f»STf T-1.5G.
2320 Mb -UO'1p=Pl'OMP-Kii ,MQ-)OT*1 •
2330 GO TO 50i*
2350 «•**.:+•'*?
2760 IF (HPST/1. r-7oro.) 4c55,5r.!!),i?n L~
2380 TO TO 5ftf
239Q 5;''J °IJO J!R=PI.'OK">-»-l . 7*US lp
2400 HO TO vO
C MftMUAL OR RM IONTOV~" 1.501
c:t r 0NT I HIM
21*50 r/PQ"P=Pl!C
2460 '/PlMp-^ F
2«»70 'vCC'-lPU^ VPUMP*;-)T TPU
2^80 C FNL'^GY
2^90 PALL ?tO COY < OTft- , TT '
2 5'10 FC>H-"!/i-= ~ 1
2^10 IF (F:'?HLf
2520 5J5 KK=f.
253Q IFC^LOP')r15,333,5?5
2550 GO TO 5?r
2560 5?2 OPTUT h2
257H F?3 f"o>;nAT(»
2580 +Fin.?,*£FT AS
2590 5^F WL'MG *•=«•.
2F-00 r-C TO ^rO
2610 b'-^ YPKV)H = 11£»7.«.^?:
2620 IF (GVu'KkH) ?Li.,
26? c r-H- HwcK(rjrrr)=c-v
264G r-0 TO 5i,£
2650 jit:. HW.< (-l^^G) -C. KHK( Yr -60,3o°
2720 5- C TF ( M T GT - 1. t * S-.G.-1MT (^'^L G , 2? ) ) 5€ ' , 5^^ ,'~,^k
2730 c.^? JJJ = 1
2740 r'0 rn -,fo
2750 f>-t (LL=LLL+1
2760 TF(LLL-7)E <^ ,-n 1,7Hr
277C r '••^i;jT(rO n^TT" -^A-rH"?
27PO C...,f^T- THA-< TUTS IS V^OT I hi AH~ OT/1^ OUT rlXCPPT OM 1Mb. OPTIilU"..
2790 ?>69 ] F ( ir-.^T'-'T (:>) ) r,;75, 3~5 , 57T
2800 c.7j -RI-.JT 5V1, NCALL|DI/i'1»»<-K,WT?T
2P10 £/l rO^VAK 1 X,J. ",Kin.3,I3,F1.6. 3,/)
2R20 575 Tc tNOWGC) 3? J ,5? ; , r H':
2630 5"0 TF(1^^1KT(5)) v 33 , l..?0 , 5«5
28^*0 [.^t> ~-:lNT 5f'f~
213
-------
2850 5*6 FO^IA f( IX,"OPTIMUM*)
2860 C GRAVITY - SU*
2870 590 IFCWTQ^-S-GfHT (NS:-.C-,?2>) £00,7'0,7uO
2880 C FILL SEGHNT KITH PU'^PfJ Ofl TA r.'IS'-LfiOE GRAVITY OATA
2890 600 SrGMNT =f
2910 ^EG'WmST-G,?) =N
2920 StGMNTOSEG, U = PUMILE"
2930 SEGN)NT( i\StG, F)-H^STAT
ScGHNHNSf G,6) =REYHOT
SEG^NTUvff G,7) =FL7YPF(ITOnT)
2960 f- l,g&9c7tt/niAH**2
2970 bEGMWT(Nr.FC,8)=3
2980 cFG'1NT(*'Sr&,°)
2990 FfGlN'TtKFtG.ia
3000 SrG(-1wT{rSrG , 11) =F
3010 SFG'WT(M?FG,1?)=Q
3020 StG.v)NT(f^5F C-,1'
3030 5EG^NT(rSFG,14)=VLTNF:
30tfO r>:G^t-JTU-.?Fr,, 15) =\
3050 VTOT=^/L I M +VFU
3060 ^rG'-tNTCMSEG, 16) =\
3070 c.FGMNT(VSf G,17)=*
3080 C.EG:-!NT(KSFG, 1-.) =WLO.'-iR
3090 iEG-1NT(f SF G,l'^
7100 ^F G^INT { f'5'E P, 2 '
3110 SFGMNT(f5FG,20)=V
3120 ~EG«1NT(HSFG,21)=WFMr,Y
3130 SEG'IMTd rEG,2?)-WTOT
31<*0 SfGMNT(^FG,2=)=PRCLHX
3150 TF(HOSTM£10,&ia,7f,r
3160 610 r'RINT 62f
3170 620 FO°.MAT(* HO^TA APPROACHING O.*,/,* HO'-^Sf ^O^C- A''?ei TH Ai?ILY SiT
3180 +1G '. FOP DIJwP IMV^STMTMT*)
3190 GO TO 7ffi
3200 CFILL ~"TG.-l~:.-jT HTH GRAVITY OATA
3210 7JP l^IAM-5rr>'M(KSrG,2)/.t2 .
3220 T = l.<3t?rV'i/niA-"i**2
3230 ?HG^NT(fSFG,fl)=QOGT*T
32r>0 ^ALL in (.TV (OP AP, 01 1t-:,^r Y,FMnor., TT Y'; '.)
3260 ^HG l^Tt^S' G, 1-D = ri(«.^^i4t-*FM''>Qr*ORt''**LVr'lAM«»5
3270 ^r. --MT (r Sr r-, 1 •') -Rr Y
3?80 •~!iG'!NT(f-5.^G,ll)=FLTYP- (TTYP >
3290 >,FG-'HT (i- Sf G, 1&) = 0.
330 n SF GMNT(h'Ff-S,lO =?t
3310 c-- Ti'-'NT('lTG,l?) =0,
3330 ^LG'^.'TC^^t-
37«»0 'r:G"!Mf(?Sr
3350 ~tr,.XNT(f.S[i",9) =-SLiP'
3360 r:rr,;'!NT(rs.FrM?3) =o.
3370 76P COMT1KMT
3380 IF (I^RINTCt)) /S?,7:
3?90 C TOTAL ^:r>J!FMTS. ?"
214
-------
3MO
765 '.ITOT - furs-G<
00 !»2-:i T = li»,?'
3t»30
3U£0
3^60
3<+70
3**8G
8?C ^r
DL M
P ^ C": p 1
m
ft/, t., <-r (
r,
i'
TL
j -?
S
t
r
>
3i»OC 'I'INVP
35CO
3510
352H
3530
3F^O
3550
3560
3570
3580
35PQ
3600
361 0
3620
3630
36^0
3650
3660
3670
3680
369C
3700
3710
37?0
3730
37^0
3750
3760
3770
3780
3790
3800
3810
3820
38*0
3fl*»0
3850
3860
3870
3880
3890
3900
3910
3920
3930
?9<»0
TOP
•;
r;
NT
r(
i ;
ii
= (
*/ —
T =
(
Nil
o
i ITT
(
I
!-
b;
^
- 1
FG
FG
££
0
ii
^
J
T
)
T
,
f-l
M
,1
-1
} _
Ib
T(
T(
T(
) =ScG
-L-TL
NTOT ,
f,'T^7.
"TOT,
• \li\jjL/
-(^TOT)^LVTI t(T)
T) /f.TG'-'MT (rnri, lo) ) *J JP .
Ic) /rL:<'IL5(NTrT)
1 °) 4-^rGyNT(N''OT,2';) 4-S-£Gk-KT(MOT,21)
- n. -jpr, T= i nr-c. s T/ s-::G?-"4T < NT ^T, ?'')*!!(•.
".'0
tt^f PC
'• R c)
T.
s
(
\
••*
'_
I
i
A
rALL
885 GP
' AL
'-]
L
f1
f,
) =
( t
c
00
(I
CO
]
<~
)
I
C
)
<
~1
Ff-
= C
- j^
TF
= A
TF
7
y
J*
(
(
,
N
n
,
?
P
T
21
T(
T*
In
FG
/sv
I';
NT 01 ,
1 • h
M T ( M
( ;. )
V ~ M , A
T) /S f'G^N'T (MCT ,?";) *!,, Q .
""IT , T) , A^S>.* Y)
r-"?A Y)
T, ^> )C =5 rr/i y (f )
.-" P ?
!>0
935 GO
1
5
L
n
Y
"
n
CALL
(1
if,
CO
(I
CO
)
7
c
)
c
r-D
= j
~f\-
= A
TF
-•>,
vi
(
3
a
,
^
(T
u*
In
FG
AY
IN
l~b
••'NT ("!
('•':> )
V JM, A
Tf)T , I) , ftRRf Y)
r>?AY )
T C, ° 0 P - A r F^ A Y ( f )
A R P
on
a
Y
1+
(t
5
)
1
~ C
3
= 17
A
»
F*
2?
Ir3
TALL COSTP(SF.:-MMT (f4
9'4'J GP
1 KG
on
P
•/:
(I
=G
)
F
=-f
DC
P
(
c>
AY
(r>)
?2> /r:L^T
TTT,T) , A-'9AY)
L:r(NTOT)
CALL COST ( TC;°r,ST» A&'PA Y)
7HP
152'? P0
1530 FO
i / 4
1 S^b F 0
1.5-0 FC1^
15i»S' FO
1553 FO'1
1553 FO
1560 FO
-K.X,
1565 FO
1570 FO
1575 FO
r;
-*
^i
'%
<
H
M
G=
AT
AT
y
<
f;f
1H
a
1
(3?X
6X»K
r^
rl
V
ft
AT
T(
OMAT
(
(
•t & T ( 3
„>
M
AT
(
DCL
IX
IX
x,
//
*
»
Y
,
»
L
V
(5
//
*c
A.',
FI
)
t ?5X,
r.RCfM
S*,FX
r ~ t T i\j
PS TAT
*T(jTiL*,
*pr:,-;
,
1
f X
MIL^
,*FPO
"MAT (7?X,*ANNU/'.L
'CtiNTS/K
?
^s
AT
-MAT
?
isao FOP
158~ TO
t-BX ,
159H FO
+ SX,
159-j FO
<-^X ,
16CFs FOP
C», «•••••
P
F
i»
M
'•!
AT
AT
AT
. 3
RMAT
F
3
.3
?MAT
*
•1
P
1
A
R
30
T(
IN
(
(
fl
IX
IX
(IX
(
5
IX
(IX
,
(
.
/
T
/)
IX
*)
,1
. .
X
»
,
»
»
,
»
X
.
G
,
AL
•»0
*on
»
*
FN
*,3X,
-!? ON
R OM
!-: :< G Y *
* COSTS ::>.'rAKL'OWN*»/,l3X,*lNVi.'ST^cNT*»/)
T*, 7X,*C"NT':/GFT*,5X,*CEh T3/&OO* ,
,*OF TOTAL*,€X,»CAPABILnY*,!.>X,*F^.OCUCTTrN*,/)
~* tX Fl1^ ^ ^ X F^ ^ 7X Fh ^ /V F^ "%}
1 0 N S * , F 1 ? . 2 , e X , F 8 . 5 , f X , P J . 1 , 7 X , F ? . ,5 )
7X,F1.?.?,7X,»llJrj.*,l'JX,Ff.^,7X,F2.'.,/)
*,6X,F10.2,21X,F»i.'<,7X,FJ-.3)
nUCTIOM*,/)
*,1«X,*UNIT rcSTS»,/,31X,*K OOLLA^S*,
*Pt.Rce^T OF TCTIL*,/)
PTPrLIN.'L*,liiX>Fin,2,'jX,F3. 3, SX, F3.7>
FMJH°«s T A T T ONS* , 1 Li X , F 1 •; . 2 , f X , F °. . ? , 8X , F 8 . * )
f27X,Fin.?,r3X,FA,j,,-5X,FP.3)
TOTAL OPF.^ATING COSTS' , 8X, Fl 3 . ? , 6X , F« . 3 ,8X ,F8 . 7 , / )
^CAPITAL
*
CA
*TO
,
.
*P>
. .
PITAL
TAL °
?:^ nr
OHtPGF. OM ^TPFLTNr'j-^X^lO.e^^XTFa.7,
CHG Otl 3UvlccTATinHS *,F11.2,6X,F;?.~,
PODUCTT O'l COFTS*,7X,F10.?,SX,F«.3,
LC*, 77X,F8.3)
215
-------
3950
3960
3970
3980
3990
4000
4020
4030
4040
4050
4060
4070
4080
4090
4iin
4120
4130
4140
4150
4160
4170
4180
4190
4200
4210
4220
4230
4240
4250
4260
4270
4280
4290
4300
4310
4320
4330
4340
4350
4360
4370
4380
4390
44CO
4410
4420
4430
4440
4450
4460
4470
4480
449C
^ --
160?
•— ~ -\ LIN 1
PRIMT
PRINT 1
r « r M i
1 A O r>
FC&,:
i
1
r
PPI^T IfilC,
PRINT 1
PRINT 1
PRINT 1
POINT 1
P2D
82F,
fi^O ,
f??,
-t
(
(
(
r i . - .} .
YrAP
NHSTAT
(PLMIL :
(TEPFA'
(CNSFf •
j. '
r
•^
•^
' • • '_- -.-•'-)
740,1^0?
(NShG) ) ,NSL G- 1 , NU'iSEG)
(NSE^) ) ,N5:rG=l jNUHSEG)
(N^rG) ) , \Sh G- 1, NUMSFG)
PRINT 1^:35, ( (SiGMNT ) , N*:K G=l , MUMScG)
(N?tG,23)) ,N'FrG = l,NU*/jfG)
(NSF.G,4) ) ,N!?rG-l,NU'1Sf.G)
(MS1' G,5 ) ) »NS*-"G- I ,slU:^c.t:~G)
{NS"G,6) ) ,NC> G = l , ••JU'ISEG)
(Nf -"."G, 7) ) ,K'Sr G=i , KiUMSEG)
(NSr.G,ll) ) ,NS£G = 1, JUMS^G)
T(MSFGi 3.) ) jN-tG-l^-iJ^SPd
(MSrG,l2)),N5LG = l,NU'k'SfG)
T (NSr.'G, 9) ) ,''»c-G-l,fJ!)>'SEG)
(t"SEG,i:-:) ) , NSrT^ljNUl-rS^G)
NFEG) ) , N5r'G=-l , MU1 ~; G )
0
X,*c^(?AMiTr= VALUES I.-1FC5 Eil* ,//)
AP OF 'ESTIM£TF*,-;X,lH>
,
2X,*RLG
.PX,*5E
G
*MLiAGE
TF?.,M FACTOR*, Fl ;1 . 4, 2 ( 5X ,F in . £*) )
j X,*SL3Pt£,
(///
,7X,*OE
FT/MILr*,3X,Pl.r.4,2(5X,FK .^t) ,///)
s
IGN PAPftalLIT Y* ,1?X,F1D.'*,!5X,
216
-------
«•* MILLION '^ALL
1HV; F:)a.JAT (3y,*£'X.p;.CTIir AVc'?4nr' r~nrUClION*,?X,FU',4,r>X,
4520 +* MTLLIO f> ALL 0"' S/T^ Y* >
4530 1350 FO *:-!AT (7X,*ACTUAL AV-^flGr ^COllf T ION* ,4 X , Fl 0 . 4 , S X,
4550 1^5-i FORMAT (?X,» MAXIMUM DAILY PROOIT T I Chi* , ? X , Fl ;j . 4 , >X ,
4560 + * :tTLLICN G ALL ON3/!?»>Y* )
457C 1360 FORMAT (SX.^MHT^IH n^ILY P-'QOUi- TlON* , >j* , PI f:. 4,r X ,
4580 -«•* MIL LICK GALLON'S/GAY*)
4590 latv FO°:-1AT (?X, * °TPELTMI I IF*";* , 1,4X , P? . 0 . 1GX , * Y~ A ?S *)
4600 1P67 r'J*:'1&T (.7X,*Pl! 5°STfiTTOf-J LIFT* , lrSX,F *. ], 1-!X,*Y^- Au'c'* )
4610 L37- FORMAT (7X , * IlJTrVt "IT t'A TE* , 16X ,F 10 . U , 5X , *FP /'CTT "M/Yr:A -:* )
4620 1880 FORMAT ("X, * IMSU^AMfF PATr* , 1->X, Fl j . 4 , "5X, *F f ACT ION/ Y- AS * )
4630 1*85 FORMAT (3X, *1'iX >A T * , ?nx , FIG . 4, 5X , *F^AC fl Ct /YE AR*)
4640 Iflgr, FORMAT (3X,*LTKITING H^flQ PUH'~ S 1 ATIONS* , ?X , Fl j . 4 , S X ,
4650 +*FT. OF FLUITi*)
4670 IOQQ FORMAT C-'X, *L&* ^''ET1 -?S rp" IH- OTTIMZEO !"»!• f. IGiN* , / )
4720 1Q3C FORMATd H'J ,?X,*GONVr YANCF T YP^:*, 7SX , Al 0, ?. ( 1 tX , A 1 0 ) )
47^0 ±q.lr> cO;V-1f.T(n r ,?y,-OPTIMUM PIP: OIAf-'f-T.R I .(;.* ,f-X, * INCHCS*, U.X, Fl? . 4,
4740 -»^(UX,F.1C.t) )
4750 1937 FORMATMHf:,2X, *r*AXir-UM P^t ? "'jr. ' CL A SS* , 1 ilX , *»3SI * , 19X , F 1 Q . L, , J- (i j> ,
4760 +F10.4))
4770 19^0 FO^-IATf 1HC ,?X,*M'JM^FR OF PIJf?P ?T AT lOM'i* , OX,*STA T TONS* , 16X , F ~ . C ,
4780 +?(17X,Fr.r»
4790 19£«4 FORMAT(iHC,?X,*Ht:AO, ^UH° STATIC t<* , 1'tX ,*F7 . OF FLUIiD* , 1 JX , F 1 '.. u ,
4800 +?(10X,F1P.4»
4810 1945 Fc?:-ta.T^AT(1HO,,?X,*Gt5IGM TNSTALLcH HP*, 13X, *HP/STA TION*, 12X, Fl? . 4,
4850 19^5 FO">lATf?Hn ,2X,*fJFSIGM REYNOLDS MUHJi^* ,^2X , r 1 a . J», :iATClHO,?y,*n.iSIGN FRICTION HF AU* , 12X, *F1 ./i-IIL 7* ,14X ,F1 C . 4 , 2 (1'•
4930 +Y,Fir:.i»M
4940 19U5 FPR;lAT(lHC,2X,*AVE''>AGr FRICTION Hf ^.n* , 11X, *FT ./"I LF. * , .1 4X , FI r . 4,
4950 +?(ir'X,F10.4) )
4960 19S8 FOR'IATClHr ,?X, *VNE :^GY P?ICh* , 2TX , * C.~NT3/
-------
5P60 --'I.iT 1. iT, nri'-ihT (NTOT,14) ,PCnC,TGPuP
5100 t-PTJ-JT It 55
5110 r-'PINT l:.f>n
5120 -"-PINT lr,6&,cFG-'MT(;!TOT,lrt) ,GPDC(lf ) ,°C(13)
5130 'PTNT Ir7f-,Sf 'GNNT;}) ,GPI-C f 2P ) ,PC (2 3)
5140 "'PINT Ic7r, SrC.^NI (NTP1 ,2? ) ,GFPf (21) t^CCZl
5150
5160
5170 r-RTlT l!:-nr, SFG"NT(NTOT,1P) , ~, PCC ( 1 9) , PC (1 9)
51flO ^PT-IT
5190 ^F IJT
5200 C PRINT SHC^T f'PINTOL'T
521C ?l^i1 IF
522C 21?C IF
52^0 217C
5?40 ?140 FOr>"'AT{ ///,7X,*nDOT*f1 2X,*QF A1?:*,!?'/,* f I AM* , i?X , * 5LOFF* , ?X ,
525C **COr4Vt YANf-P TY- ~», ^X, »CKMT S/GCC hi * , 4X , »OrLl AHS/MG VI f RQrjUf TTUN*
5260 +,/)
5270 ?150
5300
5310 21oO
5320 +13)
5330 H LOOP ,?ri
5340 3COO rONTINUf
5350
5360
5370 ^G?5 GO TO ( 7C30 , ?2b rA') (•>, i ID PL^TLT
5440 -»TH-r lll^LKTL'. (I) ,'-.LF VR ( I) , r LF Vr (I) , T f: 3 F/U (I) ,GNSFAC(I)
5450 303€ fONTINUr
5460 r-0 TO ?ltr.
