1

Profit based Unit Commitment Problem with

Power Flow Constraints
I. Jacob Raglend and Narayana Prasad Padhy1

Abstract In this paper an algorithm to solve the profit based

unit commitment problem (PBUCP) with power flow constraints
has been developed to plan an economic and secure generation
schedule. The proposed algorithm is developed from the view
point of a generation company wishing to maximize a profit as a
participant in the deregulated power and reserve markets.
Energy price plays a major role to make a decision in the
restructured system. Unit commitment (UC) in such a
competitive environment is not the same as the traditional one
anymore. This proposed algorithm introduces an efficient UC
approach that maximizes the profit of the generation company
satisfying both unit and network constraints. Repeated OPF for
the satisfactory unit combinations under the given study period
has been carried out to obtain UC solutions with both operational
and power flow constraints. The methods of payment for power
delivered and payment for reserve allocated are discussed. This
proposed algorithm has been tested on IEEE 30 bus system
scheduled for 24 hours. The solutions of the PBUCP have been
compared with the traditional unit commitment problem mainly
on its commitment schedule. The profits obtained by the above
mentioned methods are compared. The solutions obtained are
quite encouraging and useful in the deregulated environment.
The algorithm and simulation are carried through Matlab
environment.
Index Terms Dynamic Programming, Economic dispatch,
Newton Raphson, Optimal power flow, Profit based unit
commitment.

I. INTRODUCTION

N most of the interconnected power systems, the power

requirement is principally met by thermal power generation.
Several operating strategies are possible to meet the required
power demand, which varies from hour to hour over the day.
In power system operation and control, due to variation of
load demand and non-storable nature of electrical energy,
given the hourly load forecasting over a period of a day or a
week ahead, the system operators should schedule the on/off
status as well as the real power outputs of the generating units
to meet the forecasted demand over the time horizon [1-2].
The need to schedule the enough generation to meet the load

1
Narayana Prasad Padhy is an Associate Professor in the Department of
Electrical Engineering, Indian Institute of Technology, Roorkee, India247667.(e-mail: [email protected] ) and I. Jacob Raglend is a research
scholar in the Department of Electrical Engineering, Indian Institute of
Technology, Roorkee, India-247667(email: [email protected]).

in its daily cycle, led to the classical formulation of the

problem faced by a vertically integrated utility. In a simplest
formulation, unit commitment is the problem of determining
the schedule of generating units within a power system subject
to network and operating constraints [3-6].
Deregulation and restructuring of electric power systems have
created a competitive open market environment. A
restructured system allows the power supply to function
competitively, as well as allows the consumers to choose
suppliers of electric energy. Deregulation in general often
involves unbundling, in case of electric utility services
represents disaggregating into the basic parts of generation,
transmission and distribution. These operations are becoming
the responsibilities of generation companies (GENCOs),
transmission companies (TRANSCOs) and distribution
companies (DISCOs), with a central coordinator called
independent system operator (ISO), to balance supply and
demand in real time and maintain system reliability and
security [7-8]. In fact, the GENCOs, operating in an open
electricity market, are no longer bound to serve a local load,
but aim at maximizing their own profits. Unit commitment
choices are made on the expected behavior of market prices
over the time rather than by the forecasted load levels. In the
new competitive environment, each market participant finds
its own path to maximize their profit [9-12]. For the PBUCP
the GENCO runs unit commitment not for minimizing the
total production cost, but for maximizing there own profit. In
the past, utilities had an obligation to serve their customers.
All the demand and spinning reserve must be satisfied
completely. Here in restructured environment, based on the
forecasted prices and demand / reserve, GENCO can now
consider a schedule that may produces less than the predicted
load demand and reserves but will maximize the profit. The
objective of maximizing the profit is not same as minimizing
the cost. Profit depends on both revenue and the cost. If the
incremental revenue is larger than the incremental cost, more
energy may be generated for attaining more profit. Hence if
the incremental revenue is smaller than the incremental cost, it
may be less attractive to sell energy [13]. A number of
researchers have examined the concept of profit maximization
[14 - 22]. Researchers solved the price based UC problem by
an algorithm belonging to the Lagrangian relaxation approach
[23]. Researchers have formulated a new bidding strategy for
day-ahead energy and reserve markets based on evolutionary
programming [7]. Ritcher and Sheble have presented a tool for
assisting a GENCO to determine the amount of power to be
1

