S~lhan[, Vol. 8, Part 2, March 1985, pp. 179-206. © Printed in India.

Study of Bhakra reservoir operation

P S RAO* and S R A M A S E S H A N t
Indian Institute of Management, Bannerghatta Road, Bangalore 560076, India
t Civil Engineering Department, Indian Institute of Technology, Kanpur 208 016, India.
Abstract. The study outlines the methodology for working out a multiobjective reservoir
operation with conjunctive utilization of surfaceand groundwater. Irrigation and power have
been considered the two objectivesfor operation of the [Link] results of this study are
compared with those achieved with conventional operation in the past and show that there is
an increase in irrigation by 7 to 8 ~ due to conjunctive use of surface and groundwater. The
trade-offbetween irrigation and power has been developedand discussed. The transformation
surface between three objectives viz., power, irrigation and carry-over storage has been
developed and the merits of carry-over storage have been discussed.
Keywards. Reservoiroperation; multiobjectiveanalysis;conjunctive use of water.

1. Introduction

Irrigation and power are two of the most important purposes for which surface and
groundwater resources are used. The integrated operation of a complex system is
necessary for efficient management of limited resources to meet irrigation and power
demands. The importance of the conjunctive utilization of surface and groundwater in
an integrated framework has been realised (Irrigation Commission 1972) but has not
been put into practice perhaps due to institutional constraints and lack of coordination.
However, using a systems analysis procedure it seems possible to study the integrated
operation of a multiobjective reservoir system with conjunctive utilization of surface
and groundwater. A case study of the Bhakra reservoir, which has been in operation
since 1966 is described in this paper. The object of this study is to analyse the past
operation o f the reservoir within a multiobjective framework to evaluate the trade-offs
between irrigation and power demands in the past. The study also demonstrates the
methodology and procedures in multiobjective analysis within an integrated
framework.

2. System description

2.1 The Bhakra

The Bhakra system in Punjab constitutes one of India's biggest projects (figure 1). The
Bhakra dam is a 225-55 m high concrete gravity dam built on the river Sutlej. The river,
the largest tributary o f the river Indus, begins in Tibet and enters the Indo-Gangetic
plains near Bhakra. The total catchment area upstream of Bhakra is 56,980 km 2. The
rainfall in the catchment varies over the basin with an annual average of around
875 mm. Govindsagar, the reservoir formed by the Bhakra dam, has a gross capacity of
179
180 P S Rao and S Ramaseshan

bladhopur Headworks. / / ~ M ~ o p u r 8eos Link ] //

, Y =+as o+= +n';+''"®` l l =..,.I

1'/ ~ y Pow~tr Plant ~,~ ++e~
rudas -, I , ~ I.~ --

/ / / L t Bhakra Dam &¢

I I// ~, Power PLa.t~ ~ Beo, so,~,j L,°k
Hoshio"p u,r I

~'ro~,
fm-~.o~,/l/1 /
/ . Julun6ur
~ ~ ~\
\ \ ~'~"
% Power
.....
House No ](6angawa,)
car,ag-'-/{ / / ~ ~N, ~'---NangaL Hyde! Channel

Figure 1. Interconnected system of the rivers Ravi, Beas and Sudej.

9876 m cu m and a live storage capacity of 7814 m cu m above a dead storage level of
445.62 m. It covers an area of 168.35 km 2 when full. The total run-off at Bhakra for a
dependable year works out to 13,723 m cu m and that for a mean year to 16,441 m cu m.
A dependable year has been adopted by the planners as a hypothetical year in which the
mean discharge o f any 10-day period is equalled or exceeded by the corresponding 10-
day mean discharges in 67 % of the years for which data are considered.
The river flows are high for the four months from June to September due to
snowmelt and monsoon rains. Almost 47 % of the river flow of a dependable year
occurs during these four months. Water from Govindsagar (Bhakra reservoir) can be
passed through the turbines of two power houses, one on the right bank and the other
on the left bank at the foot of the Bhakra dam. Both the power houses have 5 turbines
each. Each of the generators in the right-bank power house has a maximum capacity of
120 MWwhen the head is 116 m or more, whereas each of the left-bank generators has a
maximum capacity o f 90MW when the head is 110m or more. These capacities fall
respectively to 70 MW and 53 MW when the head falls to about 80 m. Thus the power-
generating capacity varies from a maximum of 1050MW tO a minimum of 615 MW.
About 11 km downstream o f the Bhakra dam is the Nangal reservoir formed by the
28-95 m high Nangal dam. It serves as a head regulator for control of irrigation releases.
Part o f the water from Nangal is released to the Nangal hydel channel which is 64.48 km
long and has a carrying capacity o f 353"75 cumecs. The remaining water is reJeased to
the Sutlej. The Nangal hydel channel supplies water to two power houses in its path at
Ganguwal and Kotla with a total installed capacity of 154 MW. Water from the Nangal
hydel channel is then divided between the Bhakra main canal and the Sirhind canal. The
Bhakra main canal serves some o f the canals of the old western Jamuna canal system
which were merged with the Bhakra system and so receives some supplies from the
western Jamuna canal.
Downstream of Nangal, there are headworks at two places on the river Sutlej at
Ropar and Harike. At Ropar, water is diverted to the Bist Doab and Sirhind canals. The
Beas river joins the Sutlej at Harike. Water is diverted at Harike to the Rajasthan feeder
 Study of Bhakra reservoir operation 181

and the Ferozepur feeder for the Eastern and Bikaner canals and the Sirhind feeder.
Water from all these canals is distributed through the network of the branch canals of
the irrigation distribution system. The details of the system are given in figure 1.
Punjab possesses large resources of groundwater which are generally suitable for
irrigation. Due to the presence of alluvium, a large-scale programme to install tubewells
has been undertaken. The State in 1974 had about 126,000 tubewells and another
75,000 were proposed under the fifth plan (1973-1978). An area of 0-6 to 0-8 m ha suffers
from waterlogging. The installation of tubewells would also solve the waterlogging
problem to some extent.

2.2 Irrigation demands

Bhakra was primarily designed to improve irrigation over an area of 1.42 mha and
irrigate an additional 2.43 million hectares. The area served consists of parts of Punjab,
Haryana and Rajasthan. The main crops grown in this region are bajra, cotton, maize,
rice, jowar, sugarcane, oilseeds, pulses and fodder during the monsoon (kharif) season;
and wheat, gram, oilseeds, barley, potatoes and fodder during the winter (rabi) season.
Irrigation requirements given in table 1 were assessed when the project was planned

Table 1. Irrigation requirements of the Bhakra system

Daily
No. irrigation Total irrigation
of requirements requirements
Month Period days (in cumec days) (in cumec days)

Filling period (from 1st of June to 20th of September)

June 1-30 30 715.71 21471
July 1-31 31 561"13 17395
August 1-31 31 561"13 17395
September 1-20 20 711"63 14233
70494 or 6100 m c u m
Depletion period (from 21st September to 31st of May)
September 21-30 10 711.63 7116
October 1-15 15 727.71 10916
16-31 16 654.63 10474
November 1-30 30 654.64 19639
December 1-10 10 611.76 6118
t1-15 5 363.67 1818
16-31 16 345.80 5533
January 1-15 15 345.80 5187
16-31 16 363-67 5819
February 1-10 t0 390-46 3905
11-28 18 545-52 9819
March 1-31 31 545.52 16911
April 1- 15 15 345.80 5187
16-30 15 418.89 6283
May 1-31 31 703.50 21809
136534 or 11800 m cu m
Total 6100+ 11800 = 17900 m c u m

(Mehndiratta & Hoon 1973)

182 P S Rao and S Ramaseshan

and were modified only slightly since, although the irrigation demands have changed
due to the introduction of high-yielding varieties of crops which require more water.
The need to reassess irrigation requirements and revise the demands has been well
recognised.
Irrigation requirements are high from September to November owing to the water
required for the maturing of the kharif crops and the preparation and sowing of the
rabi crops. The requirements are low from December to April, except in the latter half
of February and in March when water is required for the maturing of the rabi crops.
The rabi crop requirements are smaller than the kharif crop requirements. In May and
June, the requirements are again high due to the preparation and sowing of the kharif
crops. During the monsoon season also the requirements are quite high as the areas
where such water is to be utilized have low rainfall.