547Q ^C4C« 'LAT(r.,ni)
5480 "PT-IT 111,
5490 GO TO 31£(.' •
5500 3045 "FAntD,11D FLLF ,FUt F, P.6SC L , R" T
5510 ^F?IMT 111,^LL^ ,pULF,BASrLtRrT
5520 GO TO 3150
5530 30?0 3FAO (5,111) G VCKWH,DE:GC,H J)LTM,r FF , £PS
5540 ff-»INT 1 .•* 1 » GVCKWHjHFGr ,HOL IM, :; FF , cpS :
5550 3150 IF (EOf,^) 2?sO,l5
5560 ?26C STOP
5570 JrMH
5580 SUITOUTINF OFT I H (TM JSX , VflLUb , XLf , XHI,CON:VX ,f.ON V-Vf NC ALL ,KCWGO)
5590
218
-------
5600
5610
5620
5630
5650
5660
5670
5680
5690
5700
5710
5720
5730
57*tO
5750
5760
5770
5780
5790
5800
5810
5820
5830
58tfO
5850
5860
5870
5880
5890
5900
5910
5920
5930
5953,51,51
IF (VALUE-ANSMN) 60,90,70
THISX = (THISX+XMIN)/2
IF(THISX-XMIN) l«t 0,55,11*0
IF (M) 71,71,180
K=0
IF(N-l) 100,100,110
BUHP=(-1.0)*BUMP
N=0
GO TO 28
PX=THISX
PV=VALUE
GO TO 16
IF(PV-VALUE) 18,120,18
THISX=(THISX-i-PX)/2
GO TO 130
N=0
K=PV=PX=0
GO TO 27
IF (VALUE-ANSMN) 56,55,56
PX=PV=0.0
IF(K-3)12,12,80
BUMP=2.0*BUMP
K=0
GO TO 12
BUMP=BUMP*0.25
GO TO 150
NOWGO=1
IF(ILO)220,220,230
IF(IHI)17,17,250
PRINT 2*fC,ILO
FORMAT(IX,*THERE HFRE,*I3,* ATTEMPTS TO GO BEYOND
*LOWER CONSTRAINT.*)
GO TO 17
o PRINT 260,IHI
260 FORMAT(1X,*THERE HERE,*,13,* ATTEMPTS TO GO BEYONU HIGHER CuNSTRAI
219
-------
6150
6160
6170
6180
6190
6200
6210
6?20
6230
6240
6?50
6260
6270
6280
6290
630C
6310
6320
6330
6340
6350
6360
637C
6360
6390
6400
6410
6420
6430
6440
6450
6460
647C
64BO
17
56
VAL'J: -At f p 4
THIiX-XHIN
GO TO ?^u
PX=THICX
oO TO 2.'
130 IL 0=1101
36F HUMP=(- 1»
NO TO If
2GC IHT=IHT+i
i'l, THISX,XLO,XHT,f!r*LL
-*,'?.i2.?>,* xLr -*,r;i2.•••>,* XHr =*,GI?.C,* M'/>LL
,1 J)
THIfX C'.ITSTO' YHT-Xll IIMIT~*,/,* ''UN cBO-;TLL! I'i O^IM*)
'--'Kl H ,C,vrK'HH f STtTT , lYrA?)
tP r'Lin. CMr ^GY VIA STM", VT ,"^, fN~' i;>fN:sJiL '•'WH USt
u
6500
651C
6520
6530
6540
6550
65f.C
6570
6580
6590
6600
6610
6620
6640
6650
666Q
6670
6680
• NO
fUNGTICr
*I '-ii-Nl! ON r.'.'ST ( aq)
C N;"'TT *) wr3T VTFGJI?IH («*7> WISCONTTN (!•?) WV^'-'TM^ (^9) NATIONAL.
f*ATA "MiT/l .46^, 1.6Q0,l.^f2,l.7;>n,l.r.^4,l.rtftf->,1.76ci,l.c:":i,
•s-1 . 4 11,1 . ?^C , 1 . 7!o, t.' Cli*» 1. ^,1 . £77, 1 . 4 32, 1. 4'-'i, 1 . 7:.'r! ,1 . ~::,-7 , 1 . ° ^ .' ,
<-1.9j7,
<-:. 7-^,1.^1,1. ,'q7,i.?qc,i. !?3, i/? '.^ , i. .:>r-7,i.os? ,i.f-4r>,'7.i:!A,j.?':
«•; . 7 -5 ' ,1 . 7"£i « 1. * "T4, . c'*3, 1. i'.b? tl .7 P
-------
67GC >? '••*>•:•-!= :'.ri (••Tn- )*?."! ^7'CY
6710 -rT'!.>;•;
67?0 op v-/iLr^i • (Vf>i/i,;:-))
67^o "<(•'''= ••;•!,<•• T( ".T/.T _.)*(-*.'-• .I-,-'?-} + .•
67^0 + -,!•>"-;.(• 1 *X**fc+ .••' j.-91^t 7*x* *
67SO --T.J-:;,,
6*20
695P
6960
678D r 'T'",'!:? >-"T10 PV ^I-'LTN" "nt:T ir .^'"T AT
679f c c'.. /•'•;-;.
6800 i r (••>,•;!_ Pf-:r D ;.,, t;, s
6870 :',(-• Tn lr.
6°OC '! F < ^A"Cl -3 H ;: .) v n ,1 •'! , ?
69ic ic •^.•;rA'j=ri
f,O Tr Qi-
(Continued on next page)
221
-------
SUBROUTINE MOODY (Q, DIAM, KEY, F, ITYPE)
C DARCY FRICTION FACTOR BY MOODY CHART AND COLEBROOK AND
C AND WHITE FORMULA. ALSO REYNOLDS NUMBER AND FLOW TYPE.
COMMON/MOOD I/VIS, EPS
REY=Oy(.50762082*DIAM*VIS)
R=REY
IF(R-2000.) 9,9,351
9 F=64./R
ITYPE=1
RETURN
351 ROUGH=EPS/DIAM
RUFF=ROUGH/3.7
GO= -2.*ALOG10(RUFF)
IF(R-4000.) 9,10,10
9 RN=.0006275
GO TO 100
10 RN=2.51/R
100 GN=-2.*ALOG10(RUFF-rtlN*GO)
IF(ABS((GO-GN)/(GN) -.0001) 170,170,150
150 GO=GN
GO TO 100
170 GN=(GN4GO)/2.
F=(1./GN)**2
IF(R-4000.) 11,20,20
11 R=ALOG(4000.)
REYLG=ALOG(REY)
F=ALOG(F)
R2=ALOG(2000.)
F2=ALOG(.0317)
FL=(REYLG-R2)*(F2-F)/(R2-R)+F2
F=EXP(FL)
ITYPE=2
RETURN
20 TURBF=(200 ,/(R*ROUGH))* *2.
IF(F - TURBF) 29,30,30
29 ITYPE=3
RETURN
30 ITYPE=4
RETURN
END
7190
7^30
l.'l;. tA**AV>
(7)
.L-^/A^AYCl)
' M !.•>
222
-------
8530 FUNCTION COSTM ( TTYP ,N°tG, I Y[ Af?)
8540 C THESE COST INDICES AP^ LKP PROJECTIONS FOR YEAt GT
8550 C....FOR ITYP GT. 13, NOT RfIG TON ALI7:i'. N?£G = C....
8560 C NREG CODES! 1=ATLANTA, ?-»ALTTMOPE,3 = 31
8570 C 5=CHICAGO, 6=CTHOINNCT1, 7=CL rVEL AND, 8=DALLAS, ^DENV^P, :
8580 C 11 = K'ANSAS CITY, 12=LOS ANGFLFS, 11=MJN!ar. APOLIS, 14 = NCW Os-'L;
8590 C 15=NEW YORK, lf>=PHTLCOFLFHIA , 17=PTTTSRLRGH, ie=ST, LOUIS,'
8600 C FRANCISCO, 20=SFATTLE, AND 21=XA1ICNAL.
8610 C ITYP Cm£Stl = PCT, 2=CCI, 3=STP, 4 = S, b = FTPF_LINF A')JUSTM = NT
8620 C....STP AND S MOT IN YET. ITYP f> TO 10 L^FT FCP LATf-R
8630 C....OF REGIONALIZED INDICFS....
8640 C 11 = AVG. HOURLY EARNINGS MFG ,:/HR., 12=CHcM. rKJf. PLANT C.I.,
8650 C 13=HS CHF.M1CCL PROCESS EQUIPMENT INOKX, 1U-WPI FOP
8660 C INORGANIC CHfMICALS, 15=r,0*Pr/STT!-~ "IPILINE UOST IM"cX, 1.fi = ;fL'
8670 C FNE^GY CI, 17-CONCRfTr 0AM, 1« = EA?TH 0AM, 19=Pl;M?ING PLANT,
8680 C 2C = :JIOnERN HFG MATNTFNANCF INDEX.
8690 C EXAMPLES (1,2,1969) = "C 1, 3ALTIf'CF•£. , 1 969 . (11, ,19/a)=AtfG. HRl Y .
8700 C EARNINGS, NOT REGIONAL IZf~n , lq?0 .
8710 DIMENSION SLOct(11^),OF°T(115)
8720 HA T A SLOPEV16. ?n5, 15. '»°0, if . ftp.9,1 f-. 709, ?1. 3' 2, 17. Oi'
8730 +, 14. & 92, 21. 282,14.^93,17. 072,14. 94 7,1 7.0 *)F.f Tl. 7"3,1 <
8740 +?9. 6«7,r"2. 796, 16.5^0,17.630,26.1^1!, 25. 36rf,?L.f}6?, 30.3^0,7 f, ,?vn?,
8750 +7Q.9f)2,39,716,?1.9?C,25.Q90,39.1C(-,27'.&«8,^7.75F;',32.62?,?«4.7Q6,
8760 +47.112,27.647, ?3. 3^5, 37.54», 39.206, 33. 43-J, H?. 187, b^'O . ^ , 0 . 0777.,
8780 DATA CEET/ 369. ,434. , Tib ., 426 . , k? 0 . , k°-2 .', ^76 . , 17 G . , k b J . ,'-18^. ,4 73.
8790 + 41 8., 49 C., 416. ,47C. ,530.,438.,446. ,757. ,4C4.,'*3l:;.,ul8.,!45r'.
8800 +567.,651.,B30.,77S.,!':»4.,5P4.,ft00.,ri(31.,53e:.,619.,blt].,61?.
8810 + 771. , 67^. , 654 . , 711. ,6 05., 4?* 0 . 0 ,1. 'J 39, 3. 9P 1 ,1. U? 3 ,1
8820 +?, 0.888,1.100, 0.775,Q.781, 0.900,0.360,0.9*7,1.176,1.13£.,1.0,l.'J3a,
8830 +1.0,Q.82C,n.e£t3,i.9»1.3?33»i:>5.?fl4»137.2*,3F.3'J9,72.
8R40 +10 1 .52, 6 9. ^95'3 ,"53.^/
8850 NO = 5
8860 IF (NRF&) 998,150,173
8870 150 IF {ITYP.LT.11.0R.TTYP.GT.20) GO TO 900
8880 TND = NC*?l+(ITYP-iC)
8890 GO TO IftO
8900 170 IF (NRE&.GT.21) GO TO 91.Q
8910 IF (ITYP.LF.O.OR.ITYP.GT.5) GO TO 9GO
8920 IND = 21*(TTY°-1)+^RFG
8930 180 COSTM = CF nT (TNO) +SLOFE (IND) * ( TY E AR-1 94"»)
8940 P.FTURN
8950 993 PRINT 210,NRFG
8960 RETURN
8970 9QO PRINT 901,TTYP,NRFG
8980 PETURN
8990 910 PRINT 210,NWER
9000 RETURN
9010 901 FORMAT (14HO INVALI'J TYPE ,I3,12P FOR REGION ,13)
90?0 210 FORMAT (1HO,2GHINVAlID REGION COOF ,13)
9030 tNO
223
-------
NAMES OF VARIABLES
ARRAY(7) Array used in cost computations
AVFAC Average of pressure factor (PRESFAC) over length of line
BASCL Base pressure class to which the pipeline costs are assigned, psi
CGPDM Unit investment per mile, 0/gpd mile
CKGPM Unit production cost, £/Kgal per mile
COSTN LK-R library function subprogram for cost indexes (variables
unique to this program are not contained in this list)
CNSFAC(I) Construction factor for segment I (applicable to pipeline invest-
ment cost)
CNTYPE(I) Name tor conveyance type. 1 = gravity, 2 = boosted, 3 = assisted
pump, 4 = pumped
CONSF Construction factor for pipeline in segment being computed
CPMI Current year unit investment in pipeline, $/mile
CYENI Current year electric energy cost index
CYMNI Current year Marshal and Stevens Cost Index for Chemical
Process Industries Equipment
D,DOLD Initialising values of diameter for iterative search
DEGC Expected average water temperature, degree C
DENS Density of water at expected temperature, gm/ml
DIAM Current inside diameter pipeline, feet
DIAMHI Upper constraint on diameter for OPTIM optimization, feet
DIAMLO Lower constraint on diameter for OPTIM optimization, feet
DITPL Amortization factor for pipeline, taxes, insurance, plus capital
recovery, fraction/year
DITPU Amortization factor for pump stations, taxes, insurance, plus
capital recovery, fraction/ye r
DMGMI Unit conveyance cost, $/mg mile
EFF Wire-to-water efficiency for pump stations, fraction
ELEVBCO Elevation of beginning point, segment I, feet, msl (mean sea level)
224
-------
ELEVE(I) Elevation ending point, segment I, feet, msl
EPS Pipe roughness, feet
FIRM Firming factor, ratio of installed capability to firm capability
for pump stations, fraction
FLTYPE(I) Name of flow type. 1 = laminar, 2 = critical, 3 = transition,
4 = turbulent
FMOOD MOODY friction factor
FMT Format number
FRACPR Fraction for second order adjustment of pipeline cost at other
than base pressure class
FRHBAR Friction head at QBARE conditions, feet/mile
FRHDOT Friction head under design conditions, feet/mile
GPDC(I) Unit investment, £/gpd, for investment component I
GPDP (I) £/Kgal cost for component I
GVCKWH Current year energy price imposed as an input parameter for
state and consumption quantity, £/Kwh
HDLIM Limiting head on pump stations, feet of fluid
HDSTA Total dynamic head on pump stations, feet of fluid
HPSTAF Firm station horsepower, HP/station
HPSTAI Installed station horsepower, HP/station
HWCK(I) Energy price, £/Kwh, for segment I
IPRES Index for pipe pressure class, 1 = 100 psi, 2 = 100 psi, 3 = 150 psi,
4 = 200 psi, etc.