2
sold into the energy market to achieve maximum profit [11].
This paper presents a new profit based unit commitment
problem with power flow constraints in the competitive
environment considering both power and reserve generations.
The UC schedule for the generating units considering only the
unit constraints may not satisfy the power flow constraints and
leads to insecure operation of the network. To obtain the
practical UC solutions the model must consider both the
operational and power flow constraints. Repeated OPF for the
satisfactory unit combinations under the given study period
has been carried out to obtain UC solutions with both
operational, and power flow constraints. This paper presents
the profit based UC schedule for an IEEE 30 bus system with
operational and power flow (PF) constraints scheduled for 24
hours.
II. PROBLEM FORMULATION
A. Profit Based Unit Commitment Problem
In the competitive environment, GENCOs run their own
profit-based unit commitment based on forecasted demand,
reserve and price where the objective is to maximize
individual GENCOs profit [7]. The objective of the profit
based UCP is to maximize the profit.
The objective function becomes max PF = RV TC (1)
Where PF is the profit of GENCO
RV is the revenue of GENCO.
Power Balance Constraints
N

PGi X it PDt
i =1
N

R
i =1

it

, t = 1,K, T

X it PRt , t = 1,K, T

B. Strategies for Selling Power and Reserve

In the restructured power markets, unit commitment is used
by individual GENCOs for maximizing its profit in scheduling
generation resources. All market informations are reflected in
the market price. GENCO will sell power in a spot market and
sell reserve in a reserve market. The exact scheduling plan of
power and reserve depends on the way the reserve payments
are made. In the real market various approaches are adapted
for payment. In this paper two examples of reserve scenarios
are discussed [24].
C. Payment for Power Delivered (Method 1)
In this approach, reserve power is paid only when the reserve
is actually used. The reserve price is higher than the spot
price. For this method, the revenue and cost for the GENCO
can be calculated as follows
N

RV = {(PGit EPt + RPt Rit ) X it }

(6)

(1 ) F (PGit )

TC = + F (PGit + R it ) X it
i =1 t = 1

+ ST it (1 X i (t 1 ) )

(7)

i =1 t =1

F (PGit ) =

NG

i =1

+ i PGi + i PGi2

$ / hr

Where
(2)

F(PGit) is the fuel cost function of ith generating unit at hour t

i , i , i are the cost coefficients of generator

(3)

STit - Start-up cost at tth hour ($/h)

X it is the on / off status of ith generating unit at hour t

EPt is the forecasted spot price at hour t,
RPt is the forecasted reserve price at hour t

Where
Rit is the reserve generation of generator i at hour t

PDt is forecasted demand at hour t

PRt is forecasted reserve at hour t.

PGi - Real Power generated by the ith generator.
The power balance and reserve constraints are different
from the traditional unit commitment problem. The GENCO
can now select to sell power and reserve less than the
forecasted level if it creates more profit.
Two more constraints are added as follows:
(4)
0 Ri PGi max PGi min , i = 1, K, N

Ri + Pi PGi max , i = 1,K, N

the important parameters to the profit based UC problem as

well as strategies for selling power and reserve.

(5)

The unit capacity constraints and the minimum up / down time

constraints do not change. The problem formulation of
traditional unit commitment problem and their constraints are
given in Appendix 1. All the constraints given in Appendix 1
from (30) to (41) have to be satisfied while solving the profit
based unit commitment problem. Forecasted demand,
forecasted reserve, expected spot price and reserve price are

is the probability that the reserve is called and generated.

D. Payment for Reserve Allocated (Method 2)

In this approach, GENCO receives the reserve price per unit
of reserve power for every time period that the reserve is
allocated and not used. When the reserve is used, GENCO
receives the power price for the reserve that is generated. In
this approach of payment, the reserve price is much lower
than the power price. The revenue and cost for GENCO can
be calculated from
N

RV = {(PGit EPt + ((1 ) RPt + EPt )Rit ) X it } (8)

i =1 t =1

(1 ) F (PGit )

TC = + F (PGit + Rit ) X it
i =1 t =1

+ STit (1 X i (t 1) )

(9)

3
III. IMPLEMENTATION OF PROFIT BASED UNIT COMMITMENT
PROBLEM WITH POWER FLOW CONSTRAINTS
The proposed methodology can provide a fast solution, but the
quality of solution strongly depends on the algorithm to
update lagrangian multipliers. The step by step procedure for
the proposed method has been given below.