2.3 Power demands

The releases from the reservoir were originally planned mainly for irrigation. Only
987 mcum (revised upwards to 1360 m cu m recently) was earmarked for releases for
power generation. These releases were in addition to those made for irrigation. It was
assessed that the releases for power made would yield a firm power of 282 Mw from the
Bhakra power houses during an empirically-defined dry year.

2.4 Current operating policy

Each year the Bhakra Management Board (aMa) conducts operational planning
studies, referred to as water power studies, of the system as part of the planning to
regulate the supplies from the Bhakra reservoir. The water year from 1 June to 31 May
is divided into two periods. June 1 to 20 September is the 'filling period' when supplies
are generally available due to snowmelt and the monsoons and the reservoir fills up.
September 21 to 31 May is the 'depletion period' when water is released from reservoir
storage to meet the demands of irrigation and power.
During the filling period, irrigation demands are met in full subject to the constraint
that the reservoir is not allowed to deplete. This means that demands are not fully met if
they exceed inflows. This could happen because of the delayed arrival of monsoon,
during breaks in monsoon and also in relatively dry years.
Water power studies are conducted for the depletion period using the available
utilizable stored volume of water in the reservoir and dependable year or dry year
inflows for the depletion period. Although on the basis of 10-day inflows any particular
inflow figure of a river in a dependable year would have a probable availability of 67 ~ ,
the total yearly inflow of a river at any particular site, corresponding to the dependable
year has a probable availability of about 85 ~. A dry year is taken as the driest year on
record. It is not possible to meet the full irrigation requirements of the depletion period
every year. The ratio of deliveries to demands is called the reservoir factor (RF)defined
as

RF = (Total storage available in reservoir + expected inflow during de-

pletion period +__gains or losses upto the point of offtake of
canals.)/(Total canal requirements during the depletion period.)
The water power studies of BMBindicate that the firm power level is not a constant
throughout the depletion period. The firm power level is 15-30~o higher from 21
 Study of Bhakra reservoir operation 183

September to 10 December than from 10 December to 31 May. These studies include

the possibility of carryover storage from a good year to the subsequent year. Based on
water power studies a decision is taken on 21 September regarding regulation of
supplies from the reservoir during the depletion period. The decision implies choice of
(i) ~F (ii) firm power level, and (iii) minimum reservoir drawndown level to be reached
on 31 May (this determines the carryover storage, if any).

3. Review of earlier studies

The Bhakra system has been studied and reported by economists (Raj 1960; Ansari
1968; Minhas et a11972; Reidinger 1974) and by engineers (Singh 1964; Mehndiratta &
Hoon 1973). Raj concerned himself with some economic aspects of the Bhakra Nangal
project and a preliminary analysis of selected investment criteria. Ansari studied the
economics of irrigation rates and examined the basic arguments for incurring financial
deficits on irrigation supplies. Reidinger studied the institutional rationing of canal
water in an area in Hissar district which is served by one of the Bhakra canals.
Minhas et al (1972) made extensive simulation studies of the operations of the
Bhakra system (i) to determine for the depletion period the efficient combinations of
firm power with certainty and irrigation supplied with a given confidence level, (ii) to
evaluate the increase in availability of both irrigation and firm power due to conjunctive
utilization of surface and groundwater and to compare the economics of this scheme
with that of a thermal back-up. They observed that "in the absence of a satisfactory
economic measure of the relative worth of irrigation and power, one cannot determine
an ideal operating policy". But the important question of devising such a satisfactory
economic measure has been left unanswered in the study. As such the trade-off
possibilities delineated in a probabilistic sense cannot be used by decision-makers in
choosing the point of operation of the reservoir at the end of the filling period.
Singh (1964) put forward the idea of using seasonally available secondary power for
lifting groundwater by tubewells in the general framework of conjunctive utilization of
surface and ground waters. Minhas et al (1972) noted that the tubewells, when operated
solely to augment irrigation supplies, can utilize power which would otherwise go waste
and which therefore is 'free' in the economic sense of the term (since the economic value
of goods is based on their alternative uses). Operated solely to increase the level of firm
power output, tubewells can convert useless secondary power to valuable firm power.
Singh also pointed out the role of groundwater development as an antiwaterlogging
measure.
Mehndiratta & Hoon (1973) report the methodology adopted by B}aBto regulate
supplies from the Bhakra reservoir from 1966 to 1972. This report greatly helped in
understanding the system and serves as the source material for the present study. Water
power studies of BMB,given in tables 2 and 3, reveal that according to the planning
studies for the operation of the reservoir in 1969-70 and 1970-71, considerable
secondary or dump power was generated during the early part of the depletion period.
From the studies reviewed, it is clear that ti) the sudden peak in the irrigation
demands during the rabi sowing season, which is the early part of the depletion period
of the reservoir, necessitates the release of water with high potential for power
generation for irrigation; (ii) consumers of power who get used to the higher level of
firm power supplied during this period find it difficult to reduce their consumption of
 OO

ta.

Table 2. Water power study for the year 1969-70

Gains or Total
losses requirements
Inflows between Contributions of canals Available
in river Bhakra from Sirhind to be fed Actual Actual Reservoir average
Sutlej & Ropar feeder from Bhakra release RF elevations power
Month Period (cumecs) (cumecs) (cumecs) (cumecs) (cumecs) (~0) (m) (MgO

1 2 3 4 5 6 7 8 9 10

September 23-30 318"10 84"95 711"62 519"92 85 512"36 631

October 1-10 275-75 79-29 727-72 539-27 85 510-85 650
11-15 231-18 70"79 727-72 539"27 85 509"93 655
16-20 231-18 70-79 654"64 485"65 85 509"32 579
21-31 191.19 56.63 654.64 499.79 85 507.49 590
November 1-10 170.89 56.63 654.64 499.79 85 505.36 582
11-20 152.12 42.48 654.64 513.96 85 503.22 590
21-30 137-65 28.32 654.64 528"12 85 501.09 597
December 1-10 130.06 ~2-83 21.24 611.75 498,75 85 498"65 552
11-15 111.25 ~2"83 21.24 363'67 396.55 116 497.73 438
16-20 111.25 + 2'83 21'24 345'81 399'15 122 496"82 438
21-31 114-09 +2"83 21'24 345'81 403.66 124 494"69 438
January 1-10 112.90 +5-66 42.48 345"81 410.06 133 492"55 438
11-15 114.66 +5.66 42-88 345'81 415.04 134 491-64 438
16-20 114.66 + 5.66 42-88 363"67 418-52 129 490.42 438
21-31 106.98 +5.66 42.48 363-67 424.53 130 487"98 438
February 1-10 105.37 +11.33 16.99 390-46 433"16 118 485"24 438
11-20 105.90 + 11-33 16-99 545'52 442-14 86 482-49 438
21-28/29 109-21 + 11.33 14-16 545.52 451'00 87 480-36 438
March 1-10 118.76 + 5.66 14-16 545"52 461-03 88 477"31 438
11-20 116.19 +5.66 14-16 545"52 473"29 90 473"96 438
21-31 133.74 +5.66 14.16 545"52 488-13 93 470"30 438
April 1-10 150.56 -8-50 60.88 345"81 505"43 161 466-65 438
11-15 149.12 -8.50 60-88 345.81 509'85 163 464-52 438
16-20 149.12 -8.50 60-88 362"25 528-93 160 462'38 438
21-30 186.79 -8.50 60.88 362.25 545.27 165 458"11 438
May 1-10 243.67 - 16.99 109-02 703-50 568"01 94 454-15 438
11-20 274'56 -16.99 109-02 703"50 621"08 98 449-88 438
21-31 400.26 -16.99 109.02 703"50 620-42 101 446"53 438