IPRINT(6) Printout instructions as in program description
ITDOT Holding variable for ITYPE
ITYPE Index for flow type
IYEAR Year of estimate, called current year, four digits
jjj Indicator for whether gravity versus pumped comparison is to be made
KK Index for conveyance type
LLL Index for number of comparisons, pumped line costs against gravity
line costs
LOOP Data card on which loop is to be made
225
-------
MOODY LK-R library subroutine for MOODY friction factor (variables
unique to this subroutine are not contained in this list)
NCALL Number of calls to OPTIM
NMSTAT(2) State name (two seven-character words allowable)
NOWGO Indicator in OPTIM, - 1 = optimum not yet reached, 0 or +1 = opti-
mum reached
NPCLS Number of pressure classes of pipe used in line
NREG Region number according to code in Function Subroutine COSTN
NRUN Number of runs to be made
NSEG Segment number
NSET Case number
NTOT Column of SEGMNT array for summing values of segments
NUMSEG Number of segments in the line
NUMSTA Number of pump stations
OPTIM LK-R library subroutine for one-dimensional optimization (vari-
ables within this subroutine are not contained in this list)
PC (I) Percent contribution of cost component I
PCOPCT Operating cost as percent of total cost
PLLF Pipeline life, years
PLMILE(I) Pipeline length, horizontal projection in segment I, miles
PMILE Length of segment being computed, miles
PRCL Pressure class pipe, psi
PRCLMX Highest pressure class used, psi
PRLAB Labor price in 1968, $/hr
PRLOMR Current year price of OMR on pipeline, $/year per mile
PRPUMP Current year price of pump stations, $/station
PULF Pump station life, years
PUMILE Interstation distance, pipeline miles per station
PUOMR Current year price of OMR on pump stations, $/year per station
PYEX Payroll extras factor, fraction of payroll
QBAR Value of QBARE or QBARA currently being used in computations, mgd
QBARA Average amount actually conveyed, mgd (not used in present version)
226-
-------
QBARE Expected average conveyance over the project life, mgd (total
mgd conveyed in project life, divided by total number of days
in project life)
QDOT Design capability of pipeline system, mgd
QMAX Maximum day conveyance, mgd
QMIN Minimum day of conveyance in project period (not used in present
version)
RET Interest rate, fraction/year
REY Reynolds number
REYDOT Value of Reynolds number under design conditions
RTLAB Current year labor rate including payroll extras, $/hour
SEGMNT(I,J)
SLOPE
STATE
T
TER
TERFAC(I)
TGPDC
TGPDP
TINVPM
Array to hold data concerning segment I. J is the type of data
Average slope of pipeline in segment being computed, feet/mile
Numerical code for state, as found in Function Subroutine CKWH
Internal diameter in inches
Terrain factor for pipeline in segment being computed, fraction
Terrain factor for segment I (applicable to pipeline maintenance),
fraction
Unit investment, 0/gpd of capability per mile
Unit investment, $/gpd of average production per mile
Total investment per mile, $/mile
TOPCKG
TOPCST
TX
VIS
VLINE
VPUMP
VTOT
Operating costs,
Total operating cost, $/year
Tax rate, fraction/year
2
Viscosity of fluid at expected temperature, feet /second
Line investment, $
Investment in pump stations, $
Total investment, K$
227
-------
WCCHPL Annual capital charge for pipeline, $/year
WCCHPU Annual capital charge on pump stations, $/year
WENGY Annual cost of energy, $/year
WLOMR Annual cost of OMR on pipeline, $/year
WPOMR Annual cost of OMR on pump stations, $/year
WTOT Total annual production cost, $/year
XINS Insurance rate, fraction/year
YRKWH Annual electric energy consumption, Kwh/year
228
-------
CHAPTER 5
COMPUTER PROGRAM FOR PRELIMINARY DESIGN
AND COSTING FOR ACTIVATED SLUDGE TREATMENT
MODIFICATIONS MADE TO THE GOLD PROGRAM
There exists a computer program for design and costing of
activated sludge plants, termed the GOLD Program (52) . The
original GOLD Program was modified, the ultimate modification
being named GOLD2. GOLDl was an intermediate stage (53). The
flow chart of GOLD2 is shown on the next page.
A driver program was added at the beginning consisting of a
DO loop to repeat the program with several different data decks.
The modified program is executed twice for each case, once at
Q(20)=QDOT, the design capability, and again at Q(20)=QBARE,
the expected average production. The first execution with
QDOT sets the equipment sizes required to achieve the speci-
fications at a maximum production rate QDOT. Then the stream
characteristics are determined for a flow rate of QBARE through
a plant whose equipment sizes have been fixed by QDOT. The
excess capacity factors of the original -GOLD Program which are
intended for a similar purpose are eliminated. The MLSS at
QBARE is set back to the level corresponding to the increased
detention time, approximately UBAR times the MLSS at QDOT.
(UBAR=QBARE/QDOT). No attempt was made to analyze transients
while flow rate is changing. Both QBARE and QDOT are steady
state flow rates.
The GOLD Program sizes the blowers large enough to supply the
oxygen needed if nitrification occurs, and nitrification does
occur if the calculated time required to achieve nitrification
(TAN) is less than the detention time in the aerator. In
GOLD2 it is possible for the aerator detention time to be short
enough to avoid nitrification at production rate QDOT yet be
long enough to have nitrification at the lower flow rate,
QBARE. In this case the program determines the production
rate QNIT, the largest production rate at which nitrification
still occurs, determines the size of blowers required at that
production rate and compares this blower size with that pre-
viously calculated at a production rate QDOT. The blower size
is set at the larger of these two, and a message is printed
below the cost information stating that the blowers were sized
by nitrification at QNIT.
GOLD2 uses the Engineering News Record Building Cost Index for
temporal adjustments of costs. The base period is January 1960
(BCI National = 554.4), which is the GOLD report basis. GOLD2
provides for any other BCI value and runs were made with
BCI=732, for San Antonio 1969.
229
-------
Run entry
Case entry
no
Calculate Operating Costs
Adjusting blower size and
Capital Cost
Sum up capital and
Operating Costs
Print Stream and Other
Parameters at Avg. Flow
Start DRIVER Pgm.
DO Loop
'
1
Read Data
| Start 7 DO Loops
IFLOW = 1
Set Q(20) equal to
QDOT or QBARE
LOOPS 6
Main Program Block
Calc . Stream Param
GOLD2 ACTIVATED
SLUDGE PROCESS - FLOW
CHART
Re-entry for QBARE
Calculate individual
Capital Costs
Print Stream and Other
Parameters at Design
Flow
IFLOW = 2
230
-------
The GOLD report basis for operating costs is a study dated
1966, costs presumably as of 1965. GOLD2 assumes that this
is the base (BCI National=634.4), and adjusts operating costs
with the current year BCI mentioned in the previous paragraph.
The GOLD Program places no constraints on the size of equip-
ment. GOLD2 provides for replication when a maximum limiting
size of equipment is reached, and also for at least two units
if the sizes are above a certain minimum. The constraint
values are as follows:
Use at Least Two
Equipment Maximum Size Above This Size
Settlers 30,000 sf (square feet) 2,000 sf
Aeration Tanks 1 mg up to QDOT=200
Then linear to 6 mg over
QDOT=1,200 5,000 cf
Thickeners 50,000 sf 200 sf
Vacuum Filter 800 sf 20 sf
Digesters 200,000 cf (cubic feet) 10,000 cf
Incinerator 600,000 Ibs/day (pounds
per day)
In addition, for incinerators, there is a constraint of a
minimum sized unit of 8,000 Ibs/day, via Subroutine UNITEX.
It is intended that this minimum concept and this subroutine
will later be applied to all equipment.
GOLD incinerator operating costs go through a maximum at high
Ibs/day. GOLD2 uses the unit cost at this maximum for costs
above this Ibs/day at the maximum.
The GOLD Program provides a fixed input for the fraction of the
time during which vacuum filters are operated. According to
Sewage Treatment Plant Design Manual of Practice #8 (54) small
plants operate with as little as 30 hours/week, large plants
require up to 20 hours/day. The fraction of time for the
vacuum filters (TVF) is made a computed parameter, 0.2 below
QDOT = 1,0.8 above QDOT = 100,and TVF = 0.2*QDOT**0.30103, in
between.
One of the input parameters NFORK(5) = 1 is used to eliminate
the primary settler and its attendant costs. NFORK(6) is used
to suppress a printout of certain selected parameters.
231
-------
If the GOLD Program is run with typical sewage to produce an
effluent BOD which is rather high, say 30 mgpl or more, the
resultant BOD loadings, Ibs BOD/day per 100 Ib MLSS (mixed
liquor suspended solids), become very high. This is generally
considered undesirable as a reasonable sludge volume index
cannot be maintained at such high loadings. The sludge volume
index (SVI) passes through a minimum at some intermediate
value of BOD loading (see for example 55). The GOLD Program
and GOLD2 provide a fixed SVI value and thus the model is not
sensitive to the real relation between BOD loading and SVI.
To at least eliminate excessive BOD loadings GOLD2 contains a
message and a constraint which aborts the run if the BOD
loading is greater than 70.
The vacuum filter loading (VFL), the solids concentration ratio
in the final settler (URSS) and the temperature (DEGC) are
added to the variables which can be changed from case-to-case
within the DO loops.
GOLD does not provide for ultimate disposal of the sludge residue,
GOLD2 uses a cost for disposal of incinerator ash of 1.5 $/ton
of ash. (This is not in the program listing herein.)
Instructions for running GOLD2 and a sample printout are found
beyond.
232
-------
GOLD2 RESULTS FOR SAN ANTONIO EXEMPLARY CASES
The GOLD2 Program was used to develop design and costs for
activated sludge treatment for the plants involved in the San
Antonio study which were:
Capital
Name of Plant QDOT, QBARE, Cost Annual Cost, K$
mgd mgd m$ Op. + Op.
Amort.* Only
A new plant 234 118 38.48 4721
Existing Rilling Plant 80 40.7 - - 704.
U.C. Salado Plant 24 12.2 - - 244.
Existing Leon Creek 12 6.1 - - 138.
* Operation plus amortization
The information used in the present project, so far as it has
gone, is the investment and annual cost (operation plus amorti-
cation), in the new plant, and the costs of operation only in
the Rilling, Salado, and Leon Creek plants (all San Antonio,
1969) . The concept is that the investment in the existing and
under construction plants is already sunk and does not enter into
the comparison with advanced waste treatment and reuse. The con-
tinuing costs in the three plants are only the operation costs.
It would have been possible to obtain actual operation costs for
the two existing plants from the actual records and for the under
construction plant from the engineers' design. However, to handle
the general case it is necessary to have a program which will
generate such costs, and it was so used here. It is gratifying
to note, however, how closely these computed operating costs
compare with the experienced costs. The operating costs
indicated for the Rilling and Leon Creek plants sum to
842 K$/year. The program is such that the operating costs
are entirely a function of equipment size, i.e. of capability,
and not dependent upon actual throughput. The City Finance
Department's projection of the 1968-1969 fiscal year costs for
"direct cost plus administration" (equivalent to operating cost
here) was 795 K$. Thus, the computer program figure is within
6% of the actual experienced figure.
233
-------
Those familiar with the San Antonio situation will recognize that
the capability given for the Rilling Plant is not that normally
associated with that plant. The QDOT used is 80 which is actually
close to the average production of the Rilling Plant as now
operated. The capability of 80 mgd is an estimate from the
treatment plant management of the capability when operated to
reliably produce an effluent of 18 BOD, 18 TSS. The values of
QBARE for the plants are those which would be achieved in plants
of the capability given operating under the seasonal fluctuation
for San Antonio as described in Chapter 3.
The raw sewage was taken as average San Antonio sewage, com-
position shown in Table 28. The amortization factor used was
0.07783 comprising 30 years • at 4% plus 1% taxes plus 1% insur-
ance, (and is very close to 20 years at 4.5% without insurance
and taxes or 25 years at 4.5% with 1% for both).
234
-------
TABLE 28
AVERAGE COMPOSITION SAN ANTONIO
SEWAGE USED IN GOLD2 RUNS
Name Description Value
SOC Solid organic carbon 124
SNBC Solid non-biodegradable carbon 30
SON Solid organic nitrogen 12.4
SOP Solid organic phosphorus 2.7
SFM Solid fixed matter 35
DOC Dissolved organic carbon 51
DNBC Dissolved non-biodegradable carbon 11
DN Dissolved nitrogen (organic + NH3) 20.3
DP Dissolved phosphorus 5.4
DFM Dissolved fixed matter 628
DEGC Temperature degree C 25
ALK Alkalinity as CaC03 282
TSS Total suspended solids 220
Total BOD 250
NH3 nitrogen 15.2
PO4 24.9
235
-------
INSTRUCTIONS FOR RUNNING GOLD2
This program is set up to run in Fortran IV on a CDC 6400
computer. There are about 1050 cards. The compile time is
about 25 CPU seconds and six peripheral seconds, the execute
time about two and four resp. About 15 of the 25 compile
seconds accrue from the extensive PRINT and FORMAT statements,
The program will compile with a core memory of 100,000 60-bit
words. It will execute from a compiled program or binary
deck with 23,600 core words.
The following data cards are needed.
First Card:
Next Cards:
Number of runs to be made
Numbers to be assigned to the
various runs being made, one card
for each number
Format No.
809
809
The RUN Loops re-enter at this point.
Next Card: NFORK instructions 114
NFORK(l) = lf provides sludge
drying beds instead of vacuum
filters; zero provides vacuum
filters. (Note: Neither Ref. 52
nor GOLD2 actually provide the
sludge drying bed option, so zero
must be specified.)
NKORK(2) = 1, bypasses printout of
stream parameters; = zero prints
these parameters.
NFORK(3) = 1, bypasses printout of
other plant parameters; = zero
prints these parameters.
NFORK(4) = 1, bypasses printout of
individual component costs and
prints only total costs; = zero
allows all costs to be printed.
NFORK(5) =1, eliminates primary
settler; = zero incorporates
primary settler.
NFORK(6) = 1, suppresses printout of
certain selected parameters; = zero
prints these.
Two Cards, Sewage
Parameters (11):
SOC, SNBC, SON, SOP, SFM, DOC, DNBC, 101
DN, DP, DFM, ALK
236
-------
Six CArds, Plant
Parameters (24)
URPS, XRSS, CAER 20, AEFF 20, DO CKWH, 102
AF, GSS, TRR, TSS(12), GTH, GSTH, ERR,
TSS(15), WRE, GE, 'GES, TDIG, TD,
TSS(16), SBL
BCI, SVI 801
The NCASE entry is at this point, the various cases within
each RUN being made up from combinations of the succeeding
data cards. But each RUN has the NFORK instructions and
sewage and plant parameters from the preceding nine cards.
2-11 Cards:
2-11 Cards;
2-11 Cards
2-11 Cards:
2-11 Cards:
URSS data; the first of these cards 107
gives the number of URSS values to be
explored, the remaining cards the
individual values, one per card.
DEGC data; the first of these cards 107
gives the number of DEGC values to be
explored, the remaining cards the
individual values, one per card.
MLSS data; the first of these cards 107
gives the number of MLSS values to be
explored, the remaining cards the
individual values, one per card.
DEMBOD data; the first card giving 107
the number of data items, the remaining
cards the individual DEMBOD data.
2-11 Cards:
QDOT and QBARE data; the first card
gives the number of pairs of values
the remaining cards the actual pairs
of values, one pair per card.
FRPS data, same general style as the
preceding.
802
2-11 Cards: VFL data, same as preceding.
107
107
When the data deck is set up in this way the program will produce
the number of runs specified on the first card, each run according
to its own set of NFORK instructions and sewage and plant para-
meters. Within each run the program will produce one case for
each combination of the data on URSS, DEGC, MLSS, DEMBOD, QDOT
and QBARE, FRPS, VFL7 a total number of cases in each run equal
to the product of the number of data items in the seven classes.
237
-------
Any of the sewage or plant parameters or any other suitable
parameters could be brought out into the NCASE loops by so
modifying the program or the parameters brought out could be
returned to the sewage parameters set or plant parameters set
by a reversing modification.
A sample printout follows.
238
-------
INPUT DATA, RUN 36
0000000000
124.0000 30.0000 12..4QQO 2,. 7000 35.TCOO 51.0000 41.0300 20.313° 5.4.UOO
282.0000
400.0000 ,62,00 l.OOOC
.05000 i.00000 .01000 .0778? EOQO.OJOCC
.9500 60000.UOOG 750.0000 9.0010 .76Cr
60000.0000 3.000C flOO.OOIP 9.0000 33.3000
15.0000 200.0UOC ^.(tPQC
732.00 100.00
1
3.00
1
25.00
1
6000.00
1
18.00
1
234.00 118.OQ
1
.50
1
4.90
239
-------
OOOT = 274. CO1"1
QBAPE = 110. oar
UBAR = .50!*
PAPAMFTFP
MILLION
MILL TO-
GCLL ON!S/ DA Y
GALLONS/HAY
CONDITIONS
HPOT
AT
UNITS
EFFLUENT 300 16.Cf
BCD REMOVAL 92.8?