Initialize the unit characteristics for the N unit system

with system constraints.
Find all the available states that satisfy the load demand
for 24 hours. Each state corresponds to the ON and OFF
conditions of the generator units and represented as 1 and 0.
Calculate the transitional generation cost for the states
satisfying the system constraints on their transit from the
present stage to the succeeding stage with the help of
following steps
For each satisfying state carry out the optimal power flow
solution using a hybrid Lagrangian multiplier and Newton
Raphson power flow algorithms. Prepare the data base for the
system including line data, bus data, generator data and tap
setting of the transformers, forecasted demand, forecasted
reserve, forecasted spot energy price and forecasted reserve
price.
Form Ybus using line resistance, reactance, and shunt
elements [25].
Compute PGi and QGi for each load bus using (35, 36).

(k ) for
Compute the Scheduled errors PGi(k ) and QGi

each load from the following relation

(k ) = P Sch P (k )
PGi
Gi
Gi

(10)

Q (k ) = Q Sch Q(k )
Gi

Gi

(11)

Gi

where H = T C T Rbus C * * , H is a Hermitian matrix and it

symmetrical, i.e. H = H*.
H +H*
Real part of H can be determined using [H ] =
2

B 11

B 21
[H ] = M

B N G1

B 01
2

i(k +1) = i(k ) + i(k )

V (k +1) = V (k ) + V (k )
i

B1 N G

B 22

B2NG

B NG
B 02

L
K

M
B NG NG
B0NG

B 02

B0NG
2

B 00

B 01

(18)

N
T

L(P, R, , ) = TC RV t PDt PGit X it

i =1

t =1
(19)
N
T

t PRt Rit X it
i =1

t =1

Substituting the costs and revenue in (19) gives

(1 ) F(PGit) + F(PGit + Rit )

+ STit (1 Xi(t 1) ) PGit EPt Xit

N T

L = RPt Rit

i =1 t =1

(20)

N
N
T
T

t PDt Pit Xit t PRt Rit Xit

i =1
i =1
t =1

t =1

Equation (20) can be rewritten as

P J 1 J 2
(12)
Q = J 3 J 4 V

The new voltage magnitudes and phase angles are

computed using

After getting the loss coefficient perform economic

dispatch i.e., the power generated in each ON generator unit
using Lagrangian multiplier method.
For Method 1, the Lagrangian function becomes

Using (35, 36) compute the elements of the Jacobian

matrix obtained from the partial derivatives with respect to
i(k ) and Vi(k ) .

B 12

(13)

(1 ) F(PGit ) + F(PGit + Rit )

(
)
ST
X
P
EP
1
+

X it +

it
i (t 1)
Git
t
N T

L = RPt Rit

i =1 t =1
N

(( P

(14)

i =1 t =1

t Git

(21)

+ t Rit )X it ) (t PDt + t PRt )

t =1

The process is continued until the residuals PGi(k ) and

The last term in (21) is a constant and can be neglected.

Finally the Lagrangian function can be written as

tolerance .
Calculate the loss co-efficient using the power flow
solution. The total injected power at bus i is given by
S i = Pi + jQi = Vi I i*
(15)
The loss equation becomes
(16)
*
PL = [ PG1 ]T C T Rbus C * * PG
1
The resultant matrix is complex and the real power loss is
given by
*
(17)
PL = P T [H ] PG
1

(1 )F (PGit ) + F (PGit + Rit )

T

L = + STit (1 X i (t 1) ) PGit EPt X it (22)
i =1 t =1


RPt Rit + t PGit + t Rit
For an estimated value of and , PGi and RGi are

(k ) for all load buses are less than the specified

QGi

calculated
PGit
1 1
R =

it 1 1

A it
1 B it

(23)

G1

RPt t

B it =
2 i
Similarly for Method 2 (payment for reserve allocated)
the Lagrangian function becomes

(1 ) F(PGit ) + F(PGit + Rit )

+ STit (1 Xi(t 1) ) PGit EPt Xit

N T

L = ((1 ) RPt + EPt ) Rit

i =1 t =1

(24)

