Genetic algorithm based Different Re-dispatching Scheduling of Generator Units for Calculating Short Run Marginal Cost in Deregulated Environment

ACEEE Int. J. on Control System and Instrumentation, Vol. 03, No. 02, March 2012

Genetic algorithm based Different Re-dispatching
Scheduling of Generator Units for Calculating Short
Run Marginal Cost in Deregulated Environment
1
Priyanka Roy and 2A.Chakrabarti
1
Department of Electrical Engineering, Techno India, EM 4/1, Sector – V, Salt Lake, Kolkata-700091, W.B. India.
roy_priyan@rediffmail.com
2
Department of Electrical Engineering, Bengal Engineering and Science University, Shibpur, Botanic Garden,
Howrah - 711103, W.B. India
a_chakraborti55@yahoo.com

Abstract— Proper pricing of active power is an important issue regulated environment, these are not produce any significant
in deregulated power environment. This paper presents a pricing scheme in open access system. In the context of
flexible formulation for determining short run marginal cost deregulated electricity market, many literatures show marginal
of synchronous generators using genetic algorithm based cost of wheeling transactions and non-utility generation options
different re-dispatching scheduling considering economic load
using OPF [8], calculation of SRMC for power production taking
dispatch as well as optimized loss condition. By integrating
genetic algorithm based solution, problem formulation became
into account of real and reactive powers [9]. A mathematical
easier. The solution obtained from this methodology is quite model of electric power system has been developed to
encouraging and useful in the economic point of view and it understand marginal prices for consumers and generators in
has been observed that for calculating short run marginal competitive generation market and verified with a practical power
cost, generator re-dispatching solution is better than classical system [10]. Different methods of transmission pricing for
method solution. The proposed approach is efficient in the wheeling transactions [11] is also adopted.
real time application and allows for carrying out active power This paper introduces an alternative approach to determine
pricing independently. The paper includes test result of IEEE SRMC using genetic algorithm based different redispatching
30 bus standard test system.
solution followed by calculating SRMC for each generator. Using
Keywords— Short run marginal cost, genetic algorithm, genetic algorithm, three types of optimization problems have
economic load dispatch, loss optimization. been solved i.e. one simple ELD problem, loss optimization
problem as well as loss optimization maintaining ELD constraint.
I. INTRODUCTION In each and every optimization problem, one re-dispatching
scheduling of generator units has been found. It has been
In a restructured power industry, Electricity transmission observed that calculating SRMC in open environment, re-
and distribution are considered natural monopolies, whereas scheduling gives better result compare to conventional ELD
generation and retailing are open to competition. Open access problem. One of the advantages of genetic algorithm is that it is
to the generation system and fair, cost reflective pricing of a parallel process because it has multiple offspring thus making
generation are very imperative for healthy competition in the it ideal for large problems where evaluation of all possible
power sector. Short run marginal cost (SRMC) in deregulated solutions in serial would be too time taking. Application of the
environment is calculated from optimal power flow (OPF) proposed pricing formulation is demonstrated on the IEEE30
solution. Economic load dispatch (ELD) problem is the sub bus systems. The results have also been compared with the
problem of OPF. The main objective of ELD is to minimize the conventional method of solution of the ELD problem and shown
fuel cost while satisfying the load demand with transmission that proposed re-dispatching method is more significant than
constraints. The classical lambda iteration method is the base to classical one.
solve the ELD problem. This method is used equal increment In section 2, the proposed GA based ELD formulation as
cost criterion for systems without transmission losses and well as loss optimization techniques together with SRMC
penalty factors using B-coefficient matrix for considering the expressions are derived. Section 3 demonstrates the application
losses. of proposed formulation on IEEE 30 bus test systems.
In the past few years the interest in charging methodology Conclusion follows in section 4.
has become more pronounced. Many techniques have been
adopted and used to solve this problem. Pricing method such as A. Nomenclature
postage stamp, contract path, and megawatt mile are applied [1- In the analytical model following symbols have been used:
2]. The various rate structures have sought improvements in a
wide range of social objectives including the cost of electricity
Fctotal : Cost function of an N-bus power
generation, reliability of supply, utility profits and even income systems having NG number of fossil
distribution [3-7], but these methods are only applicable to fuel units

© 2012 ACEEE 1
DOI: 01.IJCSI.03.02.45


N : Number of buses Subject to

 ,  , : Cost coefficients g ( x, u )  0  (2)