(Mehndiratta & Hoon 1973)

OO
 Table 3. Water power study for the year 1970-71
oo
o~
Gains or Total
Inflows losses Contributions requirement
in river between from of canal to Available
Sutlej Bhakra Sirhind be fed from Actual Reservoir average
at Bhakra & Ropar feeder Bhakra releases Actual elevation, power
Month Period (cumecs) (cumecs) (cumecs) (cumees) (cumees) RF (%) (m) (MW)

1 2 3 4 5 6 7 8 9 10 t,.,

October 1970 21-31 191.19 .. 96-45 654.64 410-60 77.5 492.86 439"00
November 1-10 170-89 .. 65'07 654"64 42ff75 74'9 491.34 447.8
11-20 152.12 .. 58.96 654.64 420'75 74"0 489"20 441.5
21-30 137.65 .. 55.87 654.64 420-75 73'5 487.07 434"5
December 1-10 130-06 + 2.83 53.58 611'75 396-44 74'0 484.94 399.9
11- 15 111.25 + 2.83 55.64 363'67 311'96 101-8 484.02 311"9
16-20 111.25 + 2'83 47.68 345.81 313.88 105"3 483.41 311'9
21-31 114.09 + 2.83 47.68 345.81 317-17 106.3 481"58 311.9
January 1971 I-I0 112.90 +5.66 29.08 345.81 322.00 103"I 479.76 311"9
11-15 114.66 + 5'66 41.40 345.81 325.70 107-8 478.84 311'9
16-20 114.66 + 5.66 39.04 363.67 328.27 102'5 477.92 311'9
21-31 106-98 + 5.66 53-32 363.67 332'72 107.7 475.79 311'8
February 1-10 105.37 + 11"33 63.94 390-46 339' 17 106" 1 473.35 311"8
11-20 105.90 + 11.33 48.99 545.52 345.92 74-4 474.22 311.9
21-28/29 109.21 + 11.33 70.79 545.52 352"45 79.6 468.09 311"9
March 1-10 118.76 + 5'66 84.95 545.52 359"54 82.5 466.65 311-9
11-20 116.19 + 5.66 84.95 545'52 368"20 84" 1 463'91 311.9
21-31 133.74 + 5.66 84.95 545.52 378-25 85'9 460.55 31 I-9
April 1-10 150.56 -8, 50 53.83 345.81 389'I 1 125.6 458'I I 311'9
11 -l 5 149.12 ~ 8.50 63.17 345.81 307-67 131 '6 456.59 311.9
16-20 149.12 -8.50 63'17 418'89 404.18 110-2 455.07 311'8
21-30 186.79 - 8"50 72.21 418.89 414"44 114"1 452.02 311 "9
May 1-10 243.67 - 16.99 96.45 703.50 427' 18 72"0 449"58 311 "9
11-20 274'56 -16"99 87.75 703'50 438'34 72'3 447'14 311'8
21-31 400-26 -,I 6-99 108-82 703' 50 446"30 76"4 446'23 " 311 "9

(Mehndiratta & Hoon 1973).

 Study of Bhakra reservoir operation 187

power subsequently; (iii) it has not been possible for Br~Bto satisfactorily coordinate
the releases for irrigation and power; and (iv) conjunctive utilization of surface and
groundwaters may lead to increased firm power level and reliability of irrigation.

4. Conjunctive utilization

4.1 Conjunctive-use concept

The Bhakra surface water system has been described in §2.1. There has also been
intensive development of groundwater by the public and private sectors. Such
combined use of surface and groundwaters as is now practised was not preplanned but
has come into being out of necessity. Though the concept of conjunctive utilization of
surface and groundwaters has been put forward and widely advocated, it has not been
applied systematically. The Irrigation Commission (1972) recognized this fact and
recommended that schemes for conjunctive use of surface and groundwaters in existing
irrigation schemes should be accorded high priority and that a systematic study should
be made by each State to identify areas where conjunctive use is feasible. It appears that
conjunctive utilization is considered as the planned development of surface and ground
waters for irrigation. It has not been sufficiently realised, despite the good work of
Singh (1964)and Minhas et al (1972) that conjunctive utilization enables the operating
authorities to cope with the differences in time distribution of irrigation and power
demands in an optimal way and that it is a means of converting secondary power to
valuable firm power. The model described below and applied to the water power studies
ofaMB for 1969-70 and 1970-71 seeks to emphasize the conjunctive utilization concept
and demonstrate the benefits of its application.

4.2 Conjunctive-use model

A linear programming model which is an adaptation of a model proposed by Thomas &
Revelle (1966) for the Aswan High dam on the Nile river has been developed. Figure 2 is
a schematic representation of the model.
The objective is to minimize the discharge from the reservoir during the period 23
September to 30 November. The model is based on the fact that if secondary or dump
power generated could be used to lift groundwater and supply part of the irrigation
demand, then it is possible to reduce the discharges from the reservoir during this
period.
N

Objective function: Minimize ~ xi, (1)

i=l

where i = index for the subperiod i; subscript i denotes the value of the variable in
subperiod i, N = number of subperiods, and X = release from the reservoir in
cumec days.

4.3 Constraints
4.3a Power constraint: The reservoir release in each subperiod should generate the
firm power supplied in that subperiod.
X i >~Pi/Ki (for all i), (2)
188 P S Rao and S Ramaseshan

INFLOWS

@ BHAKRA
RESERVOIR

SECONDARY I I~RIG~.TION
POWER DEMAND

'r
I
l
I
i 1
IGROUNOWATER~
L ...... J
Figure 2. Conjunctiveuse model for the Bhakra reservoir.

where P = firm power demand, and K - - p o w e r conversion factor (energy rate

function) i.e., power generated by unit volume o f water in Mw per cumec.
4.3b Irrigation constraint: The irrigation demand will be met from two sources:
(i) releases from the reservoir and (ii) groundwater lifted by the excess, if any, of power
generated over firm power level. The amount of dump power generated should be
sufficient to lift the groundwater component of the irrigation demand. This is the
modification introduced in the model o f Thomas & Revelle (1966) to adapt it to the
present problem. Whereas surplus surface water is stored in the aquifers for later use in
their model, surplus power is used to lift existing groundwater in this model,
Xi >1Ai - mi(KiXi - Pi), (3)
where A = irrigation demand, and m = volume of groundwater lifted by unit power
(cumecs/Mw). When the irrigation demand is overwhelming, this constraint is binding
and hence holds as an equality.
4.3c Continuity constraint: The continuity constraint specifies that releases upto any
time cannot exceed the sum of storage and inflows upto that time. As the reservoir is
nearly full during the early depletion period considered in this problem this constraint
will not be binding.

Xi <~ V + ~ li for all s with s = 1,2 . . . . . N, (4)

i=1 i=1
whcre V = reservoir storage volume at the beginning of the depletion period, and
I -- inflow into the reservoir.