BOD LOADING 61.1C
AER. OETEN. TIME 1.07
NITRIFICATION C
EFFLUENT NITPftTE Q.OT
EFFLUENT NH7-N ?0.£4
EFFLUENT TSS 15.71
MIXED LIOUOR SS 6HOO
SLUDGE RETURN .97
DISCHARGE 237.95
VAC. FILT. LOADING 4.9H
INCINERATOR ASH 'l.f-P
BUILDING COST INDEX = 732.0
AMORTIZATION FACTOR = .Q7783
NUMBER OF
PRIMARY SETTLERS
AERATORS
FINAL SETTLERS
THICKENERS
DIGESTERS
ELUTRIATION TANKS
VACUUM FILTERS
INCINERATORS
COMPONENT 0» ITE^
17.99 wr,/L OXYGEN
51.IP: L'VDAY °CD P£*
2.0° HOURS
C (l-Y'.S, r=NG)
0.00 M5/L NTTi=CG£N
20.+4 MG/L NITROGilN
20.43 MG/L MASS
3,116 MG/L MASS
.04 FRACTION
117.97 MILLION' GAL/DAY
4.90 GAL/HR-SF
15.57 TON/DAY
100 LF MLSS
SIZF OF FACH
I'NITS
6
10
H
2
1
c
p
?9
1
29
7?
26
6^6
91
.176
.07T
.617
'7.\7
.766
. fl 6 3
.278
vj^F
f TLLICN
K?F
KCF
KSF
SF
TCN'/TAY
GAL
CAPITAL COSTS
K-DOILAPS PERCENT
AMORT. PLUS OPERATING COSTS
K-i/YEAP CENTS/KGAL* PERCENT
PPELTM. TREATMENT
PRIMARY SETTLE0
AERATOR
AIR BLOWERS
FINAL SETTLER
SLUDGE RET. P'JHPS
CONTROL HOUSE
SLUDGE THICKFNER
DIGESTER
SLUDGE ELUTPIATION
VACUUM FILTRATION
SLUDGE INCIN.
ASH DISPOSAL
CHLORINATION
SITE DEVELOPMENT
TOT. CAPITAL COST
AS CFNTS/GPO OF
ODOT
3903.22?
1 267.9"* fl
3 1 7 2 ^ "^ 5
2^34.4^2
46 75., r 17
9Q4 .267
2271.404
7 S 46.277
C.COO
20".4'Q
9Dt .592,
2.
10.
10.
3.
13.
5.
12.
2.
•5 .
2.0.
• o.
•
2.
01
^1
14
2 P.
64
31
15
•JCT,
90
-?o
Tf1
53
If
563
328
71
1
5
6
7
2
77
07
70
00
9
80
73
•
•
*
,
•
•
•
%
•
•
*
5P,r,
677
9 If1
004
379
29 C
369
84"
674
56?
^
1
1
1
.3317
.762°
. Q Lf /^ p
• 1 6 o ^
.4038
.1772
.1634
.1333
.7747
.0229
,651f
.1678
3
11
10
6
8
1
5
3
. 10
1
12
16
q
1
.07
.95
.21
.96
.65
.52
.23
.63
.74
.49
.71
.19
..21
.94
.49
38483.772
TOTAL AMORTIZATION AND CFfRATION C
OPERATING COST ONLY
CFNTS PER POUND OF TOO PFKOVED -
CENTS PEP POUND OF TSS "fMOVEn =
240
OST
17?6.15G
c.6578
4.7219
4.0378
ino.oa
36.56
-------
100 PROGRAM GOLDflNPUT, OUTPUT, TAPF5 = IN'CUT,T4PF6 =
110 HIMENSIOM N^UNCVO)
120 niMFNSIGN ODOT (10) ,nnAPF(in) ,crncT <15) ,'ftOCr-f KIT)
DTMFNSIfN EBGO{?_) ,TPOOP.-(2) , BCOLO (?) ,OTEN (2) , NTT ( ?)',
1TOTSS (2) ,XNJ03(?) ,XNH7(2))TASH(2)
OTMFN^ICN Q (20) ,SCC(?0) ,S*J3C (?0) ,^CN (?0) ,SOF (20 ) , SF!* ( 20) ,
1 nNaC(20) ,ON(?n> ,OP(2Q) ,OFM(20),SPOC(20),CPrD(?n) ,CCO$T(1 '<=) ,
? COSTO(15) , aCOST(l^) ,ftOCO?(l"5) ,^FORK(1Q) ,A?S<12) ,FVEOD(10) ,
3 VSSC20) ,TSS(?C) ,ALK(?n) ,rPFRK(1 ?),ECF(15),P?0(10),<=PSIN(in),
" ^ QOUT
2«»0 00 25 IRLIN=1,KRUNS
250 no ?CO T=3,?0
260 fMI)=0.r
270 SOC(T)=C.O
280
290
300
310 SFM(I)=C.G
3?0 POC(I)=r.O
330 nNflC(I)=0.0
3*»0 ON(I)=0.0
350 DP(I)=0.n
360 nFM(I)=C.O
370 SRnnd)=o.n
380 OSOO(I)=0.0
390 VSS(I)=0.0
»»10 300 ALK(I)=0.0
ueo no i»oo 1=1,15
CCOST(I)=0.n
COSTO(T)=0.f1
AOCOS(I) = 0.0
cc°CTm=r. o
AOCPCT(I)=0.0
500 ASB^O.O
510 C ...... READ AND PRINT TN»UT DATA FOR PUN
520 WPITP(LIST,8t»1) NRUN(IRIJN)
530 ^PAD(KAPD,lltt) (NFOPK (T) ,T = 1 » 1 0)
5^0 WRTTE(LIST,«70) (NFQRK (T) , 1=1 , 10 )
550 RFArHKAP.0,101) SOC(?n) ,^NBC(2n) , SON (20) , SC"= (2H) ,SFM (?Q ) ,Q(1C
560 1DN3C(20) ,PN(2H) ,OP(2Q) ,DFM(20) ,ALK(20)
570 WRTf-(LTST, ^71) SOC(2Q) ,SNRC(2f))tSCN<20),SOF<20)tSFM2Q),roC(?0),
580 inNBC(20) ,DM(2n) ,DP(?0) ,
590 VSS(?0)=^OC(20) *2.13
600 TSS(?Q)=VSS(?0)+SFM(20)
610 SBOO(20>-
620
630
PFAO(KAPD,t02) URPS, XFSS, CAEP2, AEFF2, nO.CKWH ,,Af ,GSS ,
241
-------
650
660
670
680
690
700
710
720
730'
750
760
770
780
790
800
810
820
830
8UO
850
860
870
880
890
900
910
920
930
950
960
970
980
990
1000
1010
1020
1030
1050
1060
1070
1080
1090
1100
1110
1120
1130
1150
1160
1170
1180
1190
1TRR,TSS(1?) ,GTH,GSTH,FRR,TSS(15) ,WRE ,GE ,GES,TOIG, TO ,
2TSSC16>,SBL
WRITE(LTST,872> URPS, XRSS,CAER2, AEFF?,DO, CKWH, JF ,GSS,
1TRR,TSS(12> , GTH, GSTH, ERR, TSS115) ,HRE,GE, GES ,TOIG, TO ,
2TSSC16>,SBL
REAO RCI,SVI
WRITE(LIST,873) 3CI,SVI
CIFC = ECI/55't.t*
HIFO =
QCL2=8.
C.......READ AND PRINT INPUT DATA FOR CASES
REAO(KAPD,1CI7) NUP, (URS (I) , 1=1, NUR)
WRITE (LIST, 87U NUR, (URS(I) , 1=1 ,NUR>
RFAD(KAPn,107) NTHP, (DEG(I) , 1=1, NT^F)
WRITE" (LIST, 87M NTMP, (DFG ( I) , 1=1 ,NTMP)
RFAD(KAPD,107) MAS, (ASS (I) , 1=1 ,h'AS)
WRITE (LIST, «7
HRI7E(LIST,875)Nn, (ODOT(J) ,OBAPF
HRITE(LTST,ft7«t) NF1?, (PPSTN (J) , J=1,NFR)
READ (K/SPD,107) NVFl , { VFLD ( J) , J=1,NVFL)
WRITE (LIST,a7**> NVFLt (VFLO ( J) , J = 1,NVFL)
C ....... ...START 7 DO LOOPS ..........
NCASE=1
00 25 LF=1,NFR
FRPS=RPSIN(LF)
00 25 NU=1,NUR
URSS=URS(NU)
00 25 NTM=1,NT«P
OEGC=DEGCNTM)
80 D025I=1,NAS
XMLSS=ASSII)
D025K=1,NBOO
nOD5=EMPOO(K)
00 25 NV=1,NVFL
VFL=VFLO=QDOT{LQ>
GO TO 9C*t
903 0(20)=QBAPF(LQ)
90** URAR=CBARE(LO)/QOOT(LO)
908 LOOPS=6
C MIX STREAMS NINF AMD TWENTY
67 IF(LOOPS-f) 66,65,6f
65 0(9)=0.0
242
-------
1200
1210
1220
1230
124Q
1250
1260
1270
1280
1290
1?00
1310
1320
1330
66
1760
1370
1390
1400
1*410
1A20
1««50
11*60
1470
1490
1500
1520
153t)
1540
1550
1560
1570
1580
1590
1600
1610
1620
1630
1640
1650
1660
1670
1680
1690
1700
1710
1720
1730
1740
+TF.
SNRC (1) =
0(l)=0(?n)
-SN3C(l) )*1 .87
= (DOC(1)-OM«5C(1))*1.«7
PPTfARY SFTTLER PFPFOR^ANCF
IF (NFORK15)) 1091, in«l , 1092
1092 0(8)=n.
Q(2)=f3tl)
TEMP1=1.
r,0 TO 1C93
1091 0(8)=FRPS*Q(1) /URPS
1093 SOC(?)=TFPP1*SOC(1)
SNBC(2)=TEHP1*SNBC(1)
SON{2)=TFHP1*SON(1)
SOP(2)=TFfPl*SOP(l)
ONBC(2> =DN8C(i)
nN{2)=DN(l>
nP(2)=OP(l)
TSS(2)=VSS(2)+SFM
-------
1750
176P
1770
1780
1790
1ROO
1810
1820
1830
1840
1*50
1660
1870
1880
1890
1900
1910
1920
1930
1950
1960
1970
1Q80
1990
2000
2010
2020
2030
2050
2060
2070
2080
2090
2100
2110
2120
2130
2140
2150
2160
2170
2180
2190
2200
2210
2220
2230
2240
2250
2260
2270
2280
2?90
C
1000
85
86
9G5
1001
1002
1003
1006
1007
10Q8
1004
1005
906
907
70
50
42
7
19
(*
5
6
8
20
VSS(8)=SOC(8)*2.17
DB03(8)=(DOC(8)-ONRC(8))*1.S7
AF.R4TO«? PFRFCRMANCT
CEn> +09002
FMAX=FOOO
N=l
GO TO 8
FPl?r)!?=Ff/Sy-FMlN
TOL=.10
TF(rRfi>or-TOL)?l,?l,19
19 FOOO=(FMTN^F^flX)/2.n
IF(FOCO-DPOO(2))5,5,6
5 OBODCt)=DPOn{2)-FOOn
SBOHU) =SPOD(?)
GO TO 8
6 SBOOCt) =0(2)-FOOQ)».7C
8 TEMPl=(.f5*Fron/XMLAS)-XP^S
0(7)= (Q(P)*TFMPl-CrD'?*VAEl?)/(URF5?-XRSS)
20 0(5)=Q(2)-0(7)
244
-------
270C IFIN-21 ?~,?3,??
2^10 ?2 TFMP?=XFSC*Q (5) +1)055*0 (7)
2320 XHLPS=O (2> *SPOD<<*) /TFMP2/.30
2330 SBOnC^) = {XM|_AS*.685 + XMl_nS* .80)
2560
2570
2580
2360 GO TO ?.k
2370 27 TDnr)5-x^LPS*
?*t TF(Tqcn5-EOn«=) 10, ir),l«5
10 TF(N-3) 1J,1?,13
11 FMIN= CCrOF*V/5F'VQ(.'>)+XP 'XMLSS/.65
C-.Q TO
707 FO^MATf/* OFfAMD MLASS CANNOT PF H^LO, GASP
GO TO 9P&
?«»70 PC TO 7
15 IF(K'-l) 16, 18, 17
16 WRITF(KTYPF,^06» NCASF
2^00 306 FOPMAT(/» R005 OFMANO CANNOT RE ACHIEVED, CASE *,I3)
2510 GO TO 98«»
?5?0 17 FMIN=FOCD
2530 IB GO TO 7
2^0 21 CONTINUF
2600
2610 TF{ABS(TFMD?)-TOD ^ !,£»!, 51
52
. GO TO 50
2650 53 ASMAX=XMLAS
2660 GO TO 5*1
?F70 ^1 CONTTNUF
2700
2710 SNBC(7)=SNBC(5)*UPSS/XPSS
2720 TEMP2=XFS?*XfLA?/?.«*6
2730 ^ON(5>=.23't*TfH0?-KSOC(c5)-TEHP?) /10.
2750
2760
2770 SOP(7)=SOC(7)*.01
2780
2600 nOC(5) = CMPC(2) ^-OBOn(it) /1.87
2810
2820
2870 ON^CC?) =ONRC(5)
nN(5)=(0(?) *{SON(2)+nN(?))-(SCK'(5) *C (5) +SON (7) *0 (7) ) ) / (Q ( 5) +0 (7) )
245
-------
2850
2860
2870
2880
2890
2900
2910
2920
2930
29«fO
2950
2960
2970
2980
2990
3000
3010
3020
3030
3PUO
3050
3060
3070
3080
3090
3100
3110
3120
3130
3150
3160
3170
3180
3190
3?00
3210
3220
3230
32«»0
3250
3260
3270
3280
3290
3300
3310
3320
3730
331*0
3350
3760
3370
3380
3390
C
1070
1071
C
1075
917
916
91*»
C
910
26
27
911
913
918
DN(7)=ON(5)
OP<5)=(0
IS BOO LOADING GREATER THAN 70
IF (RCOLO-70.) 1075,1075,1070
PRINT 1071, 8CH5, NCAST
FORMAT (!HO,*aon LOADING APOVF 70 AT THIS
+1X,*AT
GO TO
COMOITTON? FOR NITRIFICATION
Q(6)=(Q(?}*(1.-.65*FOOO/XMLAS)+CEOR*VAER)/(URSS-1.)
FFFLUENT EDO OF*,F7.2,
RETUR=0(6)/0(2)
X«»Xl=(.l.+PTTUR)/RETUR/URSS
CNIT=.18*EXP(.116*(nrGC-15.)J
TAN=(l.+RFTUR)*(ALOG(Xt»X3)+tt.e05/
-------
3 1> 0 ft q 1 c; f -,' B r L fl G - 1
?«?70
3670
7700
3710
375P
3760
3770
3780
3800
3810
3e?o
3830
3650
3«60
3870
./n t
C r ALCUtr T r, c j
P(in)=n (7) +Q
~(f ON (7)
tn)= (pr (7)*Q(7)
{n)-(rr'M7)*o(7
(in) =sc ~ (in ) *?. "
son ( ID -TF Mot.*?nr;(
VSS(11» =^0^ (111*2.
SOP(ll) -
(1.1 .) =
0)
1 0)
ATM?- 0(1 n)*T?s (10>*J>.3T'./G5:7H
SFMd?) =
nP(l?)=pP(in>
VSSd?) -^P
aTHi.=o(ir)
g?o
3920 C
3930
*Q(")
) *0 (7)
O'" (7) *Q(?1 -t-nnp. (ft
) = (PMFC (7)*" (71 -fPMEi.c (?) *C (PU /O (10)
247
-------
3950
3970
3980
4DQO
40in
40?0
4C90
MOO
1+170
4180
4?50
4?70
4?»0
4?90
43?0
4730
4340
43?0
4360
437H
4380
44HO
4t«10
4420
4430
4440
4460
4£»70
44 BO
TnoPT=(i.-
.?) ) /CtOIG
4? TO
44 niG13=C?r7N/ (CiniG*TD-1 .)
TFMC4=( PIC 12-0 TGI?) /(SOf{l?)+rcC
r: (i?)
C (1?)
(1?)
nFM(13) =
92? V9IG=Q{3?)*TC*lr>00./7'./»q
CONTINUE
rflLCULAT
M SLUOGF
45
TFMP1=Q
Of
VSS{17> ^
7) =T
=TF'-'P1*SCN(1?)
-T«~
=7F
248
-------
** 5 1 P
'« ^ ? 0
1 0
«*660
<»710
<»8flo
4 c f o
1*970
If°80
5010
50?0
tut -i r "??
^ F '•< ( 1 '4 ) r T F ' « r. ? * ^ TT -i ( -I •? )
TtrMr:l=0(i?)/{0(i"::)tr(i7))
(17)
(1 ^)~
60
PM< 1 A
rp(i. =T'r"!FllfALK<17) + T!rM^
,ALK(15> -/si . K(l't)
Yl = in:].*flL
Y?-VSS( IF)
H rflLCULATr
nM vyaCUUM
SS (I?)
1 r)
(1 ^^ -TF
*Tf
OP
n M T C ( 1 fl ) = 0 >! 'T (1 ^ ) * T
nOH (1 8) - PPr ( I.1') * TF
Tr MC'=T5 c(16 ) /TCS(I
N (16) =
5)
nFM(i&) -
249
-------
5400
5410
5420
5430
5440
5590
IF (IFLrW-t) in-°0»lC'PO,928
5060 C VARY VF TI'4? 'WITH
5070 1080 IF {Q(?D-
5080 1081 TVF=.?