Bij h( x, u )  0  (3)
: Loss coefficients for active power
Where
 : Power factor angles of bus load
 NG  NG
 : Phase angles of bus voltages Fctotal    Fci     i PGi   2
  i PGi   i 
 i 1  i 1
PD : Real power demands
The equality constraints (2) are the power flow equations, while
PG : Real power outputs
the inequality constraints (3) are due to various operational
PL : Real loss . limitations. The limitations include lower and upper limits of
generator power capacity.
Rij : Series resistance of lines 2) Loss optimization problem
The active loss is conventionally expressed using B-
 pi : Lagrangian multiplier for active power coefficient (or loss coefficient) matrix and can be represented as,
balance at the ith bus n m

g ( x, u )  0 : Equality constraint PL    PGi Bij PGj
i 1 j 1
h( x, u )  0 : Inequality constraint
n n m
 : Tolerance limit  B00   Bi 0 PGi   PGi Bij PGj
ρi : Short run marginal cost for ith bus (5)
i 1 i 1 j 1

For a system of N-plants, the loss coefficients are given by :
Suffix i stands for ith bus while suffix j stands for jth bus. The
variables have been expressed in p.u. while the angles have cos( i   j ) Rij
been expressed in degree. Bij 
cos  i cos  j | Vi || V j |
II. GA BASED RE-SCHEDULING WITH SRMC (6)
Genetic algorithm (GA) is a global adaptive search technique n m m
based on the mechanics of natural genetics. It is applied to B00   PDi Bij PDj And Bi 0   ( Bij  B ji ) PDj
optimize existing solutions by using methods based on biological i 1 j 1 j 1
evolution presented by Charles Darwin. It has many applications Constraints are considered to be
in certain types of problems that yield better results than the min
commonly used methods. To solve a specific problem with GA, PG i  PGi  PGmax
i
(7)
a function known as fitness function needs to be constructed
and
which allows different possible solutions to be evaluated. The
n
algorithm will then take those solutions and evaluate each one,
deleting the ones that show no promise towards a result but    PGi  PD  PL (8)
i 1
keeping those which seem to show some activity towards a
working solution. 3) Loss optimization with ELD constraint
This paper proposed a new generation re-scheduling
A. Problem formulation considering power flow where loss optimization problem is taken as a fitness function
requirements within GA for GA based solution where a new constraint is included, i.e.
In this paper, comparison study of short run marginal ELD constraint, in a single optimization problem. Fitness function
cost calculation is made and analyzed with three different re- for this problem is same as discussed in equation number (5)
sheduling techniques, namely economic load dispatch, loss and (6) where as a different constraint function is incorporated
optimization within power flow and loss optimization with other constraints mentioned in equation number (7) and
technique with maintaining economic load dispatch (8). Penalty factor has been selected as the constraint function,
constraint. The respective fitness function for three different and is given by equation number (9).
optimization problems is given hereunder.
1) Economic load dispatch problem  
The objective function being the total cost of power
dFc ( PGi )  1 
generation, we have  
NG dPG i 1  PL  (9)
 P 
GC ( x, u )   Fc ( PGi )   Gi 
i 1
© 2012 ACEEE 2
DOI: 01.IJCSI.03.02.45


According to the marginal cost theory, the marginal price ρi for of four types of condition denoted in types 1 to 4. All the
active power injection at bus i is generating powers are in p.u. basis and for each condition
Lagrange multiplier and B-coefficient are kept constant for power
PL balancing. The incremental loss cost and respective short run
ρi = -
  PGi for i = 1,2,……..,n (10) marginal cost of the generators for each operating condition are
given in table IV.
i.e. for SRMC for active power at a generator bus has two
TABLE I. T EST SYSTEM PROPERTIES
components, the first is the increment cost and the second is in
proportion to the system production cost rate multiplied by the
incremental loss caused by transmitting active power to this
bus.