N
N
T
T

t PDt Pit Xit t PRt Rit Xit

i =1
i =1
t =1

t =1

Bit in (23) has been changed as follows

(1 ) RPt + EPt t

B it =
2 i

Check the slack bus power generated from economic

dispatch and the slack bus power obtained from the power
flow solution. If they lie with in a tolerance limit say 0.001,
then find the profit using (1). If they are not with in the
tolerance limit, then with the power generation obtained from
economic dispatch is given as the P specified in the load flow
analysis for the next iteration.
The same procedure is followed for all the states that
satisfy the constraints (2) and (3) for that hour.
Now tabulate all the profits obtained for all the satisfying
states for each stage and choose the maximum profit obtained
for each stage that satisfies the operational and power flow
constraints. Repeat the above steps for 24 hours with the
forecasted load profile.
Calculate the total profit gained by the GENCO by
summing all the maximum profits obtained between each
stage and print the results.

UC schedule considering only operational constraints and

including both operational and power flow constraints are
given in Table I. The profit based UC schedule considering
only operational constraints under both methods (payment for
power delivered and payment for reserve allocated) are given
in Table II. The profit based UC schedule considering
operational and power flow constraints under both methods
are given in Table III. For every hour, all the possible
combinations that satisfy the load demand and spinning
reserve constraints are selected and these states are allowed to
perform OPF. Unit commitment schedule without PFC may
not be practical, since the states must include the system
network losses and also converge for OPF. All the states that
satisfy OPF and their corresponding demand, power
generation, reserve, profit has to be stored. This procedure has
to continue for the specified time horizon. Select the state that
possess maximum profit and satisfies the unit constraints for
the entire time horizon. Finally the complete unit commitment
schedule with maximum profit for the GENCO has been
obtained. Fig. 2 shows the transitional generation cost for
every hour with and without PFC for the traditional UCP. For
both payment for power delivered and payment for reserve
allocated, the profit obtained by the UC schedule without PFC
versus time has been plotted in Fig. 3. The profit obtained by
the UC schedule with PFC versus time for both methods has
been plotted in Fig. 4. Fig. 5 gives the comparison of the
profit based UCP with and without PFC. The solution results
in this study indicate that the proposed algorithm is applicable
to the day-ahead UC calculation of large scale power systems.
The platform used for the implementation of this proposed
approach is on INTEL[R], Pentium [R] 4 CPU 1.8 GHz, 256
MB of RAM and simulated in the MATLAB environment.
300
IEEE 30 bus system
280
260
240
Load in MW

Where A it = EP t t i
2 i

220
200
180
160
140

The IEEE 30 bus system with six generating units shown in

Fig. 6 has been considered to illustrate the performance of UC
schedule with operational and power flow constraints. The
characteristics of generators, unit constraints, forecasted
demand, forecasted reserve, forecasted spot price, and
forecasted reserve price are given in Appendix 2. Probability
that reserve power is called and generated is set to 0.005 in
both methods. Loads at different buses are assumed to be
variable and have been generated using Gaussian random
noise function for the 24 hours during power flow simulations
because in practice the loads do not vary uniformly. The one
day forecasted load profile for the above case is given in Fig
1. In the proposed approach, the UCP schedule with
maximum profit for the GENCO has been obtained
considering both unit and network constraints. The traditional

120

10

15

20

25

Time in Hours

Fig. 1 Load Profile For IEEE 30 Bus

1000
Transitional cost without PF
Transitional cost with PF

900

800

T r a n s i ti o n a l C o s t in $ /h r

IV. SIMULATION AND RESULTS

700

600

500

400

300

200

10

15

20

25

Time in Hours

Fig 2. Traditional UCP with and without PFC

5
Table 1. Traditional UCP with Operational, with Operational and Power
Flow Constraints

3400
3200
3000

Period
2800

P ro fit in $ /h r

2600
2400
2200
2000
1800
Profit without PFC under Method1
Profit without PFC under Method2

1600
1400

10

15

20

25

Time in Hours
Fig 3. Profit obtained using Method 1and Method 2 without PFC
2800
Profit with PFC under Method1
Profit with PFC under Method2