5090 no TO 9?7
5100 1092 IF (0(20)-
5110 1033 TVF=.'<
5120 GO TO 927
5130 10B4
5140
51*0
5160 0(9)=Q(11)
5170
5180
5190
5 ? 1Q SN^CC^I—TEM^1*SK' 8C f 1 ^ j-fT^M^^^SK^CCl^) + T ^"M P 3 * S Nl P H
*???0 ^nN(P)~^E^c>l*^ON(ll.)4pTF^^?'*SOK(J^)^'TfrM^;^*^CN(16)
5240 cjFrH{9)="TFHDl*SFMCll)^-TEMP^>'vSF^{'*i*)
5?50 D
5270 DN(9) =TF f F1TN (11)+-TFMP^*nN ( 14)+TFMP3*ON{ 16)
5280 0
5300
533P
5340
5350 929 LOOPS = LCOPS-1
5360 IF ( LOOPC) *>fi , 66 < ^7
5370 C STOPE P£FfMrTrp? FOP
5380 46 roMTTNUF
5460 PTPNCIFLCW*=Tft*'4.0
5480 9P-T(TFlPH)-PTTUP
5500 TF(TFLOK-l)
551(1 C HALCULfTF CfiDTTAt TOST?
5520 93C CCf
5^30 IF
5540 1096 APS=0.0
5553
556P
5570 ^0 TO 10Q7
250
-------
5fOO
5610
5P30
5f50
5660
5670
5690
5700
571 n
57? n
5750
5760
5770
5780
5790
5800
5P10
58?0
5870
5880
5P90
5900
5910
5920
5930
501,0
5950
5970
59BO
5990
60CO
6010
6020
6030
6040
6050
6060
6070
6090
eion
6110
61?0
6130
TF (APS-?.)
TF (A=^-Tj.
6fli MPS = ?
GO Ta 9V?
f95 NPS=IFI
GO TO 9??
ST (? ) =
r.cosi
937 S
10°7
TF (QFS-?.)
IF (&FS-30.)
610 »JFS=?
r,o TO 9.^F
615 MFS
GO TO
(-H. 81
{?) =U°OST?*AFS*l(1CO.
(f> ) =3^ 50 . +°2!?'1 . * Q ( ^)
0 ** 7
937
IF
C-0 TO 9"70
GO TO a.rP
93 « MTH=!
qiq
-------
6150
6160
6170
61SO
6190
6200
6210
6220
6230
6? 40
6250
6?60
6270
6280
6290
6300
6310
6320
6330
63
TF (N«?FLAG) ?i»
CONTTNUF
IF -'. » , ,» ,
COSTOd?) = P..nr;*YTONS
GO TO a/4^
9^5 rOSTO(l?)=1.6.1*VTOf'!?-. 0 OHO 9* YTC'-F** ? .
9'»6 COSTO(l£')=nni''* (OC^O-Q (17)
IF(K'FCRKd)
f,0 TO
9«*9 no 941 J=l ,15
9M roSTO(J)-rn«;Tn(j)/ino'1.*CTFn
C CALCULAT TOTAL
31
CENG-3. fg* ( 1 00 '3 . /TOTCO ) **0 . 1.
ACOST (J)=AnnsT(J)*nrp
252
-------
C.71Q
6770
67°0
67°"
6P10
6C7D
7010
7p?o
7070
7^60
7070
7080
7090
7100
7110
71?0
7150
7^60
7170
7180
71°Q
7?00
7??0
op rrr^-T (, | ) = r " n c; T ( J ) * ~. r "
jr. r n--;-n .r
"! 0 •? ? J - 1 , 1. c;
J)
"•'-CIS (J) -fln^T (.)) H-rpTTn (J)
CD-^K (J) =f nrpc: ( j; /T f 7r ) /^.
T o T * n =T r T r o t n r r s ( ..n
Trn-Tc^TnsTr+fnS'f 1 ( J>
rfic>{|)f-ir!^./T^T'>''
5 f i j " n n P r - ( r ( ^> n ) * T n PP ° r - { C ( 5 ) - Q ( !• 7 > ) * T
WTSorir> (0(?^)*T^S(?r:)- (0(r, )-0 (1 7) ) ^
CWP^n (7 > *Tnn7/o c3) / T n r ?
rr~' = j. .-Trsf*c(r)/'nf?)/Tnr?
jonof-'si . -T^rpn *o (f ) /^ ( ?) /"""-^np1
1-^M (P) ) ^O f ^) /n (?) / (S
'
-" p ~ - n n <~ tv _ r !.; ~~>
•• ' -WM-I . - (<:.fU!
(?) ) .
IF (IR_
P-, R COMTTNUr
Tf(IFLOV'-j)
Of,fj '.;t U1^ ([ JCT j
^ T f?
C 0 T n q f ^
M?irr (L 1^7 , « 0- ) ° (? 3 3
op-5 MPTT- (LI?T,11->
(Li?T,^i7) Nsr.f ,n (i) ,ior (n ,snn t
l^FM(fl), VSS (8) , TSlr( ?)
(L TST , T0!+) ror (°) ,r^nn (•?) ,r•^'^^ t P).,TM (8) ,DP («) ,DFM ( «>
(LIST,
\'ST.r. ,n (5) , "en m , ^'::o'o. (5) , ^t-
i: > ,?or c?) ,
253
-------
7250
7?60
7?70
7780
7?«>C
7?00
7710
77?0
7770
771,0
7750
7760
7770
7780
7790
71,70
7*00
75?0
.7570
755P
7560
7570
7600
7670
7P60
767H
7680
77 OH
771P
77?0 C
777C
7750
776-C
7770
77PO
WPITC (LIST, 717) NST£,0{7) , ?OC (?) , cam ( 7 ) , TK ^C ( 7 ) , ^ON ( 7 ) t ST^ (7 > ,
<^M(7), VSS(7),T<>S(7)
WPITF (LIST, 3Hi») 90", (7) ,01011(7) ,PNPC<7) ,HN (7) ,1P (7) , C-FM (?)
N?TA=1Q
WP I IE (LIST, 717) NSTA,0(1Q) ,SOC(1P) , OOC d?) ,nnCD (1 0) ,CN"?C ( 10) , O'dQ ) , OP (13) ,PCM ( 1 H)
NSTA=11
W9ITF(LI<:T ,717) NSTfi,0(1l) ,SnC(ll) , SHOO ( 11) ,<5N9f: ( 11. ) , SCN ( 11) ,
SOn(ll) ,SFH(11) , VSS (11) , TC<; (ll )
W9ITF(LIST,-'ri») PCCdl) ,0-3 OH ( 1 1 ) , rvar ( 11 ) , f> ( 1 1 ) , no d 1 ) , npM ( j j >
WRITFd.IST,?!'?) NSTP , p ( 1?) , (15) ,nP(l^) ,nF»'(
t'STS-16
K!?ITF(LI«;T,'?17) Nc Tf 1 0 (16) .^OC (1 6) , S°nn ( If-) ,SNT- ( 1 6 ) , SCN dc )
1 SOD(15) ,SF^f If ) ,V?3d6) ,TSSd?) <
WRITT(LTST,7r/») HOC (1 5) , QPCn ( 1? .) , CN = C ( 1'.) , TK d^ ) , np (i t. ) ,OP" (
MSTA=9
WPTTrdisT.^j^) ricTp ,n(Q) ,scc
1SFM(«?) ,<'??(°)»TC?{P)
5T,7r'4) OPQ(T)
C ) « ^nOrl ( ? D .
, ON PC ( ? 1) , H
WPIT£
HP TT- (LIST, 7 17) VST A , WD ,SO^ (
1«) ,T5S(tf»)
f op
clfl^T F A ~ « "t Tf 3 c ..........
CCMTTN'jr
V,'PTTr(|_TcT, j Or ) XN-! . /• S , v^L" c , XL Nr S , X VL OS , X^L IS , XliLSS , V OF S ,
T,RrTUf,r:^',C:""'%r FTOC ,ij=-c^, (JDCTC ? vnc c jCc c, fl PC , f.^S , A^S ,GT h , r^T
MP 1 7 r (L T^T , i i ? > T n T r , T n , \/n jr, , c •? r ir , r i. n in , r r r i -, , CH 4 C F , C r ?r r , vc L ,
254
-------
7800 1TVF,M/F
7810 WPITr(LT^T,ll?)
7820 1AIRCF ,rFP<~* , r?T7r
7830 WRTTrai^T,! IS) PfCt. "*, t L K ( 1.) , «L " (1 ? > , AL«< (1 3 ) , *,L K (1 ^
7850 HP:
7860 IF
7870 10!*2
7880 WRTTF (LIST, 701 ) flr
7890 10^3 W°TTF (LIFT,831)
7900 95P TFdFLOV'-l) cr> 1, 9C n , T1? 1
7910 Q5P TFLCW^?
7920 ^0 TO Prl
7930 951 fPN!TINUc
79«»0 C . ..' FPTf^T OPCTjOpn.p^,^?^0
7950
7960
7970
7980 TF (NFOFK(o))
7990 C. ......... P0 TKT NFO^MF) PF-'^ftlK ^"LrCTFr PADA*-'FTF?S
8000 9BC Wt?TT-(L35;
8010 WRTTF(LI5r
8020
8030 WPTTF (LIFT, 055) T-^nnp-d ) ,Tanncr(3)
80**0 WOIT" (L TFT, 8C7) QpHLH (1 ) , R CPLr (? )
8050 WPITf (LT?T, ssfHQTTNd) ,PTF "'(2)
8080 WPITF(LIST,.°Fl)yMHT'(l),XNH7(?)
S'JPO WPTTF(LT,rPERK(it),ACCcCT(M
83t*0 WPTTF (L TFT, P 17) CC^ST ( ft ) , PPPCT ((r),ACPO?(b),'"DCpK(t),ACCPCT(f)
255
-------
fi35H , WPITr(LIST,Pia) CCOFT ( 7 > , OCDCT (7 ) , AC-CO S ( 7 ) , f P£R K ( 7 ) ,ACCnC T ( 7 )
8380 WRTTF (LIFT, «2?) cnsT(ii) ,cfcpCT ( ID ,-soros
85*»0 WPITi- (LTcT,fnO)
8550 TF(KBFLfG)1
8560 983
8580
8590 25 CC^
8600 STOP
8610 101
8630 103 FOPMflTfFfi.?)
8650 1 9HyLNn. £*, 6'X, 6H W?F - Ft
8770 112 Fn?MAT{ P'X, 7HTOTC- - Fi^.^^XjC
8780
8790 ? 3X, 9HCH'*rr'n - £\ n . u, 7y , 1KCO?CCP = Fln .'4/°X,"6HVFL = E
8Bin 113
8820 1 3Xj9HCrrP"'^ - c in . L , ^y , 7H C AFC = p 1 T . <* /
8?30 2 ?5y,7y,oHflFFF?>"> - F1^.(i,5X,7H5rFF — "I"*,
8850
8P60 11«»
8870 115
88RO
8890 116 Fn°MAT(lCX,
256
-
-------
89?0
fl'VP
8°8Q
9?00
9010
906H
9P7C
9080
9f!90
9100
9110
91?0
91«*0
P15P
Q1?P
9?QO
9?7C
Q300
931P
9^70
9380
9^90
qt.00
9MO
= 513-.'l,?X,10HALK(l~) = clQ.'+,?y,10HAt.!«15) =
f 31JN NO. *,T?,*, rO?F NC. *,I3,11X,
1*L KR-GOI. D? ACTIv^Tc-p ^L'JTGE PcCrr"s> CALCUL M ION*, 17X ,*HF'
?//)
"in .<•)
. 1C7C»,
2 5X, 7HV 5^,9 X, T
31 a FCE?vftT(iy
3 n a T 0 P M A T { ? V , 1 ? H r M {, p c T r A, - T n •• I , i ;t x , T .1 n . '4 , 6 X ,
,
n
10
," 1 0. 6 , *>X , F 1 1.'*, 7 X, T 1 n .
K i X,lHe:i-!L ,T'«, Of- ??,(»
= F10.*-,?X,7MCTGC = -^"!
r = EI n.^jpy^HnrR = pin.««)
701 -o";-'!AT( F x, nH/vp - n_r.ir)
801 FC^-AT (?rn.?)
PO? FOPMiT(I7/(?F".?»
«'">3 FORMAT(/* "II. CWr^ ST7C ?CT ay MTTPTFIcaTT OH fT OMTT = "Fn.""
pnc ir0OMAT(* rONTITTONS a^ CFSTGN FLOW S3T", nPTT = * , F o . 3 ,
1* HILLITN ^flLLOMS/O^Y*, //)
806 FHR^aT( * COk'niTjn»!? flT rXFrCT^r flVERAGE FLHW, QEfl=E =
i* MTLLTCM GALLONS/H/JY*,//)
8G7 FODf"aT(lCy,*Tfl = *,Fin.«.,6y,*T/»M = * , E 1C . t; , ty, *r-:OOLC - *, =
*,F?.,1
808
8"q
81 P
811
81?
PI?
81 5
817
=• F i 1 . L
F n T f A P V 5:; c T Tl_ r ;
r- r f /s T ?'"
=' FJ 0. fe , 5X , 7HCMT =
, F9 . ?
C^f PCT.
POL HOUS
c 0 ° M A T ( •*
FO^^ATC*
FC°yAT(* FLUCC.F nPYTMf: "Fn?* , F 1 r . 7 , ? H . ? , Fl " .3 , F < 1 . it ,Fq . ?)
8?p pnoM/ST(* fML C?TNaTTnKj * ,F 1 r . •» ,F« . ? » Fl r ."», Fl 1. tf,Fc . 2)
8?7 FOf^MATC TOT. TAPJTf.L TOST *, Fin.-)
8?8 ^OP'-iAH* TOTftL ,A-MC°TI7A TTOM AKP OFr9ATTON CPST* , F 11 , T, p^j . ^ , pq . o>
FORMAT (* CFMTS '~!rP ~'nu>;n OF TS^ pi •'CVtD = *,F-a, ,f»)
R7i FODVAT (iHi)
p-r? priow/ij^/?* PUTI.PTMG ^OST T
257
-------
9450 840 ""0 3>'
9470 851
9480 P^?
9490 853 FORMAT ( T7X , *CONDT TT-MS AT*)
9500 8F4 FODHAT(*
9510 855
95?0 856
9540 R58 FOP:-«flT(* ACI3. HFTFM. T JVF* , Fl n . ^ , F 1 0 . ? < * Kf")Upc".*)
9560 P60 ^O^ATC FcFlUtNT MJT-\STT* ,F1 ~!. 2, ril. ?, * HG/L
^580 86? C09"AT(* tFFLU'NT T^^» , Fl<. . ?. ,F 1 r . ? , *
9590 863 FOP^AM* *-TXET LTOUP-? S?* , F9 . ° , Fl fl . 1, *
9600 870 Fn°MAf(!X,1111)
9610 871 FO^'ftT (lX,OFf«.4,4X,F8.4/1.X,?«:.«<.t)
9F20 87?
9630 87?
9640 874
965P 875
9670 +*Q^flRE *»F1Q«^)
9680 877 FfiPMHTC^yt'OFr^STINr; Cr^T rMLY*,l6X,Fll..3,Fl1.4,F°.?)
9690 878 FQP^ATC 1HO * *MJM!?FC ^F *,'HX,*c;T7r TF F ACM* , c X , *UM!T^*
9700
9710 87Q
9720 88n FORMAT (in ,*FiN'AL ?rTTir<=c;*,^x,iJ4,ex,pio.',7x,*sF*
9730 881 FORMAT (1H ,*THTCKr^c»<:* , °X, Hi , f X , F n . -* , 7y , '•
9740 P82 ^OP^AT (1H , *CTGFSTrDS* »lnX, 14 ,f X, Fl .1. ", 7X ,'
9750 883 FOO^ST (1H , *F. LUTPIf TTCN T ANKS* , ?X ,14 , 6X ,F 1 C . 3 , 7X , *SF* )
9760 884 rO!?'1AT (1H ,"VACUUM ^IL •'•IP ft* , c X , IJ., f X, FIT . T ,7X, *SF* )
9770 885 FCPMAT (JH ,*CLUDG~. »- TIJPN* , Fl 7 , ?, F 1 * , ?, 3X , *F? A<~T TON* )
9790 887 FORMAT (JH ,*VAC. FILT. LOA^T^'T:' ^S.^.FlO.c^X^GAL/HR-S1
9800 888 FCRMAT (in
9P10 rNH
9820
qp30 AMAXI=A NOTE: Some minor changes have been subsequently
9840 IF (A -P) 1C, 20,21 made to format specifications between
9850 10 AMAXl=P 879and888.
9P60 20
9R70
9880
9890 C FAIL-SAC" n-^T^: MTTH F\rNrTIr.N
9000 C SI7F. ANO L^A^T NUMP"C PF UNTTS TC FPCOUCc AT L'AST QTOT'L
9910 C WITH CN- LIVTT Q'JT. ^T if AST TWC 'J^ITS, rXCrFT IF TWO CF VTN.
99?0 C AVAILA^L? ST?? GTV- N'OPr THAM T~'T^L' TlOTiJ, TH£N ON-F UNIT, ".<
993H r, OFSTG*' Tf N^T FATL-"Ar'r.