4.4 Data and assumptions

The model is used to study the regulation of the Bhakra reservoir during the depletion
period for 1969-70 and 1970-71. Data from the water power studies of BMB (tables 2
and 3) are used in the study.
 Study of Bhakra reservoir operation 189

Turbine efficiencies of the order of 0.85 and pump efficiencies of 0-7 are common. The
product of the turbine and pump efficiencies is assumed to be 0-6. Tail race elevation
(357 m) is taken from Singh (1965). The power conversion factors Ki are computed
from the data in tables 2 and 3. They are assumed constant during the subperiods. As
the length of the subperiod is short (8 to 10 days), the variation in the reservoir elevation
is small and the head is high, the effect of this approximation is negligible.
The depth to groundwater is assumed to be 15 m in October and 17 m in November.
Minhas et al (1972) mentioned that in Haryana and Punjab groundwater level rises to
15 to 25 m (50 to 80 ft) below ground surface even for deep tubewells. In many areas
groundwater is available much nearer the surface at depths of 3 to 4 m especially after
the rainy season in October and November. The assumption of 15 to 17 m depth to
groundwater is conservative.
The firm power level and reservoir factor adopted by aMa for the early depletion
period up to November 30 are respectively 552 MWand 0'85 in 1969-70 and 400 Mw and
0"75 in 1970-71. These values are used in this study.

4.5 Discussion of results

The model is applied to study the regulation of the Bhakra reservoir for 1969-70 and
1970-71 during the early depletion period when irrigation alone is constraining. The
results given in tables 4 and 5 show that the power used for lifting groundwater is much
smaller than the secondary power generated according to BMB studies and that the
actual releases of water from the reservoir are reduced by conjunctive use. For example,
it is seen from table 4 that during 11-15 October the secondary power generated is
103 Mw whereas the model suggests 9.9 Mw of power for lifting the groundwater
component of the irrigation demand resulting in a saving of 364.35 cumec days of water
in the reservoir. It is found that 2800 cumec days (i.e. 243 m cu m) of water could be
saved in the reservoir during the period 23 September to 30 November (table 4). This is
7.87 % of the total releases made during the same period (table 2). A similar value of
7.28 ~o is computed for 1970-7! from table 5.
The water so conserved in the reservoir can be used either to expand the irrigated area
and help enhance food production under the rabi crop or alternatively it can
subsequently help reduce the depletion rate of the reservoir when power demand is
constraining and it has a twofold effect. The conserved water increases the storage in the
reservoir and also increases the head available for power generation. At the end of the
depletion period, the reservoir could end up at a higher elevation. Although 7-8 ~ o f
saving of water appears small, it corresponds to 243 m cu m of water and is higher in
volume than the evaporation loss of 155.63 m c u m from Bhakra reservoir during
1969-70. Since irrigation is generally not constraining during the remainder of the
depletion period a 7 to 8 ~o increase in irrigation supply during this period can increase
the total output of irrigation by a corresponding percentage and this is significant.

5. Muitiobjective analysis

5.1 The concept

Multiple purposes should be treated as multiple objectives when there is conflict
between them as the nature of their consequences is noncommensurable and the
linkages between the purposes and the objectives they set out to serve are not known.
 Table 4. Early depletion period operation of Bhakra reservoir 1969-70.

Releases
Secondary suggested
power in by Model Available Power for Water saved in reservoir
BMB Study Power Groundwater
Month Period (MW) cumecs cumec days (MW) (MW) cumecs cumec days

September 23-30 79 464.25 3714 563.4 11"4 55-67 445.36

October 1-10 98 469.7 4697 565.9 13"9 69-57 695-70
11-15 103 466'4 2332 561.9 9"9 72.87 364.35
16-20 27 465.4 2327 556.0 4"0 30-40 101 .:55
21-31 38 472.5 5197 557.9 5'9 27-29 300"19
November 1-10 30 479.0 4791 557.0 5'0 20-79 207.90
11-20 38 485'6 4856 558.2 6.2 28"36 283"60
2 I-31 45 488.0 4880 552.0 0 40"12 401"20
Total: 32793cumoc days Total: 2799-55
eume¢ days.

Table 5. Early depletion period operation of Bhakra reservoir 1970-71

Releases
Secondary suggested Power Water saved
power in by model Available for in reservoir
BMB study power groundwater
Month Period (MW) cumecs cume¢ days (MW) (MW) cumecs cumec days

October 21-31 39.0 382 4199.3 406-0 6.0 28-60 314.6

November 1-10 47.8 384 3841.5 408.5 8.5 36.75 367-5
11-20 41.5 390 3898-6 408.4 8.4 30.75 307-5
21-30 34.5 395 3946-3 407.2 7.2 25.75 257.5
Total 15885.7 Total 1247.0
 Study of Bhakra reservoir operation 191

There has been sharp competition for and conflict in the use of the Bhakra reservoir
storage for irrigation and power (Mehndiratta & Hoon 1973; Reidinger 1974). In the
planning stage Bhakra was considered more of an irrigation project and the power
benefits were secondary. A firm power of 282 MW was expected from Bhakra power
houses in a dry year. With the rapid increase in power load in the region, the position
has completely changed. The power demand overrides other interests of the project
during most of the months and it has not been possible to satisfactorily coordinate the
releases. It is therefore worthwhile to consider irrigation and power as the two
objectives to operate the Bhakra system and to study the complementarity and conflict
between the two for the six years for which data are available.

5.2 The model

Thomas & Revelle (1966) used a linear programming (LP)model and for a given storage
in Aswan High dam derived the transformation curve between irrigation and power by
maximizing power for different levels of irrigation demands. A similar approach is used
to study the planned and actual depletion period operation of the Bhakra reservoir. The
LP model is as follows:
Objective function: maximize P, (5)
where P = firm power level in Mw.

5.3 Constraints
5.3a Irrigation constraint: The reservoir release in each subperiod should meet the
irrigation requirements to be supplied from the reservoir,
Xi >1Ai, with i -- 1, 2 . . . . . N, (6)
where N - - n u m b e r of subperiods during the depletion phase of the reservoir,
X = reservoir release in cumec days, and A = irrigation requirements to be supplied
from the reservoir in cumec days.
5.3b Power constraint: The reservoir release should generate the committed firm
power in each subperiod.

Xi/>--
P'flini with i-- 1,2 . . . . . N, (7)
gl
where fl~ = firm power coefficient in subperiod i, n~ = number of days in subperiod
i, and K = power conversion factor(Mw/cu mec).
5.3c Continuity constraint: The total volume of water released from the reservoir
during the depletion period should not exceed the total utilizable storage in the
reservoir at the beginning of the depletion period and the inflows in the depletion
period.
N N
x, v+ E 1. (8)
i=1 i=1

where V = volume of reservoir storage utilized during depletion period in cumec days,
and I = river inflow into the reservoir.
 Table 6a. Availability of water in filling period

Stored supply I,,2

Starting reservoir available at end
level at the Closing reservoir of filling period
Contribution beginning of" level at end of up to dead storage
River from Sirhind Contribution Releases filling period filling period reservoir level
inflows feeder from wJc* made in meters/ in meters/ 445.62 meters
Year (mcum) (mcum) (mcum) (mcum) Date of occurrence Date of occurrence (m cu m)

1966- 67 11114 913 382 4367 439.83 4-6-66 509.26 14-9-66 6463
1967-68 11397 1172 308 4305 443-80 3-6-67 512.23 21-9-67 6944 ta.
1968~9 9819 950 271 4317 454-74 31-5-68 506.76 12-9-68 6056
1969-70 12002 962 74 5230 451.03 27-5-69 513.30 21-9-69 7142
1970-71 7722 1011 247 3256 444.92 25-6-70 495-98 3-10-70 4527
1971-72 11077 999 358 3910 443-23 28-5-71 512.76 16-9-71 7043

(Mehndiratta & Hoon 1973) * West Jamuna canal

Table 6b. Availability of water in depletion period

Reservoir level Reservoir level

at beginning of at end of Stored supply
depletion period depletion period utilised during River Ravi-Beas Total supply
(in metres)/ (in metres)/ depletion period inflows contribution utilised
Year Date of occurrence Date of occurrence (m cu m) (m cu m) (mcum) (m cu m)

1966-67 509.26 14-9-66 443'30 3-6-67 6599 3848 604 11051

1967-68 512'23 21-9-67 454-74 31-5-68 6352 4576 1011 11939
1968-69 506-76 12-9-68 451-03 27-5-69 5711 4317 1493 11521
1969-70 513"30 21-9-69 444.92 25-6-70 7179 5033 1369 13581
1970-71 495.98 3-10-70 443'23 28-5-71 4663 3071 1604 9338
1971-72 512"76 16-9-71 445"54 16-6-72 7043 5205 1850 14098

(Mehndiratta & Hoon 1973)

 Study of Bhakra reservoir operation 193

5.4 Data and assumptions

5.4a Number ofsubperiods: The water and power studies of BMBgenerally divides the
depletion period from 23 September to 31 May into 29 periods (table 2). An LP package
on the time-sharing computer terminal at the Harvard University Centre for
Population Studies limits the total number of constraints to 50. To facilitate its use, the
depletion period is divided into 24 subperiods by aggregating some of the subperiods in
the waterpower studies.