2600

2400

P ro fit in $ /h r

2200

2000

Load
MW

UC
Schedule
with
Operational
Constraints

101101
1
166
101101
2
196
101101
3
229
111101
4
267
111101
5
283.4
111101
6
272
111101
7
246
111101
8
213
111101
9
192
110101
10
161
110101
11
147
110101
12
160
110101
13
170
110100
185
14
110100
208
15
110100
232
16
110100
246
17
111100
241
18
111100
236
19
111100
225
20
111100
204
21
111000
182
22
111000
161
23
111000
131
24
Minimum total generation
cost

Minimum
Generation
Cost
$/hour

492.3680
514.4198
623.0319
898.2148
766.8436
728.0110
642.2775
539.1919
476.9293
432.1891
363.5154
399.3866
427.6901
497.5949
538.8728
616.7335
663.7953
738.3682
609.1311
573.9276
508.6984
524.2561
378.1085
295.4216
13249
$/day

UC
Schedule
with
Operational
& Power
Flow
Constraints

101101
101101
111101
111001
111011
111011
111011
111011
111011
111011
111011
111011
111011
111011
111011
111011
111011
111011
111011
111011
111011
111011
111011
111010

Minimum
Generation
Cost
$/hour

506.0666
535.2034
799.0942
838.9670
987.8135
765.0130
670.6727
558.4164
491.5142
398.9967
359.6488
396.1338
425.1014
469.9788
542.1670
621.9727
670.6739
653.0992
635.7233
598.2173
529.3179
460.8624
398.9961
338.9695
13652
$/day

Table 2. Profit based UC Schedule with Operational Constraints for

Method 1and Method 2

1800

1600

1400

1200

10

15

20

Period

Demand
MW

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

166
196
229
267
283.4
272
246
213
192
161
147
160
170
185
208
232
246
241
236
225
204
182
161
131

25

Time in Hours

Fig. 4. Profit Obtained using Method 1 and Method 2 with PFC

3500

3000

P ro fit in $ /h r

2500

2000

1500

1000

Profit with PFC under Method1

Profit with PFC under Method2
Profit without PFC under Method 1
Profit without PFC under Method 2
0

10

15

20

Time in Hours

Fig. 5 Comparison of Profit based UCP with and without PFC

25

UC
Schedule
with
Operational
Constraints
Method 1
110010
111011
110100
101011
101011
110100
111011
100010
101101
110010
111011
100100
101111
110010
111001
100110
101011
110001
111111
110010
111001
100110
101111
110000

Maximum Profit

Profit
$/hour

2032.7
1822.5
1920.0
2296.7
2022.9
3022.7
2670.5
1906.1
2027.1
2047.7
1599.4
1834.9
1854.2
1875.4
1572.6
2217.1
2076.0
2451.1
2716.9
2063.2
1868.0
2255.2
1657.8
1575.9
49386
$/day

UC
Schedule
with
Operational
Constraints
Method 2
011111
111011
111011
111111
111011
111011
111111
011111
011111
011011
010011
011111
011111
111011
111011
011111
111011
111011
111111
011111
011111
011011
011011
011111

Profit
$/hour

2872.4
2614.1
2086.0
2728.7
2876.1
3194.0
3264.2
2865.0
2583.0
2704.2
2400.9
2864.6
2530.0
2551.4
1988.0
2773.0
2697.4
2949.1
3200.6
2944.0
2642.7
2841.1
2438.3
2565.4
65174
$/day

6
Table 3. Profit based UC Schedule with Operational and power Flow
Constraints for Method 1and Method 2

Period

Demand
MW

UC Schedule
with
Operational
and Power
Flow
Constraints
Method 1

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

166
196
229
267
283.4
272
246
213
192
161
147
160
170
185
208
232
246
241
236
225
204
182
161
131

110010
111011
110100
111011
111010
110101
101011
100011
111100
110011
101010
100101
111010
110011
101100
101011
110010
111101
110010
101011
100101
111010
110011
101101

UC Schedule
with
Operational
and power
Flow
Constraints
Method 2

Profit
$/hour

110010
111011
110100
111011
111010
110101
101011
100011
111100
110011
101010
100101
111010
110011
101100
101011
110010
111101
110010
101011
100101
111010
110011
101101

1903.1
1672.4
1841.8
2180.1
2052.6
2785.5
2475.0
1724.7
2069.0
1757.7
1333.9
1709.9
1899.0
1532.8
1789.5
1852.5
2080.1
2511.8
2401.6
1986.9
1905.6
2095.0
1418.7
1519.2
46498
$/day