9940 C
9950 GUN IT = rt-TK'
9960 IF n*nFOT'J-?*Ov!IN!) 5",70,7n
9970 50 »'UNTTS - 1
9980 MPITT (6,rq)
9990 5F FORMAT (+ TFSIG'J NOT FAIL-^n^?, VTA F.Xr«TTICM. *)
258
-------
1000P e-rTijpk
10H10 7r f'UN'TT^ -. ^«XO( T"T C-^nrTlJ/r -IflVf.-! !=-!) ,Ih'T ( OQO TU/0 N C X + 1 .-1 r - 1-
100?Q TP (NU^TTf-?)
100?0 HO t'UNTT" = ?
TF
lonen 100 OD^IT = ^^'T^;
10P70 tin ''CT l)i.i N
1C 080 CMO
259
-------
NAMES OF VARIABLES
In the original GOLD Report (52) the variables are listed and defined, and
in addition there are some variables in the program listing which are not listed
and defined. The following listing of additional variables does not include those
occurring in the list of variables or in the program listing in Reference 52. In
other words, it includes only those variables which were added during the GOLD1
and GOLD2 modifications.
AOCPCT(i) Percentage of the total amortization and operating cost attributable
to the ith process.
BCI Building Cost Index.
BODLD(i) BOD loading for IFLOW = i, Ib/day of BOD in Stream 2 per 100 Ib
MLSSin Aerator.
CCPCT(i) Percentage of the total capital cost attributable to the ith process.
CIFC Cost index factor for capital costs.
CIFO Cost index factor for operating costs.
CPBOD Cents/lb of BOD removed, amort. +op. cost.
CPTSS Cents/lb of TSS removed, amort. +op. cost.
DEG Array to hold DEGC values.
DTEN(i) Value of TA when IFLOW = i (hours) (Aerator detention time).
EBOD(i) Effluent BOD concentration for IFLOW = i, mg/1 oxygen.
IFLOW Flow index, IFLOW = 1 when Q(20) = QDOT, IFLOW = 2 when Q(20) =
QBARE.
IRUN DO loop index for runs.
KARD Logical unit designator for card reader,(5) in system used.
KRUNS Number of runs to be made.
KTYPE Logical unit designator for printer, (6) in system used.
LF DO loop index for FRPS loop.
261
-------
CHAPTER 6
CONVENTIONAL OR REUSE?
A COST COMPARISON FOR MUNICIPAL REUSE
IN SAN ANTONIO IN THE YEAR 2000
COST OF INDIVIDUAL COMPONENTS OF THE SYSTEM
Note: The comparative costs developed in this chapter are for
facilities and operations to meet the 2000 capability and
average production, but installed and operated at 1969 cost
levels. It is obvious that if a complete system were installed
in 1969 not all the separate facilities would have a 2000 target
year. Some might be built for 2020. Others might be built for
1980 with a plan for supplementing them in 1980. This would
bring about differences in cost which might affect the compar-
ative costs. However, to explore staging of this complex system
is far beyond the resources of this project, which seeks pre-
liminary comparisons.
Surface Water Reservoirs
It had been the decision of the AACOG Steering Committee to
allocate to water supply the costs of owning and operating
reservoirs in proportion as the conservation storage is to the
total storage. It was understood that this gives about the
same results as the incremental cost method. This policy was
used in costing on this project although we did not find the
time to explore the quantitative validity of it.
Previous studies have developed methods for generalizing on
required reservoir sizes and costs in various regions of the
country and for various yields as expressed as fractions of
average stream flow (40, 41, 43). Application of this general
estimating method gives costs of owning and operating reser-
voirs in Texas, 1969, in the range of about 0.6 to 2.5£/Kgal
of yield when this is at a 100 mgd level, and 1-SC/Kgal at a
10 mgd level, the range in the figures being the range between
yields of 5% and 80% of average stream flow.
Information was obtained on the estimated costs and yields of
the three specific reservoirs involved in the Cuero scheme
(37, 56, 57, 58). The data and computations are presented in
Table 29. The table shows storage volumes necessary to allocate
costs in the above policy. The percent allocated to water
supply is based on the • conservation and flood con-
trol storages. This has the effect of proportioning the sediment
storage among the two actual water uses. The percent allocated
to water supply for Cibolo is the average for two sets of storage
figures for Cibolo which differed slightly. The investment was
adjusted to 1969 by the USER Earth Dam Index. The withdrawals
for this project were those developed in Chapter 3, and shown
is the percentage of the yield which this represents. The OMR
costs shown are taken from the generalized costs developed in
263
-------
TABLE 29
RESERVOIR COSTS
Cuero I + II
Cibolo
Goliad
Capacity, Kaf
Sediment
Conservation
Flood control
Sub-total C + FC
Total
% allocated to W.S.
Yield 2020, mgd
Investment, 1969, m$
Withdrawal this project
% of yield
OMR, !969K$/yr
Capital recovery, K$/yr
Total annual cost
Allocated to W.S.
Cost, 0/Kgal
Annual cost, S.A. supply K$/yr
S.A. share of investment, m$
50
2,816
843
3,659
3,709
76.7
216.8
150
127.7
58.8
150
6,840
6,990
5,360
6.75
3,150
67.7
28
172
218
390
418
45.6
21.3
34.2
21.3
100.
70
1,560
1,630
744
9.55
744
15.6
42
958
702
1,660
1,702
57.7
102.1
61
113.1
100. +
111
2,780
2,890
1,670
4.49
1,670
35.2
264
-------
References 40 and 43. OMR costs for Cibolo are found in
Reference 58 in magnitude of about three times the generalized
OMR cost. However, the generalized costs are retained because
they come from actual or estimated OMR costs on a reasonably
large number of reservoirs. The capital recovery is based on
100 years life, 4.5% interest, no taxes and no insurance. The
total annual cost was allocated to water supply in proportion
as the percent of storage allocated to water supply. The
costs allocated to water supply were allocated to the San
Antonio supply in proportion as the withdrawal for the San
Antonio supply was to the total yield.
It is seen that the costs so allocated and based on actual con-
struction cost estimates are about twice those which would have
been obtained from the generalized study. Of course, a high
accuracy cannot be expected from a generalized prediction of
reservoir size for a given firm yield and reservoir costs as a
function of size considering the topographical and hydrological
variability over the nation. Thus, it is preferable to use
actual cost estimates for actual reservoirs at known sites, as
is done here, if these are available.
Pipeline Conveyance
Chapter 4 presents a computer program for costing conveyance by
pipeline. This was used to compute the cost for the following
lines under flow conditions as given in Table 33, beyond:
Cuero-Cibolo-Hildebrand
Goliad to Victoria
Rilling to Hildebrand
Hildebrand refers to the distribution storage tank on Hildebrand
Avenue in San Antonio which is the terminus of other conveyance
systems that have been investigated. Rilling is the site of
the Rilling Road sewage treatment plant.
The parameter values used were as follows:
Elevations
Cuero 242.5 ft msl (mean sea level)
Cibolo 400.1
Hildebrand 830.0
Goliad 200.0
Victoria 50.0
High point between
Goliad and Victoria 220.0
Rilling 579.75
265
-------
Distances
Cuero to Cibolo 34.2 miles
Cibolo to Hildebrand 38.3
Goliad to Victoria 28.35
Rilling to Hildebrand 10.0
Construction and Terrain Factors
Cibolo to Hildebrand 1.05
Rilling to Hildebrand 1.2
Others 1.0
Pipeline life 75 years
Pump station life 25 years
Interest rate 4.5%
Tax rate 1%
Insurance 1%
Labor price 3.00 $/hr
Payroll extras factor .45
Temperature 21.5°C
Head limit on pump
station 300 ft (feet)
Wire-to-water efficiency .75
Pipe roughness 0.0003 ft
For most of these pump station systems the energy price became
approximately IC/Kwh; except for the pumped-assisted segment
into Victoria where because of the low energy consumption the
price became about 4C/Kwh. However, the effect on conveyance
cost is minimal because the actual energy consumption is so
small. The complete printout for the Cuero-Cibolo-Hildebrand
line has been shown in Chapter 4. '
Comparison o_f Conveyance Costs with Previous Engineering Study
For Texas Water Development Board (TWDB) there had been prepared
a preliminary engineering study (36) which included the Cuero-
Cibolo-Hildebrand pipeline conveyance system. The engineering
report, which also covered the Lake Austin to San Antonio
conveyance system, largely canal, produced costs which are much
lower than those used in the present study, in £/Kgal 6.15,
4.92, and 4.34 for 100, 200, and 300 thousand acre feet per year
(Kafy) respectively, with electricity at 3 mils, 200 Kafy is
approximately 179 mgd and may be compared with the Cuero-
Cibolo-Hildebrand costs in this study for 149 mgd of 13.4<:/Kgal.
The ground rules given by TWDB for their study were quite
different from those judged realistic for the present study.
Some of them were as follows:
266
-------
TWDB This Study
Utilization factor 1.0 Ca .5
Pipeline life 50 years 75
Pump station life 50 years 25
Interest rate 3.5% 4.5
Insurance 0 1.0
In lieu of taxes 0 1.0
Energy price 3 mils Ca 10 mils
Pump efficiency .85 (.75)
Also, TWDB took slightly different elevations for the pipeline
termini as follows:
Cuero 230 242.5
Cibolo 390 401
In addition, of course, the TWDB year is 1966 or 1967, the
present study year 1969. The present case contains the gener-
alized pump station and pipeline costs for Texas while the TWDB
study used their own set of costs, some of which were specified
to them by TWDB. Taken together, with the fact that the present
study QBARE is 149 compared to the TWDB 179 most of these
differences in ground rules act to make the TWDB unit cost
less than the present study cost.
To check the two systems under conditions as nearly identical
as.possible the program was run with data corresponding to the
TWDB cases for 100, 200, and 300 Kafy. That is, changes in
the data were made for elevations, QDOT and QBARE, equipment
life, interest rate, taxes and insurance, energy price and pump
efficiency (wire-to-water efficiency taken as .806). All other
data were left the same as they had been in the Cuero-Cibolo-
Hildebrand case including the terrain factors and construction
factors, the water temperature and the year, 1969. It was not
possible to use the TWDB year 1967 because the cost index
projections used are not valid prior to 1968. The comparison
is shown in Table 30.
While it was not possible to compute the case for 1967 it is
possible to roughly adjust the TWDB data to 1969 by utilizing
the Building Cost Index for San Antonio in the two years, 732
and 629. The TWDB figures adjusted in this rough manner are
shown in the corresponding columns in the table in parentheses.
It is seen that thus adjusted the TWDB and present study costs
agree within a maximum difference of about 7%. The present
study is about 6.5% low at 100 Kafy (thousand acre feet per
year), about 7.3% high at 300 Kafy and within about .35% at
200 Kafy. It happens that 200 Kafy is the case closest to the
real case studied in this project. TWDB has 5.72C/Kgal
the present study has 5.70.
267
-------
TABLE 30
COMPARISON TWDB (1967) VS. PRESENT STUDY (1969)
CUERO-CIBOLO-HILDEBRAND
100 Kafy
Pipe diam, in. (inches)
Maximum pressure class
Number pump stations
Head, pump station
Total Costs
Investment, m$ 1967
Investment, m$ 1969
Total production cost,
K$/yr
Total production cost,
C/Kgal 1967
Total production cost,
$/Kgal 1969
200 Kafy
Pipe diam, in.
Maximum pressure class
Number pump stations
Head, pump station
Total Costs
Investment, m$ 1967
Investment, m$ 1969
Total production cost,
K$/yr
Total production cost,
<=/Kgal 1967
Total production cost,
<=/Kgal 1969
300 Kafy
Pipe diam, in.
Maximum pressure class
Number pump stations
Head, pump station
Total Costs
Investment, m$ 1967
Investment, m$ 1969
Total production cost,
K$/yr
Total production cost/
$/Kgal 1967
Total production cost,
<=/Kgal 1969
Segment 1
Cuero-Cibolo
Present
TWDB Study
66 58.3
250 150
1 2
383 266
37.5
(43.6) 32.0
1,912 2,190
6.15
(7.15) 6.72
84 77.0
250 150
1 2
434 257
56.5
(77.4) 54.2
3,371 3,719
4.92
(5.72) 5.70
102 93.7
250 150
1 2
399 212
73.7
(85.7) 75.1
4,565 5,138
4.34 -
(5.05) 5.45
Segment 2
Cibolo-Hildebrand
Present
TWDB Study
66
200
2
352,348
84
200
2
360,365
102
200
2
335,335
58.3
150
3
285
76.5
150
3
283
89.5
150
3
271
268
-------
The TWDB pipe sizes are somewhat larger than those of the
present study and the total horsepower of the pump stations
somewhat less. The TWDB total investment is greater than in
the present study and the operating cost less. Part of the
difference in operating cost comes about because the TWDB
costs for OMR on pipeline and OMR on pump station are con-
siderably lower than the present study costs. For both of
these the TWDB study used the ground rules laid down by TWDB.
We believe the present study costs for these are nearer to the
experienced costs. TWDB is able to use a smaller number of
pump stations because it allows a pump station head higher
than the present study constraint. If the present study con-
straint for limiting head on pump station had been 430 feet
instead of 300 feet the number of pump stations would have
been the same as the TWDB number. However, a limiting head
of 400 feet would not have changed the present study. The
higher limiting head allowed by TWDB in part explains the
higher maximum pressure class which they use, but some of this
also could be engineering judgment on the service required
regarding backfill conditions, water hammer, etc. which are
not plugged into the present study computer program (and which
are probably beyond the capabilities of a computer program).
Ground Water Withdrawal
Table 31 shows the investment cost experience on four San
Antonio well stations constructed between 1959 and 1968. Two
of the stations are secondary stations not having high service
pumps or reservoirs, and two are primary stations for which it
is seen the unit investment in reservoirs and building make a
large contribution to the total investment. The primary and
the secondary stations do not lie on the same unit investment
curve. The investment relation for the primary station is:
Investment, $ = 237,281*(mgd well pump capability)**.55984
269
-------
TABLE 31
INVESTMENT IN RECENT SAN ANTONIO WELL STATIONS
Station
Maltz-
berger
Wurzbach
34th
Street
Basin
Installed, capability, ragd
Well pumps 10 20 30 75
High service pumps 0 0 33 90
Number of wells 2437
Number of high service
pumps 0046
Number of reservoirs 0011
1968 investment, K$ 222.6 210.1 1,593 2,660
Unit investments, C/gpd
Reservoirs 0 0 1.05 .42
HS pumps 0 0 .42 .43
Wells and well pumps 1.98 .81 .67 .75
Site preparation, bldg,*
general .00 .02 2.56 1.71
Total including land 2.23 1.05 5.30 3.56
*building
The 1968 costs for the entire City Water Board system shows a
unit energy consumption of 1,44 Kwh/Kgal for the well stations,
an OMR cost for these of ,248£/Kgal and an energy price of
.802£/Kwh.
The computations for costs were made with a small computer
program (GWBXR, not listed) using the above investment relation,
a firming factor (installed capability/firm capability) of 1.4,
energy use of 1.44 Kwh/Kgal, energy price of 0.8
-------
References 40 and 43 present a generalized method for estimating
the costs of ground water production. The costs there are 1962
National and include in the investment only the well and the
pump, not the high service pumps, building and land. The cost
there given for a ten well station with an installed capability
of 75 mgd is 2C/Kgal, but if the investment used there is
adjusted to 1968 San Antonio and adjusted for the investment
increase attendant upon land, building and high service pumps,
a factor as shown in Table 31 of about 4.8, then the so adjusted
generalized costs for the comparable San Antonio station come
out to 3.5C/Kgal. Those wishing to use the reference for a
general prediction should multiply the capital recovery costs
there given by about fivefold to get correct costs for well
stations with buildings and high service pumps.
Water Treatment
If surface water were used it would be necessary to treat it in
a conventional water treatment plant by coagulation, sedimenta-
tion and rapid sand filtration. Reference 59 presents results
of a comparative cost engineering audit on treatment plants
of this type in sizes of about .5 and 8 mgd. Some additional
data on costs in large size plants was used to interpolate the
costs in a 230 mgd plant operated at 149 mgd average produc-
tion. Adjusted to San Antonio 1969 by the Building Cost Index
these project an investment of 18.55m$ and an annual production
cost of 3,240 K$/yr, about 6
-------
Balancing Storage
The monthly average curves of water distributed and sewage
collected such as presented in Chapter 3 show no average month
in which the water used is less than the sewage collected. If
this were also the day-to-day situation the advanced waste
treatment plant could operate on the sewage collected and the
difference could be made up by a fluctuating ground water
pumpage.
However, if there occur individual days on which the sewage
collected is greater than the water used then something must
be done with that day's excess "raw material." If the system
comprises an advanced waste treatment reuse portion in parallel
with a conventional treatment discharge portion then presumably
the day'e excess could be shunted to the conventional treatment
plant and discharged. However, if the system is a complete
recycle-no discharge one then a place must be found for the
day's excess. Not only must a place be found because of the
no discharge-no pollution requirement, but also if any
appreciable number of days' excesses are wasted, additional
input from ground water must be supplied. Thus, the total
system would have to be enlarged to produce this excess which
is wasted. To avoid these consequences requires a balancing
storage of some sort and of some magnitude..