5.4b Irrigation requirements: The irrigation requirements in cumecs are given in

table 1. During the depletion period, the irrigation requirements may be satisfied only
partly as defined by the reservoir factor (RF). They are computed for values of RF
ranging from 0.6 to 1 at increments of 0-05.
The expected inflow during depletion period has two components: (i) the inflow into
the reservoir and (ii) the inflows of the Ravi and Beas rivers that contribute through the
Sirhind feeder and supply part of the irrigation requirements of the Bhakra command.
The Sirhind feeder (Ravi-Beas) contributions vary from year to year and the data are
shown in table 6. While the break-up of these contributions period-wise is available for
1969-70 and 1970-71 from tables 2 and 3, an assumption is needed for the other four
years. The proportion o f the total flow in each subperiod for these years is assumed to
be the average of such proportions for 1969-70 and 1970-71. Gains or losses upto the
point of offtake of canals are available for each of the subperiods from tables 2 and 3.
5.4c Firm power level: The BMBchose for operation the results o f water study no. VI for
1969-70 (table 7), study no. 1 for 1970-71 (table 8) and study no. VIII for 1971-72

Table 7. Summaryof water power studies for 1969-70

Starling Minimum
reservoir reservoir
level in Pattern of level in Reservoir
Study No. meters inflow meters factor % Power

I 513.28 Dry year 445.62 75 22-9-69to 31-12-69550 MW

1-1-70 to 20-5-70450 MW
21-5-70 to 31-5-70 289 MW
II 513.28 Dry year 445.62 75 22-9-69to 31-12-69512 MW
1-1-70 to 31-5-70441 MW
III 513.28 Dry year 445.62 97 Minimumpower 282 MW
was ensured and releases
were proposedin the interest
of irrigation
IV 513.28 Dependableyear 445.62 78 22-9-69to 31-12-69500 MW
1-1-70 to 30-4-70450 MW
1-5-70 to 31-5-70 500 MW
V 513.28 Dependableyear 445.62 80 22-9-69to 10-12-69524 MW
I 1-12-69to 31-5-70450 MW
VI 513.28 Dependableyear 445-62 85 22-9-69to 10-t2-69552 MW
11-12-69 to 31-5-70438 MW
VII 513.28 Dependableyear 445.62 100 Minimumpowerof 282 MW
was ensured and releases
were proposedin the interest
of irrigation

(Mchndiratta & Hoon 1973)

194 P S Rao and S Ramaseshan

Table 8. Summaryof water power studies for 1970-71

Starting Minimum
reservoir Pattern reservoir
Study level in of level in Reservoir
No. metres inflows metres factor % Power

I 494,76 on Dependable 445.62 75 Up to 10th Dec.

I st October year 1970--400 MW 11th
December 1970 to
31st May 1971---
312MW
II 494.76 on Dry year 445.62 75 Up to 10th Dec.
31st October 1970--400 MW 1lth
December 1970 to
31st May 1971--
300 MW

(Mehndiratta & Hoon 1973)

(table 9). The fll values corresponding to the same were used in this study for the
respective years. The fli were found to be respectively 1"26 in 1969-70 and 1-282 in
1970-71. For 1966 to 1969 for which no data are available, a distribution similar to
1969-70 and 1970-71 with a fli value o f 1.27 is used.
5.4d Power conversion factors: The power conversion factors K i for 1969-70 and
1970-71 are deduced from the data in tables 2 and 3 respectively and the relationship
between K~ and reservoir elevation is derived. For the other four years, the
approximate reservoir elevations for each of the subperiods are estimated from the
anticipated and actual depletion curves in figures 3 (a to f) and the power conversion
factors corresponding to these elevations are used in this study. These depletion curves
generally correspond to RE values in the ranges of 0.75 to 0.85. For other RF values, the
depletion curves would be somewhat different and the reservoir elevations and hence
the power conversion factors would also be different. However, for simplicity, it is
assumed that the conversion factors are independent of the R~ value and the same
conversion factors as in the water power studies of BMa are used in each of the
subperiods. F r o m the solution thus obtained, it is possible to determine the reservoir
elevations implied and hence determine a new set o f values of K~ and repeat the
procedure till convergence is obtained. However, the irrigation requirement is
constraining generally only from 23 September to 10 December when the reservoir
elevation is very high and the depletion rate is low. Such refinement is not considered
here.
5.4e Sutlej river inflows: This study assumes the dependable year inflows for
planning as in the case of water power studies of aua. They are obtained from
Mehndiratta & H o o n (1973) and Lamba & Prem (1975) for all years except for 1971-72
for which there is a discrepancy between the two reported values. The dependable year
inflow for that year is corrected by interpolation. These are shown in table 10. For
actual operation, actual river inflows (table 6b) are used.
 Study of Bhakra reservoir operation 195

Table 9. Summary of water power studies 1971-72

Starting Minimum Reservoir

reservoir Pattern reservoir factor
Study level in of level in up to
No. metres inflows metres 10th Dec. 71 Power

I 512.06 Dependable year 445-92 8570 512MW up to 10th

December 1971 and
444 MW thereafter
up to 31-5-72
II 512.06 Dependable year 451-41 85~o 512 M W u p t o 10-12-
71 and 430 MW from
11-12-71 to 31-5-72
III 512.06 Dependable year 456-90 85 ~/o 512 MW up to 10-12-
71 and 414 MW from
11-12-71 to 31-5-72
IV 512.06 Dry year 445.92 81 o~ 484 MW up to 10-12-
71 and 432 MW from
11-12-71 to 31-5-72
¥ 512.06 Dry year 451.4t 81 ~o 484 MW up to 10-12-
71 and 4t9 MW from
11-12-71 to 31-5-72
VI 512.06 Dry year 457.20 81 So 484 MW up to 10-12-
71 arid 402 MW from
11-12-71 to 31-5-72
VII 512-05 Dependable year 44623 75)o 512 MW or more up
with month of May to 10-12-71 and
as per driest year 425 MW from t 1-12-
71 to 31-5-72
VIII 512.06 Dependable year 452.32 85 0~ 500 MW up to 31-12-
71 and 430 MW for
1/72, 3/72 and 5/72
and 420 MW for 2/72
and 4/72
IX 512.76 Dependable year 452.32 85~ 500 MW upto 31-12-
71 and 430 MW for
1/72, 3/72 and 5/72
and 420 MW for 2/72
and 4/72.
X 512.76 -do- 445.62 85 o,~ -do-
XI 512.76 Actual 445.62 85 ~o -do-

Note:
(1) Study Nos. I to VIII are BMa Water Power Studies for 1971-72 (Mehndiratta & Hoon 1973).
(2) Study No. XI is one of actual operation.