Maximum Profit

V. CONCLUSION
This paper presented a profit based unit commitment
problem in a restructured power system. An approach to solve
profit based unit commitment problem by accommodating
operational and power flow constraints has been presented.
This algorithm would give realistic results as entire unit and
network constraints are included. The proposed algorithm
finds the most economical scheduling plan for GENCO by
considering both power and reserve generation. Spot prices
and reserve prices in the market are the important parameters
while solving the profit based unit commitment problem. Two
reserve payment methods are presented and simulated in an
IEEE 30 bus system. Since exhaustive enumeration technique
is used, it guarantees the optimality of the solution. By this
approaches, the GENCOs can maximize its profit. This
method can be extended to larger systems. Further the profit
based unit commitment problem can be solved incorporating
more constraints such as environmental emission constraints
and security constraints. From the results, the UC schedule
obtained maximizes the profit when solved by the approaches
presented in this paper. The results achieved are quite
encouraging and indicate the viability of the proposed
technique to deal with future UC problems.

Profit
$/hour

1911.3
1684.0
1847.2
2191.5
2061.0
2794.5
2483.5
1731.3
2076.5
1765.8
1338.4
1714.9
1910.8
1541.9
1794.9
1859.0
2087.2
2522.3
2409.3
1993.9
1911.8
2102.5
1426.6
1526.7
46686
$/day

VI. APPENDIX 1
Problem Formulation of Traditional UCP
Unit commitment is an optimization problem of
determining the schedule of generating units within a power
system with a number of constraints [1]. For a given power
system network, the optimization cost of generation is given
by the following equation.

Table 5. Comparison of Profit based UCP with Operational, with

Operational and Power Flow Constraints
IEEE 30 bus system
Method 1 with Operational Constraints
Method 1 with Operational and Power Flow Constraints
Method 2 with Operational Constraints
Method 2 with Operational and Power Flow Constraints
G

NG

G
5

G
8

13

F (PGit ) =
16

10

17

(25)

Where
TC is the total production cost for the UC schedules.
N G is the total number of generator units in the network.
FC is the total fuel cost of generators.
Total fuel cost of generation FC in terms of control
variables generator powers can be expressed as

G
12

TC = Min f i (FC ) + STit + SDit

i =1 t =1

Profit in $/day
49386
46498
65174
46686

11

28

NG

i =1

+ i PGi + i PGi2

$ / hr

(26)

where i , i , i are the cost coefficients of generator

PGi - Real Power generated by the ith generator

14

18

19

20

22

21

27

30

23

24

25

SDit - Start down cost at tth hour ($/h)

The start up cost is described by:
STit = TSit Fit + (1 e( Dti ASit ) BSit Fit + MSit

26

15

STit - Start-up cost at tth hour ($/h)

29

(27)

th

where TS it - Turbines start-up energy at i hour (MBTu)

Fit

Fuel input to the ith generator

Fig 6. One Line Diagram for IEEE 30 Bus System

Dit - Number of hours down at tth hour

AS it - Boiler cool-down coefficient at tth hour

min
max
QGi
, QGi
- Minimum and maximum value of reactive

power allowed at generator i.

PGi , QGi - Real and reactive power generation at bus i

th

BS it - Boiler start-up energy at t hour ($/h)

MS it - Start-up maintenance cost at tth hour ($/h)
Similarly the start down cost is described by
SDit = kPGi

PLi , QLi - Active and reactive power loss at bus i

Vi - Voltage magnitude at bus i, i - Voltage angle at bus i

(28)

where k is the proportional constant and the total production

cost is optimized with the following constraints.
Equality Constraints:
Power balance
NG

PGi = PDt + PRt + PLt

Yij - ij-th elements of Y-bus matrix

MVA f ij - Apparent power flow from bus i to bus j

MVA

connecting bus i and j.