TABLE II. PRODUCTION UNITS’ PROPERTIES

Table III shows different types of rescheduling, where first three
scheduling are computed using genetic algorithm whereas forth
one is classical method economic load dispatch. For each
scheduling, system operating loss is also computed. It has been
observed that for GA based scheduling, loss is less compared
to classical method. Moreover, in between three types of
scheduling, loss is minimum in third type of optimization, where
not only system loss is optimized, ELD condition is also fulfilled.
In table IV, Incremental loss cost as well as short run marginal
cost are calculated for each scheduling denoted in table IV. It
has been observed again that incremental loss cost is much
lesser in GA based scheduling compared to classical one.
Figure 1. Flowchart of the proposed optimization procedure Between three GA based optimization, proposed scheduling
where loss optimization and ELD both are maintained, computed
III. ANALYSIS AND COMPARISON WITH EXAMPLE
lesser incremental loss cost compared to other GA based
To examine the validity of GA model for proposed re- optimization. As soon as loss cost is decreased, short run
dispatching scheduling of the power generation of the marginal cost of the generating units as well as total plant’s
participating generators as well as SRMC calculation in the marginal cost is increased. This observation can be conveyed a
deregulated electricity market, IEEE 30 bus test system has been good economic message to the generation companies in
considered. The respective test system details are given in table deregulated environment to determine a new open access pricing
I and II. Table III shows different re-dispatching scheduling scheme for generating units.
T ABLE III. different re-dispatching scheduling of generator units

© 2012 ACEEE 3
DOI: 01.IJCSI.03.02.45


TABLE IV. I NCREMENTAL LOSS COST AND SRMC FOR GENERATING UNITS

[3] .Jaskow, “Contribution to the theory of marginal cost pricing,”
IV. CONCLUSION the bell journal of economics, vol9, no1, 1976.
[4] A.Kahn, The economics of regulation, principles and
This paper proposes a genetic algorithm based different institutions, newyork, john wileyand sons, inc, 1971.
power generation scheduling, i.e. economic load dispatch [5] G.Morgan, S.Talukdar, “electric power load management:
scheduling, loss optimized scheduling and a new scheduling some technical, economic, social issues, proc of IEEE, vol 67, no 2,
where loss optimization as well as ELD constrained are maintained feb, 1979.
for calculating optimized short run marginal cost for each [6] M.Munasinghe, “principles of modern electricity pricing”
generating unit with incremental loss cost. For each optimization proc of IEEE , vol 69, no 3, march, 1981
problem LaGrange multiplier and B-coefficient of the system are [7] Electric utility rate design study, EPRI, 1979.
[8] R Mukerji, W Neugebauer, R P Ludorf and A Catelli.
kept constant. These features can be used for further studies of
‘Evaluation of Wheeling and Non-utility Generation (NUG)
recovering generating cost by gencos in open access. The GA Options Using Optimal Power Flows’, IEEE Transactions on
based solution has been found suitable for calculation of SRMC Power Systems, vol 7, no 1, 1992, pp 201-207.
of synchronous generators in deregulated environment which [9] A A El-Keib and X Ma. ‘Calculating Short Run Marginal
is fundamentally different from those existing literature. Costs of Active and Reactive Power Production’.IEEE Transactions
on Power Systems, vol 12,no 2,1997,pp 559-565.
REFERENCES [10] I J Perez-Arriaga and C Meseguer. ‘Wholesale Marginal Prices
in Competitive Gener ation Markets’, IEEE Transactions on Power
[1] M.C.Caramanis, R.E.Bohn and F.C. Schweppe, “ optimal spot Systems, vol 12, no 2, 1997, pp 710-717.
price:practice and theory,” IEEE trans on power app and syst, vol [11] Y R Sood, N P Padhy and H O Gupta. ‘Wheeling of Power
101, no 9, sept, 1982. under Deregulated Environment of Power System - A Bibliographical
[2] R.D.Tabors, F.C.Schweppe and M.C.Caramanis, “utility Survey’, IEEE Transactions on Power Systems, vol 17, no 3, 2002,
experience with real time rates,” IEEE trans on power apps and pp 870-878.
syst, vol 101, no 9, sept 1982.

© 2012 ACEEE 4
DOI: 01.IJCSI.03.02.45

Genetic algorithm based Different Re-dispatching Scheduling of Generator Units for Calculating Short Run Marginal Cost in Deregulated Environment

Recommended

More Related Content

What's hot (18)

Similar to Genetic algorithm based Different Re-dispatching Scheduling of Generator Units for Calculating Short Run Marginal Cost in Deregulated Environment (20)

More from IDES Editor (20)

Recently uploaded (20)

Genetic algorithm based Different Re-dispatching Scheduling of Generator Units for Calculating Short Run Marginal Cost in Deregulated Environment