5.4f Stored volume: For planning studies, live storage available above dead storage
level (DSL)is used in the different studies. In case carryover was planned for, carryover
storage was estimated from volume elevation curves and deducted from live storage
above DSLto give utilizable storage with carryover. It is presumed that the losses due to
evaporation and infiltration from the reservoir are considered by BMB in the
specification of demands or in specifying inflows. The total volume of water for
196 P S Rao and S Ramaseshan

utilization during the years is assumed to bc the sum of dependable year inflows and the
utilizablc storage from the rcservoir as in thc water powcr studics of [Link] studics of
actual operation the volume of water actually utilized during the dcplction period is
used. The details of the data are given in table 10.

i i i I ~ i i i J
510.54 -.Max. R.L.509.~6 on 14.9,~6" 510.54 .L. 512.23 on 2.9.67

495.3o 495.30 [ ~

z
480.06 . . . . \~
.J
Anticipated ~ \
depletion curve \ \
~ e a l Oeplefion Actual deplebon
~ 464.82
uJ curve 464.82 curve \\~

\\\
449.58
449.58 \
M'In.R.L.4S4.74 on 31 56~
441.95 441.96 - _. I r
PERIOO PERIOD iocTINOvIOEC~JM~[FEBIMAF~Iz,Pr~I~AY

a. 1966 - 57 b. 1967- 68

i i i i f
510.5~ x. R t 506.76 on 12 9.68 510 54 on 21.9.69-

v3
LO
E 495.3o 4~30
w
z
J
~ 480.06
Anticipated ..~ . Antiopated \

~ 464.B2
(~tplehon curve ~\
46~82
\
cur ve
uJ

L
nr

449 58 449.58
vim. RL 451.03 on 27 5.69 \ Mtn. R L 444.92 on 2

441.96 - I 1 t~- 441.96

PERIOD ,~]~ OCT IN0VJDECIJANl FEBI~ARIAP~IMAY PERIOD 0CTINOVD
I ECIJANIFEB[MARIAPRIMAYI JUNI ,

C- 1968 - 69 d. 1969- 70
 Study of Bhakra reJervoir operation 197

5105a 9 71.

i l i l I
495.30 5 0 n 31070
~r 4 ~ 30

W
z
w 480.06
,,-i, ~8o 06
>
Z
W

g
~ 46~.82 Anbc~patcd
deplebon curve ~, 464 82 curve
Actual depletion ~\

449.58

441.96 " I I I t . . . . J-
441.96 I I ~ I I _
PER~D PERK)D 3C" i NC' 'IDFC , AN~ FEg!MaP'j a ,F~IMAV ]JL~,

¢ 1970-77 f 197]-72

Fig,~re 3. a, b; e, d; e, f; Anticipated and actual depletion curves. (Mehndiratta & Hoon

1973)

5.5 Derivation of irrigation and power transformation curves

The maximum level of firm power that can be developed for a given RF is determined by
solving the LP problem. The values of RF used are 0.6, 0-65, 0-7, 0.75, 0"8, 0-85, 0.9, (}95
and 1-0. The results o f the studies o f the planned and actual utilization of the reservoir
storage from 1966 to 1972 are given in table 11 and the corresponding transformation
curves are plotted in figure 4.
Carryover storage was an additional variable besides irrigation and power in
1971-72. The details and results of the studies with various carryover storages
considered in the planning studies in 1971-72 are given in tables 12 and 13. The
graphical representation of the concept of transformation surfaces and the resulting
transformation surface for Bhakra are plotted in figures 5 and 6. The transformation
curves, between irrigation (RF) and carryover storage, and between firm power and
carryover storage, are given in figures 7 and 8 and explained later in § 5.7b.

5.6 Analysis of transformation curves

The transformation curves for the planned and actual operation of the reservoir for six
years are generally similar in shape. The curves for 1966-67 are a little different in shape
at higher ~r levels and this is presumably due to very low Ravi-Beas contributions in the
depletion period of that year (table 6b).
The difference between the curves for planning and operation in each year is due to
the difference in the assumed depletion period inflows and actual inflows. The actual
inflow is much smaller than the assumed inflow in the dry year 1970-71; slightly smaller
 Table 10. Supplies planned for and actually utilized

oo
Stored Depletion Total supply Actual
supply period inflows planned for Stored depletion Total supply
above (dependable utilization supply period actually utilized
D.S.L. year) utilized inflows
Year (m cu m) (m cu m) (m cu m) (cumec days) (m cu m) (m cu m) (m cu m) (cume¢ days)

1966-67 6463 4145 10608 122777 6599 3848 10447 120914

1967-68 6944 3812 10756 124490 6352 4576 10928 126481
1968-69 6056 4231 10287 119062 5711 4317 10028 116064
1969-70 7142 3812 10954 126782 7179 5033 12212 141342
1970-71 4527 3454 7981 92372 4663 3071 7734 89513
1971-72 6463* 4059** 10522 121743 7043 5205 12248 141758

* Minimum reservoir level is 452.32 m; ** Corrected value.

Table 11. Maximum firm power for different reservoir factors.

1966--67 1967-68 1968-69 1969- 70 1970-71 1971-72

Reservoir Firm power Firm power Firm power Firm power Firm power Firm power
factor (MW) (MW) (MW) (MW) (MW) (MW)

Case 1 Case 2 Case 1 Case 2 Case 1 Case 2 Case 1 Case 2 Case 1 Case 2 Case 1 Case 2

0"60 430-8 424.25 445.3 465.0 413"6 401"2 461.9 514-9 320"9 310-1 447.7 521.3
0"65 430-8 424-25 445'3 465"0 413.6 401"2 461"9 514-9 320"9 310.1 447-7 521"3
0"70 430"8 424-0 445"3 465.0 413"6 401.2 461-9 , 514'9 318"8 305"8 447.7 521"3
0'75 428"1 420"7 445.3 465"0 413.6 401~2 461-9 514"9 308'2 288.4 ~147.7 521"3
0"80 420'9 411-8 442"5 464.4 412"5 397-8 461"0 514"9 279-4 250"5 446"3 521-3
0"85 404.7 390"3 432'5 459.6 404.0 386-0 454"4 514"9 227-6 N.F. 437"6 521"3
0.90 353"0 339"0 414"5 442.6 383"5 350-8 439'8 512.0 N.F. N.F. 422"5 520'5
0"95 263"1 N.F. 367"8 397'9 334"2 299"3 409.0 503"7 N.F. N.F. 402.4 513-7
1.00 N.F. N.F. N.F. 339"7 N.F. N.F. 354"1 488"0 N.F. N.F. 352-0 499"9

N.F.: Not feasible; Case 1: Study based on stored supply planned for utilisation and dependable year inflows during the depletion period;
Case 2: Study based on utilized stored supply and the actual inflows during the depletion period.
 Study of Bhakra reservoir operation 199

1.0

0.9

n,.

:\.\ \\ ';
~0.8 ,~-67--. t ~L\ ~1 I~ ,oo7-~ I ',
,.<

~ 0.7
n, , il Irl~ i,
, il II I! i i

0.6
Planned Operation
Actuat Operation

0.5 _1
200 300 400 500
FiRM POWER (MW)

Figure 4. Transformation curves for the Bhakra reservoir (1966-72)

IRRIGATION

i
o2 ~b

i," I

' C
r ;
,/ ls3 POWER

t31k / . . . .

D Oec~sio~ Polrff.
for Operation

Figure 5. Graphical representation of transformation surface.