(29)

i =1

Inequality Constraints:
Minimum up-time
0 < Tiu No. of hours units Gi has been on
Minimum down-time
0 < Tid No.of hoursunits Gi has been off

APPENDIX 2
Forecasted Demand, Reserve Spot Prices, Reserve Prices for
IEEE 30 Bus System

(30)
(31)

Period

Forecasted
Demand
MW

Forecasted
Reserve
MW

Forecasted
Spot Price
($/MWhr)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

166
196
229
267
283.4
272
246
213
192
161
147
160
170
185
208
232
246
241
236
225
204
182
161
131

15
18
20
25
28
27
25
21
19
16
14
16
17
18
20
23
24
24
23
22
20
18
16
13

10.55
10.35
9.00
9.45
10.00
11.25
11.30
10.65
10.35
11.20
10.75
10.60
10.55
10.35
9.00
9.45
10.00
11.25
11.30
10.65
10.35
11.20
10.75
10.60

Maximum and minimum output limits on generators

min
max
PGi
PGi PGi
Ramp rate limits for unit generation changes
PGit PGi (t 1) URi as generation increases

(33)

PGi (t 1) PGit DRi as generation decreases

(34)

where PDt - Demand at

(32)

th

t hour
PRt - Spinning reserve, PLt - Total system losses
Tiu - Minimum up-time, Tid - Minimum down-time

URi - Ramp-up rate limit of unit i (MW/h)

DRi - Ramp-down rate limit of unit i (MW/h)
Power Flow Equality Constraints:
Power balance equations
Nb

V j Yij cos ij i + j = 0

PGi PLi V i
j =1

Nb

QGi Q Li + V i
j =1

V j Yij sin ij i + j = 0

(35)
(36)

Power Flow Inequality Constraints:

min
PGi
min
QGi

Vi

max
PGi PGi
, i = 1,........, NG
max
QGi QGi
, i = 1,........, NG

min

Vi

Vi

MVA f ij MVA

(37)
(38)

max

, i = 1,........, N L

imin i imax
f ijmax

f ijmax - Maximum rating of transmission line

Max.
(MW)

Min.
(MW)

1
2
3
4
5
6

200
80
50
35
30
40

50
20
15
10
10
12

Ramp
Level
(MW/Hr)
50
20
13
9
8
10

Forecasted
Reserve
Price
($/MWhr)
Method 2
0.4220
0.4140
0.3600
0.3780
0.4000
0.4500
0.4520
0.4260
0.4140
0.4480
0.4300
0.4240
0.4220
0.4140
0.3600
0.3780
0.4000
0.4500
0.4520
0.4260
0.4140
0.4480
0.4300
0.4240

0
0
0
0
0
0

2.0
1.7
1.0
3.25
3.0
3.0

0.00375
0.01750
0.06250
0.00834
0.02500
0.02500

(39)
(40)

, i = 1,........., N TL

Unit
(No.)

Forecasted
Reserve
Price
($/MWhr)
Method 1
31.65
31.05
27.00
28.35
30.00
33.75
33.90
31.95
31.05
33.60
32.25
31.80
31.65
31.05
27.00
28.35
30.00
33.75
33.90
31.95
31.05
33.60
32.25
31.80

(41)

where N b - Number of total buses

N L - Number of load buses
N TL - Number of transmission lines

Unit
(No.)
1
2
3
4
5
6

Min
Up
Time
(Hr)
1
2
1
1
2
1

Min
Down
Time
(Hr)
1
2
1
2
1
1

Shut
down
Costs
($)
50
60
30
85
52
30

Cold
Start
(Hr)

Initial
unit
status

2
1
1
1
1
1

-1
-3
2
3
-2
2

Start up Costs
Hot
Cold
$
$
70
74
50
110
72
40

176
187
113
267
180
113

min
max
PGi
, PGi
- Minimum and maximum value of real power

allowed at generator i.
7

8
REFERENCES
[1]

N.P.Padhy, Unit Commitment- A Bibliographical Survey, IEEE Trans.