It is very possible, of course, that in sSan Antonio this balanc-
ing storage could be accomplished by reinjection of the AWT
plant product back into the aquifer. San Antonio is sitting
on top of a tremendous no-cost storage reservoir.
But most cities are not so fortunately situated as to have a
convenient and readily accessible underground storage reservoir
at hand. Note that it is not necessary that the city obtain
its major supply from ground water. An aquifer storage arrange-
ment could be worked out if the aquifer is available even though
the city has a surface water supply.
The alternative to aquifer storage would be a- .surface reservoir
for storage. In any event there would be some cost involved
for this balancing storage.
This project has not attempted the assessment of that balancing
storage cost primarily because it has not been found possible
to work up the extensive data which would be required to deter-
mine the magnitude of the balancing storage. Involved is a
study similar to a reservoir sizing study, for which 'the
techniques are known. However, the fluctuating input comprises,
instead of the data on day-by-day stream flow, the data on day-
by-day excess of sewage collected over water used. This, of
course, will have many zero days or negative days which count
as zero days in the input regime.
272
-------
This project collected day-by-day data on sewage flows and
water flows for a number of corresponding years for San Antonio.
The sewage flow data were used in a number of statistical
studies including consecutive days of low flow and consecutive
days of high flow such as are involved here. The water data
were not used because the investigation promised to be a rather
major one.
If the quantities involved in each excess event are small the
balancing storage could possible be of the order of distri-
bution storage, some of which is normally available on an
existing system. The expected frequency of excess events
would be greater in the wintertime than in the summertime since
in the summer the water use is some three times the sewage
collected. It is quite likely, however, that the maximum
storage required for an event, which determines the cost, will
occur in the springtime months since these are the months of
highest rainfall in San Antonio.
Regardless of these speculations this project has not included
the cost of balancing storage because it has not determined
the magnitude of the storage requirement. It is left as an
unknown cost element which is applicable in the reuse scheme,
but not in the conventional scheme.
Demineralization and Disposal
As has been described in Chapter 3, the solution to the recycle
problem is not yet completely worked out and therefore a com-
plete material balance on the reuse scheme cannot be worked
out. It was shown in Chapter 3 that the blend would meet the
500 ppm TDI specification during only two months of the year
if there were no demineralization and no discharge.
With a given system any ion might be subject to buildup in the
recycling water, that is, in the blend. The RECYCLE Program
will provide for computations on about 50 of these and any one
of them may prove to be the controlling ion. For illustration
let attention be directed to TDI as a controlling parameter.
To reduce the TDI to 500 mgpl from the monthly level shown in
Chapter 3 it would be necessary to either discharge some water,
that is to increase the makeup, or to demineralize some of
the return. Our aim in this section will be to make some very
rough approximations to the cost of achieving 500 mgpl in the
blend under the San Antonio monthly average conditions given
in Chapter 3.
273
-------
If the blend is lowered to 500 mgpl by increasing the discharge
that is by increasing the makeup over the levels given in
Chapter 3, which will generate a discharge, major changes will
have to occur in the system. The amount of increase in the
makeup would be quite high because the makeup in the example
taken is already at 437 mgpl and therefore not very useful as
a diluting water. Also, the ground water supply is not far
from its limiting constraints. As will be shown the need for
discharge fortunately happens to be the least in those months
where the ground water withdrawal is the highest. Neverthe-
less, if the makeup has to be increased very much it will soon
come up against the ground water constraint and thereafter will
require a surface water supplement. This, of course, would
considerably change the entire picture with respect to reuse.
Demineralization seems the preferable alternative if it does
not prove too expensive.
It is not possible to arrive at the steady state figures for
demineralization without going through the recycle program.
However, as a simple approximation Table 32 shows the quantity
which would have to be demineralized, by electrodialysis, in
order that the blend of Chapter 3 meet the 500 mgpl TDI
specification in each month. This has been calculated by
simply computing the fraction of the return that would have
to be demineralized at 45% removal per stage in order that the
blend become 500 mgpl (but not taking into account that the
next return would then have a different composition).
TABLE 32
QUANTITIES TO BE DEMINERALIZED (45% REMOVAL)
FOR 500 mgpl IN BLEND
mgd
Month 1st Stage 2nd Stage Total Return
1 177 46 177
2 162 0 171
3 144 0 172
4 76 0 164
5 96 0 186
6 44 0 180
700 168
80 0 183
9 82 0 189
10 119 0 178
11 176 24 176
12 175 86 175
QDOT 327 159 350
Annual amount 104 13 177
processed -r 12
274
-------
During two months it would not be necessary to demineralize
at all since the blend is already below 500 mgpl. During
three months a 45% removal from the return flow would not be
adequate and it would be necessary to demineralize all of the
return in a first stage and some of the return in addition in
a second stage. This, of course, could actually be done by
arranging some of the stacks in series to double-stage the
demineralization. The cost computation here, however, com-
putes the costs as if the operation were carried out in two
one-stage plants. The design capability for the ED (electro-
dialysis) plant is taken as 1.85 times the highest monthly
average flow. This is actually not correct. The 1.85 factor
gives the maximum day in ratio to the maximum month at the 90
percentile level for the total sewage flow. The correct pro-
cedure would involve determining this ratio for each month and
taking that month giving the highest product of ratio times
amount demineralized. This would give a lower QDOT and a
lower cost.
The table is to be read as follows: Using January as an
illustration, the average total return is 177 mgd. All of
this would have to be demineralized in the first stage and
46 mgd would require demineralization in a second stage,
leaving 131 mgd to by-pass .the second stage. In May, the
fifth month, out of 186 mgd average return only 96 mgd would
have to be demineralized in the first stage leaving 70 mgd
to by-pass demineralization entirely.
Previous work on electrodialysis costs has developed the invest-
ment and operating costs for one stage (also for up to six
stages) working on water of average ease for demineralization
by electrodialysis and in a warm climate such as at San Antonio.
The data and costs are from Ionics, Inc. installations and
experience and the one stage cost actually applies to demin-
eralization from 900 mgpl to 500 mgpl TDI. The costs for the
equivalent 45% removal become slightly higher as the starting
TDI becomes lower, considerably higher as it becomes as low
as 300 mgpl, and become lower as the concentration becomes
higher than 900. However, the cost changes in the starting
concentration ranges concerned here are negligible compared to
the approximateness of the plant to which the costing is applied,
For a single stage the cost relations are well represented by:
Unit investment, C/gpd, _
(year, region) ~
CYCEI CYBCIREG 1 ,
109.7 CYBCINAT tAF l.259739+.0356967*LN(QDOT) J (1)
Unit operating cost, C/Kgal _
(year, region) at UBAR = .9
CYBCIREG ^ 1
687 EXP [ .42406+.0277584*LN(.9*QDOT)] (2)
where CYCEI = current year Chemical Engineering Chemical Plant
Cost Index
CYBCI = current year BCI, regional and national
275
-------
The operating cost equation correlates the operating cost in
plants of different capabilities in which QBARE is 90% of
QDOT. It is not correct to simply apply this relation to
situations in which QBARE is the variable in a plant of fixed
QDOT capability. In a plant of fixed capability some of the
operating costs at UBAR = .9 are fixed and independent of the
QBARE and of those that are dependent some are dependent
linearly, others not. Cost studies on electrodialysis (e.g. 60)
indicate that of the operating costs, for a 10 mgd plant for
example, about 10% are fixed and independent of throughput in
a given size plant such as labor, lighting, heating, etc. and
90% are independent upon throughput roughly linearly such as
energy, membrane replacement, etc. A small computer program
was written which would take the production schedule in
Table 32 and compute investment and operating costs in which
the operating costs according to equation 2 were aportioned
as 90% fixed and 10% proportional to QBARE. The resulting
investment, 1969 San Antonio, is 42.8 m$ and the annual pro-
duction cost 7,072 K$/yr. This amounts to 18.6<=/Kgal of feed
to the first stage.
When the RECYCLE Program is put to work on demineralization it
will be found that the quantities to be demineralized become
less than indicated here which has simply taken the required
demineralization for one pass to lower the concentration of
the return coming from a blend such that it will give 500
mgpl in the blend. However, the blend from which this return
came had a concentration greater than 500 mgpl. Not so much
demineralization will be required on the return that it is
generated from a blend already at 500 mgpl.
In the other direction the cost will be greater by virtue of
the limited recovery in demineralization by electrodialysis.
Electrodialysis, and also reverse osmosis, are limited in the
recovery which can be achieved which is controlled by the
concentration built up in the reject liquors. Recoveries of
95% are projected for the San Antonio water. At this rate an
additional amount of makeup equalling about l/20th of the
quantity demineralized would be required. This makeup increase
would amount to about seven mgd average over the year with 14
mgd being required in the winter months. The production costs
of this extra water would have to be added to those shown in
Table 33 beyond, but would not increase the total very much.
The 14 mgd requirement in the winter months fortunately comes
when the makeup demand is low and the total will not exceed
the ground water constraint.
Some means of disposal will have to be found for the reject
liquor which will have a concentration of the order of 10,000
mgpl in the winter. A pipeline to dispose of this to the Gulf
of Mexico would cost about 19 million dollars and the annual
disposal costs in it would be of the order of 1,100 K$.
276
-------
The reader can clearly see that the figures and costs given
for demineralization and disposal are of the very roughest
sort and only included to give an idea of their general magni-
tude. A number of alternative demineralization and disposal
methods could be considered.
Sewage Conveyance
Early in the project AACOG (Alamo Area Council of Governments)
had a need for costs of sewers. Neither the AACOG need nor
this project involved the small sizes of sewers in the collec-
tion system, since for this project the collection system would
be required for either the conventional system or the reuse
system. Accordingly, it was decided that the cut-off point for
the study would be a 12 inch sewer, which has a capability of
about two mgd. Sewer costs as a function of depth and diameter
were worked out as described beyond for those purposes. The
sewer costs were not actually used in the project because at
the present stage we have only compared a single AWT plant
with a single conventional plant both tentatively at the Rilling
site. Accordingly, the conveyance system, i.e. the trunk and
interceptor lines, would be the same for both alternatives.
There was prepared (61) a correlation of bid tabulations from
about 20 bids throughout the nation on concrete reinforced
sewers all deregionalized and adjusted to a WPC-S National
Cost Index of a 127.04 corresponding to March 1968. The bid
items correlated were simply those items giving per linear
foot costs for various sizes of sewer pipe. In addition to
these items the total bid normally contains also such addi-
tional items as paving, structures, rock excavation, highway
crossing, engineering, etc.
The data, supplied as curves for 15, 21, 27, 33, 48 and 60
inch sewers and various depths of cut, were correlated by
multiple regression analysis on the Cartesian equations, the
best resulting predictive equations being as follows:
for 12' to less than 30":
In (S/foot) = 1.48019+0.0412173*10+0.0268722* -'DEPTH
o- = 1.05
for 30" and up:
In ($/foot) = 2.28733+0.0227985*10+0.0134408*DEPTH
cr = 1.028
277
-------
The o" ratios shown above are not the true & ratios for the
original data points. Rather they are the o~ ratios for the
selected points of the curve used in generating the multi-
variate regression. They thus indicate how closely the pre-
dictive equations reproduce the curves rather than how closely
they reproduce the original point data. (The o" ratio is the
anti-log of the standard deviation in log units and corresponds
to the ratio of the 84th percentile point to the 50th per-
centile point.)
Obviously the total cost of an installed sewer per linear foot
must be obtained from the above predictive equations by multi-
plying the predicted $/foot cost by some factor which incor-
porates the other bid items. To develop such a factor, data
were obtained on recent bids for sewers in San Antonio,
specifically the Alpha, Beta, Gamma and Delta segments of
the Olmos outfall and two versions of the Salado outfall with
different joint types, the average of the two lowest bids being
used in each case. The ratios of the local bids to the pre-
dictive equation (for San Antonio area in the year of bidding)
fell'in two groups, one group around 2.0 and the other around
1.5 (actually 1.98, 2.33, 2.18, 1.44, 1.48, 1.49).
Various manipulations were made in the attempt to correlate
these ratios. The conclusion was that the ratio was not
correlated with the amount of rock excavation and not with the
average depth of cut but it did follow a measure "degree of
city streets" measured by the square yard of base supplied for
pavement per linear foot of sewer. The ratios around 1.5
occurred when this measure was less than 0.2; those around 2.0
when this measure was above 0.6. Further exploration indicated
that it was not the cost of the pavement itself or the culture
involved which brought about the difference in the ratios.
Rather the bidders had increased the per linear foot items on
their bids to bring about the difference.
It was concluded that sewer costing for the project, for San
Antonio, would be accomplished by multiplying the predictive
equation by 2.0 or 1.5 depending on the qualitative judgement
of the terrain as being largely in city streets or largely in
open country, respectively.
The final results for San Antonio sewer costs are accordingly
expressed as follows:
278
-------
if ID 12" to less than 30":
unit cost $/l.f. =
STREET* [ EXP(1.48019+0.0412173*10+0.0268722*Depth)]
* WPC-S(Yr, Region)
127.04
if ID 30" and up:
unit cost, $/l.f. =
STREET* [ EXP(2.28733+0.0227985*10+0.0134408*Depth)]
* WPC-S(Yr, Region)
127.04
STREET = 2.0 if largely through city streets,
1.5 if largely in open country.
EXP = e raised to the power indicated in parentheses.
ID = nominal inside diameter of pipe, inches.
DEPTH = average depth of cut below surface, feet.
WPC-S(Yr, Region) = EPA sewer cost indexes for the indicated
year and indicated region.
YR = year of prediction
REGION = the one of the 20-Cities regions surrounding the
ENR (Engineering News Record) 20 cities cost
indexes which has been set out by U.S. Public
Health Service for regionalizing cost indexes.
127.04 = the WPC-S national index for March 1968. Basis
of the original correlation.
It will readily be recognized that this correlation of sewer
costs, based on only 20 bids, must be purely provisional.
Through the Construction Grants activity EPA must have much
more data on individual sewer installations which could be
used for an analysis which would give much more secure pre-
dictive equations, of the same stature as those developed for
pipelines and mentioned in Chapter 4.
279
-------
COMPLETE SYSTEM COSTS
Table 33 shows the summary of the components for the conventional
and the reuse systems, the quantities as developed in Chapter 3
and the costs as developed in the immediately preceding sections.
The costs for the Cuero to San Antonio conveyance are a little
high since they have been computed for conveyance of the whole
149 mgd from Cuero. Actually 127.7 is conveyed from Cuero to
Cibolo, and 149 from Cibolo to San Antonio. The total cost
for the 149 mgd of surface water delivered is 25.7C/Kgal and
delivered and treated 31.6C/Kgal.
The cost of the new ground water supply is 3.4£/Kgal, and
3.8C/Kgal for the slightly smaller quantity involved in the
reuse scheme. For production from the existing ground water
facilities only the operating costs are shown, 1.4£/Kgal,
since the capital investment is already sunk and must be borne
with either scheme. On this basis, i.e. eliminating the amorti-
zation on the existing ground water facilities, the total make-
up water delivered and treated is 341 mgd for the conventional
case at about 15C/Kgal and 164 mgd for the reuse case at about
2C/Kgal. The differential between the conventional and the
reuse schemes amounts to 17.4 m$/yr.
For conventional sewage treatment in the existing or under con-
struction plants again only the operating cost is shown since
the capital cost must be borne by both schemes. The so computed
cost for conventional sewage treatment is 5.8 m$/yr.
Advanced waste treatment is applied to 177 mgd at the unit cost
of 23.8C/Kgal, total cost 15.3 m$/yr in a plant which requires
a capital investment of 86 m$. The return conveyance for this
water is accomplished for 3.3^/Kgal.
The very rough figures on demineralization by electrodialysis
show a cost of about ISC/Kgal. The relatively small fraction
of reject from electrodialysis, still amounting to seven mgd,
adds 1.1 m$/yr for conveyance to the Gulf for disposal.
The balancing storage which may be required in the reuse scheme
is not accounted for.
In total, the items accounted for involve an investment of about
264 m$ for the conventional scheme at an over-all 19.SC/Kgal for
a total annual cost of 24.4 m$. The corresponding investment
for the reuse scheme is considerably less, only 169 m$ but the
annual cost is greater, 26.8 m$/yr at 21.6£/Kgal•
280
-------
TABLE 33
COST SUMMARY
Share of reservoir cost Cuero
Cibolo
Surface water conveyance
Cuero and Cibolo to S.A.
Surface water Goliad to Victoria
Share of reservoir cost Goliad
Total surface water, delivered
Water treatment
Total surface water delivered & treated
GW withdrawal, new
Old, operating only
Total water delivered and treated
Sewage treatment, new
Old, operating only
AWTLCC
Return conveyance Rilling to Hildebrand
Balancing storage
Demineralization (very rough) 1st stage
2nd stage
Disposal to Gulf (very rough)
Total
Conventional
QDOT
mgd
230
230
141.4
230
id
107
333
234
80
24
12
-
id -
-
=i —
QBARE
mgd
127.7
21.3
149
113.1
113.1
149
149
149
47
145
341
118
40.7
12.2
6.1
-
-
-
-
341
New
Capital
m$
67.7
15.6
67.7
15.1
35.2
201.3
18.6
219.9
5.7
0
38.5
0
0
0
0
0
0
0
264.1
Annual
Cost
K$
3,150
744
7,268
1,157
1,670
13, 989
3,240
17,229
584
742
18,555
4,721
704
244
138
0
0
0
0
24, 362
tf/Kgal
6.8
9.5
13.4
2.8
4.0
25.7
6.1
31.6
3.4
1.4
(14.9)
11.0
4.7
5.5
6.2
AWT
QDOT QBARE
mgd mgd
-
_
-
-
102 39
333 125
164
-
350 177
350 177
Reuse
New
Capital
mf
0
0
0
0
5.5
0
0
86.1
15.7
Annual
Cost
K$
0
0
0
0
533
640
1,173
0
15, 352
2,138
0/Kgal
3.8
1.4
(2.0)
23.8
3.3
required, not considered yet
19.5
327 104 )_
_ r \
159 li /)
(14) (7)
341
(42.8)
(19)
169.1
(7, 073)
(1, 100)
26, 836
(18.6)
(44)
21.6
-------
Reuse, in these first approximating computations, proves more
expensive than conventional supply for San Antonio, but the
surprising thing is how close it comes to being competitive,
the difference in cost being about 10%. This is so close
that any real economic choice would have to await a more
refined estimate.