 T ~ l e 12. Details of water power studies for 1970-71

Study Study Study Study Study Study Study

t
I II II1 VIII IX X XI

Reservoir level at the beginning of 512.06 512.06 512.06 512.06 512.76 512.76 512.76
depletion period (m)
Reservoir level at the end of 445-92 451.41 456.90 452.32 452.32 445.62 445'62
depletion period (m)
Stored supply utilized during 6,871.61 6,521'69 6,162.52 6,462.68 6,634 7,043 7,043
depletion period (m cu m)
Dependable year 4,059 4,059 4,059 4,059 4,059 4,059 5,205*
inflows (m cu m)
Total supply utilized (cumec days) 10,931 10,581 10,222 10,522 10,693 11,102 12,248
Total supply utilizable (mcum) 126,475 122,476 118,270 121,743 123,726 128,458 141,758"*
Carryover storage (mourn) 19.43 349.92 709.08 409.93 408-93 0 0
Carryover storage as percent of 0.3 4.94 10.32 5.95 5.95 0 0
total storage of 6871.61 m c u m

* Actual inflows; ** Total supplies utilized

 Study of Bhakra reservoir operation 201

Table 13. Maximum firm power for different retervoir factors and carryover storages
(water power studies for 1971-72)

Firm power (MW)

Reservoir
factor Study 1 Study 2 Study 3 Study 8 Study 9 Study 10 Study 11"

0-60 465-1 450-2 434-9 447-7 455.0 472.4 521 3

0-65 465.1 450-2 434.9 447.7 455"0 472.4 521.3
0-70 465-1 450.2 434.9 447.7 455.0 472.4 521.3
0-75 465.1 450-2 434.6 447.7 455-0 472.4 521.3
0.80 464.7 449.1 431.9 446-3 454.2 472.3 521.3
0-85 461.3 441.2 419.3 437.6 447.9 469.8 521.3
0.90 448-2 426.3 403-2 422.5 433.5 458.6 520-5
0-95 432.2 407.7 369.2 402.4 416.4 4432 513.7
1.00 400.9 359.0 314.1 352.0 372.5 420-5 500.0

Carryover
storage
(m cu m) 19.4 349.9 709.1 408.9 408.9 0 0

* Actual operation

in the below average years 1966-67 and 19684i9; slightly larger in the above average year
1967-68 and very much larger in the wet years 1969-70 and 1971-72. It shows that
using dependable year inflows for planning every year irrespective of how wet the year
has been and how high the reservoir has filled up led to very conservative planning in
wet years and the opposite of it in a dry year. It is not sound practice. It seems possible to
delineate an average year with a transformation curve between the above average and

RESERV(~R FACTOR

//

0.9

0.8 /

0.7 D~cis,on Point for

Operation
D1.D2 Parcto Admissible
Points
~_"~ 0"6~"00 400 500

Figure 6. Transformation surface for the Bhakra reservoir.

202 P S Rao and S Ramaseshan

I0

O.O

~- 0.8
u_

o
0.7

• Firm Power z,20 MW

0.8 A F*rm Power 437.6 MW

0.5
0 0 ,5 I.o
CARRYO',/ER STORAGE (b ¢u m )

Figure 7. Transformationcurve at constant firm power (1971-72).

• RF 0 . 8 5
v
o~ A RF 0 905

o 4s0

==
U_

Z,00
0.5 ,b
CARRYOVER STORAGE ( b ¢u m )

Figure 8. Transformationcurve at constant reservoirfactor (1971-72).

below average years for which the assumption o f dependable year inflows in depletion
period would be correct. Forecasting models for the stochastic inflows (Krishnasami
1976) are required to predict depletion period inflows for use in planning.
The points o f operation D chosen in the planning studies in 1969-70 and 1971-72
(figure 4) lie inside the respective transformation curves indicating that the planned
operation is inferior and nonoptimal. The point D could be moved on to the curve at Dt or
D 2 and the operation would be Pareto-superior to that at D. A higher firm power level
would be attained for the same Rr o f 0L85 at D1 while a higher Rr would be obtained for
the same firm power level at D 2. RF planned was never higher than 0.85 and so it seems
reasonable to suppose that the points DI would be the decision points o f BMB for
operation in 1969-70 and 1971-72.
The RF planned for was 0.75 in the dry year 1970-71. The decision point chosen in
planning is not feasible with respect to the actual transformation curve. The firm power
 Study of Bhakra reservoir operation 203

level planned for was stuck to and the adverse impact of the shortage must have been
borne by irrigation only. So the actual RF provided would have been lower than 0-75.
Irrigation supplies at RFof I were feasible in 3 out of 6 years and at RFof 0"9 tO 0"95 in
2 out of 6 years. In the dry year an RF of only 0-8 was feasible.
Firm power levels planned for varied from 312 MWin a dry year to 438 Mw in a wet
year. In actual operation it seemed that 510 to 520 raw of power was generated in the wet
years of 1969-70 and 1971-72 and 280 to 300Mw in the dry year of i970-71. The high
order of fluctuations in firm power from year to year can be reduced by carryover
storage from wet to dry years. For example, a carryover of about 1500mcu m storage
from 1969-70 to 1970-71 would have facilitated operation in both years around the
level of an average year (1968-69) with firm power of about 400 to 420 raw. The
recommendation of Minhas et al (1972) that a high dead storage level should be
maintained for high levels of firm power (which are possible in years of good inflows)
implies, in a way, the concept of carryover storage. Carryover storage was planned for
in 1971-72 possibly because of the experience of the previous two years, but in actual
operation this sound policy was not carried out.
Irrigation and power are almost wholly complementary up to an RFlevel of about 0.9
in years of good inflows but the complementarity could not be exploited because of the
uncertainty of depletion period inflows. This again points to using appropriate
streamflow forecasting models for planning reservoir operation.

5.7 "Top-down" and "bottom-up" planning

The transformation curves provide the frontier on which an optimal decision point for
operation should lie. If the social welfare function or the preference function of the
decision makers is available, the optimal point is obtained as the point of tangency
between the transformation curve and the welfare function or preference curve
respectively and the slope of the tangent gives the trade-offbetween the objectives. The
slope of the tangent to the transformation curve gives the marginal rate of technical
transformation between the objectives at that level of attainment of the objectives. The
slope of the tangent to the preference curve provides a measure of the relative marginal
economic worth of the two objectives (ratio of the shadow prices) at that level of
attainment of the objectives. At the decision point they are equal and the converse is
also true.
Well-defined preference functions do not exist. Shadow prices of the objectives vary
from time to time and depend on the levels of attainment of the objectives. They need to
be derived from an overall economic framework by a central planning organisation and
given to the project operators for use in decision-making. This is "top-down" planning.
If the ratio of shadow prices of the objectives are so defined exogenously (along the
transformation curve) the point on the transformation curve where the slope equals this
ratio defines the decision point. If it is not available from decision makers it is possible
to arrive at a decision point by interaction with the decision makers using techniques
like SWTmethod or Paretian analysis. In the absence of these, a "bottom-up" planning
procedure seems possible.
5.7a Analysis and discussion of implied trade-offs: The results presented in § 5.4 are
analysed and discussed in the framework of "bottom-up" methodology to elicit the
trade-offs implied in the choice of the decision points in the planned and actual
operation of the reservoir during six years.
204 P S Rao and S Ramaseshan