On Power Systems, vol. 19, no. 2, pp. 1196-1205, May 2004.
[2] Gerald B. Sheble and George N. Fahld, Unit Commitment Literature
Synopsis, IEEE Trans. on Power Systems, vol.9, pp.135-1205, Feb.
1994.
[3] G.P. Granelli, P. Marannino, M. Montagna, F. Zanellini, Monte Carlo
based Unit Commitment Procedures for the Deregulated Market
Environment, Int. Jour. Of Electric Power and Energy Systems, to be
published, 2006.
[4] A.J.Wood, B.F.Wollenberg, Power Generation, Operation and
Control, John Wiley & sons, Inc., 1996.
[5] J. Batut, A. Renaud, Daily Generation Scheduling Optimization with
Transmission Constraints: A New Class of Algorithms, IEEE Trans. On
Power Systems, vol. 7, no. 3, pp. 982 989, 1992.
[6] R. Baldick, The Generalized Unit Commitment Problem, IEEE Trans.
On Power Systems, vol. 7, no. 1, pp.465 475, 1995.
[7] Pathom Attaviriyanupap, Hiroyuki Kita, Eiichi Tanaka, Jun Hasegawa,
New Bidding Strategy Formulation for Day-ahead Energy and Reserve
Markets based on Evolutionary Programming, Int. Jour. Of Electric
Power and Energy Systems, vol. 27, pp. 157 167, 2005.
[8] L. .L. Lai, Power System Restructuring and Deregulation Trading
Performance and Information Technology, New York: Wiley, 2001.
[9] N.P.Padhy, Unit Commitment Problem under Deregulated Environment
A Review, Proc. IEEE Power Engineering Society General Meeting,
vol. 2, pp. 1088-1094, July 2003.
[10] G. B. Sheble, Computational Auction Mechanisms for Restructured
Power Industry Operation, Dordrecht: Kluwer, 1999.
[11] C. W. Ritcher Jr., G. B. Sheble A Profit Based Unit Commitment GA
for the Competitive Environment, IEEE Trans. on Power Systems,
vol.15, no. 2, pp.715-721, 2000.
[12] P. Attaviriyanupap, H. Kita, E. Tanaka, J. Hasegawa, A New Profit
Based Unit Commitment Considering Power and Reserve Generating,
Proc. IEEE Power Engineering Society Winter Meet, pp. 1311 - 1316,
February 2002.

[13] Mohammad Shahidehpour, Hatim Yamin, Zuyi Li, Market Operations

in Electric Power Systems Forecasting, Scheduling and Risk
Management, IEEE Wiley Inter Science, 2002.
[14] G. Gross, DJ Finlay, Optimal Bidding Strategies in Competitive
Electricity Markets, Proc. 12th Power System Computational
Conference, PSCC, pp. 815 822, August, 1996.
[15] G. Gross, DJ. Finlay, G. Deltas, Strategic Bidding in Electricity
Generation Supply Markets, Proc. IEEE Power Engineering Society
Winter Meet, pp. 309 315, February 1998.
[16] AK. David, Competitive Bidding in Electricity Supply, IEE Proc.
Generation, Transmission and Distribution, vol. 140, no. 5, pp. 421
426, 1993.
[17] D. Zhang, Y. Wang, PB. Luh Optimization based bidding strategies in
the deregulated markets, IEEE Trans. On Power Systems, vol. 15, no. 3,
pp. 981-986, 2000.
[18] F. Wen, AK. David, Optimal Bidding Strategies and Modeling of
Imperfect Information among Competitive Generators, IEEE Trans. On
Power Systems, vol. 16, no. 1, pp. 15 - 21, 2001.
[19] F. Wen, AK. David, Coordination of Bidding Strategies in Day ahead
Energy and Spinning Reserve Markets, Int. Jour. Of Electric Power and
Energy Systems, vol. 24, no. 5, pp. 251 261, 2002.
[20] AJ. Conejo, FJ. Nogales, JM. Arroyo, Price Taker Bidding Strategy
under Price Uncertainty, IEEE Trans. On Power Systems, vol. 17, no. 4,
pp. 1081 - 1088, 2002.
[21] K. Bhattacharya, MHJ. Bollen, JE. Daalder, Operation of Restructured
Power Systems: Kluwer Academic Publisher, 2001.
[22] X. Guan, YC. Ho, F. Lai, An Ordinal Optimization based Bidding
Strategy for Electric Power Suppliers in the Daily Energy Market, IEEE
Trans. On Power Systems, vol. 16, no. 4, pp. 788 - 797, 2001.
[23] A. Borghetti, A. Frangioni, F. Lacalandra, CA. Nucci, Lagrangian
Heuristics Based on Disaggregated Bundle Methods for Hydrothermal
Unit Commitment, IEEE Trans. On Power Systems, vol. 18, no. 1, pp.
313 - 323, 2003.
[24] E. H. Allen, M. D. Ilic, Reserve Markets for Power Systems
Reliability, IEEE Trans. On Power Systems, vol. 15, no. 1, pp. 228 233, 2000.
[25] Hadi Sadat, Power System Analysis, Tata McGraw-Hill Publishing
Company Limited, 2002.