Various criticisms, of course, will be directed for or against
one or the other of the alternatives.
Most important the Cuero-Cibolo supply is probably not the
cheapest supplemental supply for San Antonio. Since this
comprises over half the total cost, if the alternatives do
indeed prove to be cheaper then, of course, the substitution
will lower the over-all cost of the conventional system.
Working in the other direction is the very likely possibility
that by the year 2000 a standard conventional treatment plant
will no longer produce an effluent that is allowable for
discharge. This will mean that the sewage treatment process
will have to be carried on a higher level of performance,
possibly involving actually some of the AWT processes in a
tertiary stage. Such extra performance would, of course,
involve additional costs.
Although the conventional sewage treatment plant designed and
costed in this study does have the capability to meet the
maximum day's demand under steady state conditions, it is
very likely that even so designed the plant cannot be operated
to continuously meet the standards under the fluctuating flow
conditions actually encountered. The AWT counterpart, however,
in the physical-chemical process used, is much more capable
of rapid operating changes to handle fluctuating demands...
which was one of the reasons for choosing it.
One of the reasons for the occasional (and indeed more than
occasional) poor performance of conventional biological treat-
ment plants is the high degree of flow fluctuation resulting
from storm flows in combined sewers. EPA has a whole series
of projects investigating the economics of separating combined
sewers so as to be able to segregate to storm flow portion and
in some cases to store it for bleeding off to the treatment
plant thus evening out the flow and improving the treatment
plant performance. The "uncombining" of combined sewage systems,
the provision of a separate storm sewer system, particularly the
provision of damping storage for storm flows requires a tre-
mendous investment. Work of many years ago in Boston, for
example, came to the conclusion that storm flows could not be
handled even with sewer pipes ten times the normal size. To
make conventional biological sewage treatment achieve 90%
removal 100% of the time would require a very large additional
expenditure for some such scheme, not taken into account here.
282
-------
It is quite likely that the cost of the AWT process will be
reduced by further technological development of this relatively
new process. Information on performance of activated carbon
treatment obtained since the design of the AWT process strongly
indicates that the filters may be dispensed with and this will
cut about l.SC/Kgal from the cost. If demineralization is to
be required anyway this will achieve removal of inorganic ions
including NH4 and in that case possibly the clinoptilolite
ion exchange stage of AWT could be dispensed with with a
resultant saving of some 7£/Kgal. However, this is not so
easily accomplished since the month-to-month fraction of the
AWT product demineralized fluctuates from 100% to 0%.
This project studied a reuse system in which the advanced waste
treatment was accomplished in a single plant, thus requiring
a return conveyance expense. This return conveyance expense
could be cut to practically zero and the sewage conveyance
cost reduced by having a number of decentralized AWT plants.
However, the numbers suggest that this probably would not be
economic because the return conveyance cost is only about 14%
of the AWT cost and it is likely that the increase in unit
cost in the smaller size plants would be considerably more
than this difference.
Finally, some will probably wish to look for the recovery of
the "values" from the seven mgd of 10,000 mgpl reject from
the demineralization.
Over the entire scheme, of course, looms the hard fact of the
present day that public acceptance of reuse would be less than
wholehearted. Indeed, although direct reuse is practiced in
one or two places in the world it is very unlikely that direct
reuse with processes at their present performance would be
accepted as safe by the public health authorities. This project
has compared the economics in the year 2000 on the assumption
that by that time the processes will have proven their safety
and the product will have achieved public acceptance.
283
-------
CHAPTER 7
ACKNOWLEDGEMENTS
The support of the project by the Water Quality Office
Environmental Protection Agency and the help provided by
Mac A. Weaver, Project Officer, and Roger Shull, Program
Coordinator, is acknowledged with thanks.
Project Director for Alamo Area Council of Governments was
C. Thomas Koch and assistance with the San Antonio data was
provided by Weldon Hammond.
Louis Koenig-Research of San Antonio was retained by the
Alamo Area Council of Governments to assist in the technical
matters of the project. Staff members thereof contributing
being: Paul Foerster, Louis Koenig, Jane Brymer, L. K. Cecil,
Larry Jureski, Justin Smith, Andrea Pesseto, Tazewell Dozier,
and Sharon Fletcher.
285
-------
CHAPTER 8
REFERENCES
1. Wessel, Henry E. New Graph Correlates Operating Data
for Chemical Processes. Chem. Eng. 59. No. 7, p. 209-
210. July 1952.
2. Smith, Robert. Cost of Conventional and Advanced Treatment
of Waste Waters. FWQA, Cincinnati, Ohio. July 1968.
3. Bishop, Fred. Personal communication to L. K. Cecil.
April 1970.
4. Seiden, L. and Patel, K. Mathematical Model of Tertiary
Treatment by Lime Addition. Robert A. Taft Water Research
Center. Report No. TWRC-14. Sept. 1969.
5. Mulbarger, M.D. and Grossman, Ernest III. Personal
communication. May 1970.
6. Cecil, L. K. Consultation and personal communication 1970
based on experienced performance of Accelators and related
recalcination equipment.
7. Crow, William B. and Wertz, Claude F., in Techniques and
Economics of Calcining Softening Sludges, Joint Discussion,
JAWWA. 52, 326-332. March 1960.
8. Mulbarger, M.C., Grossman III, E., and Dean, R. B. Lime
Clarification Recovery and Reuse for Waste Water Treatment.
FWQA. Cincinnati, Ohio. June 1968.
9. Infilco, Inc. The Viscomatic (R) Lime Slaker. Bulletin
255C. 1963.
10. Smith, Robert and McMichael, Walter F. Cost and Performance
Estimates for Tertiary Waste Water Treating Processes.
FWQA. Cincinnati, Ohio. June 1969.
11. Infilco Sales Bulletin, "Estimating Data, Phase I, Filters,
Accelators, Accelo-Biox," No. T-85-62 dated 9/13/62.
12. Battelle Memorial Institute. Pacific Northwest Laboratories
Ammonia Removal From Agriculture Runoff and Secondary
Effluents by Selective Ion Exchange. Robert A. Taft Water
Research Center Report No. TWRC-5, FWQA. Cincinnati, Ohio.
March 1969.
13. Mercer, B. W., Ames, L. L., Touhill, C. J., Vanslyke,
W.J., and Dean, R.B. Ammonia Removal From Secondary
Effluents by Selective Ion Exchange. Journal WPCF 42,
R95-R107. February 1970.
287
-------
14. Snow, Richard A. and Wnek, Walter J. Ammonia Stripping.
Mathematical Model for Waste Water Treatment. Report
No. IITRI-C6152-6 for FWQA. (n.d. but period ends 12/4/68) .
15. Cecil, L.K. Private communication and design suggestions
resulting from discussion with Basil Mercer, Battelle.
Based on Battelle, Tahoe and Blue Plains experience.
16. Louis Koenig-Research. Report in progress on ion exchange
program EPA Contract.
17. Mills, H.E. Costs of Process Equipment. Chem. Eng.
March 16, 1964. p. 138-139.
18. Guthrie, K.M. Data and Techniques for Preliminary Capital
Cost Estimating. Chem. Eng. March 24, 1969. p. 126.
19. Clarke, Loyal, arid Davidson, Robert L. Manual for Process
'Engineering Calculations. 2nd Ed. McGraw Hill. 1962.
20. Chilton, Cecil H. Cost Data Correlated, in; Chilton,
Editor: Cost Engineering in the Process Industries.
McGraw Hill. 1960.
21. Infilco Bulletin Sheet 106.
22. Allen, J. B., Clapham, T.M., Joyce, R.S., and Sukenik, V.A.
Use of Granular-Regenerable Carbon for Treatment of Secon-
dary Sewage. Engineering Design and Economic Evaluation.
Report to PHS. October 1. 1964.
23. M.W. Kellog Co. Appraisal of Granular Carbon Contacting,
Phase I. Evaluation of the Literature on the Use of
Granular Carbon for Tertiary. Waste Water Treatment,
Phase II. Economic Effect of Design Variables. Robert A.
Taft Research Center. Report No. TWRC-11. FWQA. May 1969.
(continued in Reference 24).
24. Swindell-Dressier Co. Appraisal of Granular Carbon
Contacting. Phase III Engineering Design and Cost Estimate
of Granular Carbon Tertiary Waste Water Treatment Plant.
Robert A. Taft Water Research Center. Report No'. TWRC-12.
FWQA. May 1969.
25. Cooper, J.C. and Hager, D.G. Water Reclamation with Granular
Activated Carbon. Chem. Enq. Progress' Symposium Series, 78.
Vol. 63. 1967. p. 185-192. AIChE. New York.
288
-------
26. Hanke, Steve H. , "Demand for Water Under Dynamic Conditions,"
Water Resources Research, 6, 1253-1261, 1970.
27. Hanke, Steve H., "The Demand for Water Under Dynamic
Conditions: A Case Study of Boulder, Colorado," 270 pp,
Center for Urban Enaineering Studies, Boulder, Colorado,
1969.
28. Linaweaver, F. P., Jr., Geyer, John C., and Wolff, Jerome
B., "Final and Summary Report on the Residential Water
Use Research Project," 87 pp, Dept. of Environmental
Engineering Science, Johns Hopkins University, Baltimore,
Maryland, July 1966.
29. Howe, Charles W., "Municipal Water Demands," In: Fore-
casting the Demands for Water, Policy and Planning Branch,
Dept. of Energy, Mines and Resources, Ottawa, 1968.
30. Wells, W., Personal Communication, 1970.
31. City Water Board, San Antonio, Texas, Water Statistics
Year Ending December 31, 1969, (Private Documents, City
Water Board, San Antonio).
32. Masse, Arthur, Advanced Waste Treatment Research Program,
USPHS, 2 pp, 1/29/63.
33. Durfor, Charles N., and Becker, Edith, "Public Water
Supplies of the 100 Largest Cities in the United States,
1962," USGS Water Supply Paper 1812, GPO, 1964.
34. Wells, W., Data in sewage treatment plant files collected
by operators and by TWQB.
35. Hammond, Weldon, Geologist, AACOG, Personal Coiairmnication,
1970.
36. Turner, Collie and Braden, Inc., Preliminary Engineering
Study of Alternative Conveyance Systems, Lake Austin to
San Antonio and Cuero-Cibolo to San Antonio for Texas
Water Development Board, March 1967.
37. Texas Water Development Board, The Texas Water Plan,
November 1968.
289
-------
38. Koenig, Louis, "Disposal of Saline Water Conversion Brines
- A Orientation Study," OSW R&D Progress Report No. 20,
1958.
39. Koenig, Louis, "Economic Boundaries of Saline Water
Conversion," JAWWA, 51, 845-62, 1959.
40. Koenig, Louis, "The Cost of Conventional Water Supply,"
Unpublished report under OSW Contract 14-01-0001-298, 1964.
41. Koenig, Louis, "Summary Report, Cost of Conventional
Water Supply," Unpublished report for OSW (summarizing
the preceding report), 1965.
42. Koenig, Louis, "Further Studies on Ultimate Disposal of
Advanced Treatment Waste," A report for the Advanced
Treatment Research Program, USPHS (unpublished), Aug. 1966.
43. Koenig, Louis, "The Cost of Conventional Water Supply,"
Chapter 11 in Spiegler, K.S., ed. Principles of Desalination,
Academic Press, New York, 1966 (now in revision).
44. Koenig, Louis and Jureski, Larry, "The Cost of Conveying
Water by Pipeline, Presented at ASCE Denver Water Resources
Engineering Conference, May 19, 1966.
45. Moody, L.F., "Friction Factors For Pipe Flow," Trans.
A.S.M.E. 66, 671- , Nov. 1944.
46. U.S. Bureau of Reclamation, "Friction Factors for Large
Conduits Flowing Full," Engineering Monograph No. 7,
USER, Denver, 1962.
47. Koenig, Louis, "The Cost of Pipelines in the United States,"
Submitted for publication.
48. U.S. Bureau of Reclamation, "Cost Estimating Procedure
(for Pump Station Operation and Maintenance Costs),"
Part 3 of Lockwood, Andrews, and Newnam, Inc., "Cost of
Transporting Water by Pipeline," Texas Water Development
Board Report 42, March 1967. TWDB, pp 127-138.. .and private
communication USER, 1970.
49. Federal Power Commission, Typical Electric Bills, an
annual publication, FPC, Washington, D.C.
50. National Electric Rate Book, Federal Power Commission,
Washington, D.C.
290
-------
51. Turner, Collie/ and Braden, Inc., Preliminary Engineering
Study of Alternative Conveyance Systems, Lake Austin to
San Antonio and Cuero-Cibolo to San Antonio for Texas
Water Development Board, March 1967.
52. Smith, Robert, "Preliminary Design and Simulation of
Conventional Waste Water Renovation Systems Using the
Digital Computer," FWPCA Water Pollution Control Research
Series WP-20-9, March 1968.
53. Koenig, Louis, "Operations Research and Logistics for
AWTRP," The Activated Sludge Process, Effect of Parameter
Improvement on Costs, and the National Benefit Resulting
Therefrom, Interim Report No. 5, FWPCA Contract No.
14-12-48, March 1970.
54. Sewage Treatment Plant Design Manual of Practice #8, WPCF
Manual of Practice No. 8, Water Pollution Control
Rederation, Washington, 1959.
55. Eckenfelder, W. Wesley, Jr., and Ford, Davis L., "Laboratory
and Design Procedures for Waste Water Treatment Processes,"
Technical Report EHE-10-6802 CRWR-31, Center for Research
and Water Resources, The University of Texas, 1968.
56. Texas Water Development Board. A Summary of Preliminary
Plan for Proposed Water Resources Development in the San
Antonio River Basin. TWDB, July 1966.
57. Texas Water Development Board. A Summary of Preliminary
Plan for Proposed Water Resources Development in the
Guadalupe River Basin. TWDB, July 1966.
58. U.S. Bureau of Reclamation. Plan of Development for Cibolo
Project, Texas. 1967. Summary Sheet.
59. Koenig, Louis, "The Cost of Water Treatment by Coagulation,
Sedimentation and Rapid Sand Filtration." JAWWA. 59.
pp 290-336. March 1967.
60. Mason-Rust. An Engineering Evaluation of the Electrodialysis
Process Adapted to Computer Methods for Water Desalination
Plants. Saline Water Conversion Progress Report No. 134.
OSW. Feb. 1965.
61. Michaels, Bob, EPA. Private communication, 1968.
291
-------
1
Accession Number
w
5
2
Subject Field & Croup
05D
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
Alamo Area Council of Governments, San Antonio, Texas
Title
BASIN MANAGEMENT FOR WATER REUSE
]Q Authorfc)
Koenig, Louis
1 6 I
- '
EPA WQO Grant No. 16110 EAX
21 Noto
Citation Final report submitted to EPA January 1972, 285 pp, 35 figures, 35 figures,
22
32 tables, 61 references
Descriptors (Started First)
23
Costs,* tertiary treatment,* water reuse,* activated sludge,*
water conveyance,* activated carbon,* water demand, return flow.
25
Identifiers (Starred First)
San Antonio, Texas,* lime treatment, clinoptilolite ammonis
removal.
27
Abstract
Computer programs were developed for the preliminary design and costing of
wastewater renovation by the lime-clinoptilolite-carbon processes of
advanced waste treatment; for activated sludge treatment; and for pipeline
conveyance of water. These together with methods of algorithms or lesser depth
for other processes were used to cost water supply and waste treatment under
conditions expected in San Antonio in the year 2000 for two extreme alternatives,
one importation of surface water according to the Texas Water Plan and
conventional water treatment, waste treatment and disposal by discharge; the
other completely closed recycle, discharging no waste water and reusing all the
waste water after treating it to make it reusable. The unit costs for these two
extremes were about 20c/kilogallon of water used and the reuse scheme was
only 10% more costly than the conventional scheme, i.e. well within
the expected error of the estimates. It was shown that the seasonal!ty of the
water consumption in the face of non-seasonality of the sewage produced has an
important bearing on the design and cost of reuse systems.
Abstractor
Louis Koenig
Innliti
"iion Louis Koenig, Research (A Corporation)San Antonio,Texas
WR:ID2 (REV. JUUY 1369)
WRSIC
SEND. «,TH COPY OF DOCUMENT. TO: W*T .« R»?URC? .SC^NT JP £j,*!5g«M*T '°N
WASHINGTON. D. C. 2024O
OU.S. GOVERNMENT PRINTING OFFICE: 1972 484-484/128 1-3
------- |