The trade-offs betw¢cn irrigation and firm power in the planned and actual operation
for the six years are given in table 14. The trade-offs implied in planned operation vary
from 200 MW/RI: in average years to 360 MW/Rv in a dry year. Since the results
indicate a higher trade-off for irrigation in a dry year, the value system adopted in
planning is consistent.
In 1969-70 (figure 4), the trade-offs of 200 MW/RF at Dt and 500 MW/Rr at D 2
provide the bounds for the marginal rates of transformation implied in the choice of
decision point D for operation. The BraBhas used a maximum R~ of 0"85 for operation
(tables 7, 8 and 9). Assuming, therefore, an Rv of 0.85 corresponding to D~ the trade-off
implied is 200 MW/Rv. The trade-off on the transformation curve of actual operation
at the level of D~ corresponding to RF of 0-85 is 0-0 MW/Rr. This implies that, in actual
operation, there is complete complementarity between irrigation and firm power up to
Rr of 0"85 and that the irrigation supply could have been increased upt0 Rr levels of 0.9
or even upto 1 with a small loss of firm power. Even at an RVof 1"0, 50 MW more of firm
power than that planned for would have been produced (table 11).
In the dry year 1970-71 (figure 4) the trade-offbetween irrigation and power implied
by the decision is 360 MW/xF. The planned operation at point D is not feasible for
actual inflows. If an Rv level of 0"75 was maintained, a firm power level of 288-4 MW
would have been reached on the transformation curve for actual operation and it would
have meant a trade-off of 490 MW/Rv at that point. This is larger than the trade-off in
an average year. During the early part of the depletion period from 3 October 1970 to
about 15 December 1970 the operation was as per the plan, i.e. at point D with RVof 0"75
and firm power level of 400 MW (figure 3e). The adverse effect of the shortage seems to
have been borne by the latter part of the depletion period from 15 December 1970 to 31
May 1971. The average power generation for the year 1970-71 was 335 MW. It can be
deduced from these that the Rr level achieved was certainly less than 0-75 and must have
been somewhere between 0.65 and 0.7 with a trade-off of 160 MW/RF or less. This, in
turn, implies a higher value for power than irrigation compared to an average year. This
indicates not only a contradiction between planning and operation but also an
inconsistency in economic evaluation. If, as pointed out earlier, the reservoir was not
depleted as it was in 1969-70 but some storage was carried over to 1970-71, it would
have greatly cushioned the adverse impact due to'poor inflows in 1970-71 and enabled
the achievement of a higher RF for irrigation.

Table 14. Trade-off between firm power and

irrigation

Trade-off (MW/XV)
Irrigation
level Planned Actual
Year (RF) operation operation

1966-67" 0-85 450 600

1867-68 0"85 260 210
1968-69 0-85 240 360
1969-70 0"85 200 0
1970-71 0"75 360 490
1971-72 0-85 230 0

* Ravi-Beas contribution is low as compared to other years.

 Study of Bhakra reservoir operation 205

5.7b Transformation surface for 1971-72: The concept of the transformation surface
amons the three objectives has been illustrated in figure 5. ABC constitutes the surface
bounded by the three coordinate planes. D is the decision point for operation. If it is on
the surface, it is Pareto-optimal. Curves a2a3, b263 and c2c3 pass through the decision
point and lie respectively in planes ala2a 3 parallel to the coordinate plane BOC, blb2b 3
parallel to OAC and c~c2c3 parallel to OAB. The trade-off implied by decision point D
between any two of the objectives can be derived as the negative of the slope of the curve
in the plane parallel to the coordinate plane defined by the same objectives.
For 1971-72, water power studies for planning the reservoir operation were
conducted by SMBassuming different levels of irrigation and firm power supplies and
different carryover storages. They are listed in table 9. Four of the studies, viz., I, II, III
and VIII assumed an irrigation level at RF 0-85 and dependable year inflows during
depletion period but they differ in that they assume different carryover storages. Details
of these studies are given in table 12.
The maximum firm power levels for different reservoir factors and carryover storages
are given in table 13. The transformation surface among the three objectives, irrigation,
firm power and carryover storage for the four planning studies is shown in figure 6. The
eMB decided to operate the reservoir according to study No. VIII with a carryover
storage of 408.93 mcum, a firm power level of 420 MW and RF of 0"85. A firm power
level of 437-6 MW was possible for an RFof 0"85and carryover storage of 408-93 m cum
(table 13, figure 4). Hence the decision point D (figure 4) is inside the transformation
curve and though feasible is not Pareto-optimal. For the same carryover storage, any
operation between D 1 (RFof 0"85 and firm power of 437.6 MW) and Dz (RFof 0-905 and
firm power of 420 MW) will be better than the point D chosen and will be Pareto-
superior. The trade-offs implied by choice of DI or D2 as decision points can be
obtained from the transformation curves in figures 4, 7 and 8 and are respectively 230 to
350 MW/RF between irrigation and power, 4824 to 5976 m cu m/Rr between carryover
storage and irrigation, and 60-6 to 67.6 MW/b cu m (billion cubic metres) between firm
power and carryover storage. As the SVlevel of 0.85 was always adhered to as an upper
limit, it appears that D~ would have been the choice. Hence the trade-offs implied by DI
are relevant indicators of decision makers' values.
The planning studies of BMS (table 9) assumed that the reservoir would reach an
elevation of 512-06m on 16 September 1971. But the reservoir actually filled to an
elevation of 512.76 m on that date with 171 mcum more storage than planned for in
these studies. Study Nos. IX and X (table 9) are revisions to account for this increase.
Study No. IX assumes carryover storage of 408.9 m cure as in study No. VI and
corresponds to planned operation. Study No. X assumes no carryover storage as in
Study No. XI and corresponds to actual operation. They are reported in tables 12 and
13. The transformation curve for actual operation (figure 4) indicates that irrigation
could have been supplied at an RFof 0-9 instead of 0.85 without any loss of firm power
and this too at a power level of 521.30 MWas against 420 Mw planned for. An increase in
ar to 1 would have resulted in a loss of only about 20 MW.
If the reservoir was operated according to the decision on 16 September, 1971 the
carryover storage would have been much more than what was planned for because the
excess of actual inflows over the assumed dependable year inflows, would have been
stored, and carried over to the subsequent year. The reservoir elevation at the end of the
depletion period would have been about 8 m higher than the planned elevation of
452-32 m. In actual operation the reservoir was drawn down to about the dead storage
206 P S Rao and S Ramaseshan

elevation o f 445-62 m. Thus 1746 m cu m o f water consisting o f excess storage and

excess inflows was used over and above the planned volume of 10,522 m cu m and all the
excess was used in the interest of power alone and the trade-off implied is 0.0 MW/RF
which considered irrigation above RF of 0"85 tO be completely worthless. Such a value
judgement needs justification. Thus, carryover storage, though planned for, was not
realized in practice. This was particularly striking after the dismal experience o f the
previous year.

6. Summary of results

The analysis of the results o f the study of the Bhakra reservoir operation from 1966 to
1972 indicates that (i) the depletion period inflows are correlated to the filling period
inflows and planning should be based on appropriate estimates o f depletion period
inflows; (ii) an RF o f 0-85 or above can be attained in an average year and an gF o f 1 can
be obtained at a loss of firm power of about 80 raw in an average year; (iii) planning and
operation should be based on consistent estimates of economic parameters; and
(iv) carryover storage will serve to reduce year to year fluctuations in firm power and
irrigation levels.

References

Ansari N 1968 Economics of irrigation rates--a study in Punjab and U ttar Pradesh (Bombay: Asia Publishing
House)
Irrigation Commission, Reports of the Irrigation Commission 1972 Vols. I, II, lII. Ministry of Irrigation and
Power, New Delhi
Krishnasami M 1976 Decoupled multivariate time series models for multisite streamflows Ph.D. thesis, IIT,
Kanpur
Lamba S S, Prem K S 1975 Integrated development of rivers Sutlej, Beas, and Ravi for optimum utility of
water in north-western areas of India, Presented at the IWRA Second World Congress on Water
Resources, New Delhi, Reprint Vol. 3, p. 79
Mehndiratta K R, Hoon R N 1973 J. Central Board oflrrigation and Power India 30:349
Minhas B S, Parikh K S, Srinivasan T N 1972 Sankhya, Indian J. Stat. B34:105
Raj K N 1960 Some economic aspects of the Bhakra Nangal Project-a preliminary analysis in terms of selected
investment criteria
Reidinger R B 1974 Econ. Dev. Cult. Chanqe 23:79
Singh H 1964 J. Central Board of lrrigation and Power, India 21:471
Singh J 1965 in Proceedings of the sixth regional conference on water resources development in Asia and thefar
East, Water Resources Series, United Nations, p. 281-295
Thomas H A Jr, Revelle R 1966 Manage. Sci. 12:13-296