P.

Subburaj
Research Scholar, National Engineering College, K.R.Nagar, Kovilpatti, Tamil Nadu, India. subbunec@yahoo.com

J. Electrical Systems 2-4 (2006): 198-207 Regular paper Distribution System Reconfiguration for Loss Reduction using Genetic Algorithm

JES
Journal of Electrical Systems

K. Ramar
Professor and Head National Engineering College, K. R. Nagar, Kovilpatti, Tamil Nadu, India. kramar_nec@rediffmail.com

L. Ganesan
Professor and Head, A.C.College of Engineering and Technology, Karaikudi, Tamil Nadu, India. drlgtny@yahoo.com

P. Venkatesh
Associate Professor, Thiagarajar College of Engineering, Madurai, Tamil Nadu, India venkatesh_paramasivam@yahoo.com

Feeder reconfiguration is defined as altering the topological structures of distribution feeders by changing the open/closed states of the sectionalizing and tie switches. In network reconfiguration for loss reduction, the solution involves a search over relevant radial configuration. In this paper, a method, based on genetic algorithm (GA) to determine the minimum configuration is presented. A genetic algorithm (GA) is a search or optimization algorithm based on the mechanics of natural selection and natural genetics. Since GA is suitable to solve combinatorial optimization problems, it can be successfully applied to problems of loss minimization in distribution systems. Test results are included to show the performance of the proposed method.

Keywords: Network reconfiguration, Genetic algorithm, Loss minimization, Radial distribution system, Power flow.

1. INTRODUCTION Electrical distribution networks are built as interconnected and meshed networks. However, they are arranged to be radial in operation. Their configurations may be varied with manual or automatic switching operations so that, all the loads are supplied and reduce power loss. Reconfiguration also relieves the overloading of the network components. Feeder reconfiguration is performed by opening sectionalizing (normally closed) and closing tie (normally open) switches of the network. These switching are performed in such a way that the radiality of the network is maintained and all the loads are energized. A normally open tie switch is closed to transfer a load from one feeder to another while an appropriate sectionalizing switch is opened to restore the radial structure. The problem to be addressed is, to determine the status of the network switches such that the reduction in power loss is achieved. A number of papers have appeared on the general topic of feeder reconfiguration as it applies to normal operating conditions. An early work on loss reduction through network reconfiguration was presented by Civanlar et al. [1] which described a formula to estimate the loss change resulting from the transfer of a group of loads from one feeder to another feeder.

Copyright JES 2006 on-line : journal.esrgroups.org/jes

J. Electrical Systems 2-4 (2006): 198-207

This is done through the closing of a single tie switch and the opening of a single sectionalizing switch. Baran et al. [2] described a reconfiguration methodology for loss reduction and load balancing based upon considering branch exchange type switching. Shirmohammadi et al. [3] described a technique for the reconfiguration of distribution networks to decrease their resistive line losses and included results pertaining to large scale system examples. Lubkeman et al. [4] presented an expert system using heuristic rules to shrink the search space for reducing the computation time. However, only a feasible solution can be obtained for knowledge based methods. Chiang et al. [5], [6] proposed new solution methodologies using the simulated annealing algorithm for the network reconfiguration. Goswami et al. [7] presented a heuristic algorithm for the reconfiguration of feeders. Kochi Nara et al. [8] proposed network reconfiguration techniques for minimum loss configuration using genetic algorithm (GA).Kim et al. [9] proposed a neural networkbased method with mapping capability to identify various network configurations corresponding to different load levels. Borozan et al. [10] proposed an algorithm for calculating Zloop matrix using the ordered network elements. Taleski et al. [11] proposed a method to determine the network reconfiguration with minimum energy losses for a given period. Jeon et al. [12] presented the simulated annealing algorithm with Tabu search for loss reduction. The Tabu search attempted to determine a better solution in the manner of a greatest - descent algorithm, but it could not give any guarantee of the convergence property. Chin et al. [13] presented a ranking index method to determine the distribution network reconfiguration problem for loss reduction. Morton et al. [14] presented a bruteforce solution for determining a minimal-loss radial configuration. The graph theory involving semi sparse transformations of a current sensitivity matrix was used, which guaranteed a globally optimal solution but needed an exhaustive search. Lin et al. [15] presented a refined genetic algorithm (RGA) to reduce losses. In RGA, the conventional crossover and mutation schemes were refined by a competition mechanism. Veerareddy et al. [17] presented a two stage approach for determining the network reconfiguration, which involves determining the loop for maximum loss reduction and distance center technique. Prasad et al. [18] presented a fuzzy mutated genetic algorithm for optimal reconfiguration of radial distribution systems. This method involves a new chromosome representation of the network and a fuzzy mutation control for an effective search of solution space. Hong et al. [16], [19] presented a method based on genetic algorithms (GA) and fuzzy multi objective programming for determining the network reconfiguration in distribution systems. A vertex encoding based on Prufer number was adopted in GA for encoding the chromosomes. Sivanagaraju et al. [20] presented a method to determine the voltage stability of radial distribution systems by network reconfiguration. This paper discusses the problem of reducing power losses in distribution feeders via feeder reconfiguration. The genetic algorithm is successively applied to the loss minimum reconfiguration problem. In the proposed algorithm, strings consist of sectionalizing switches, tie switches and fitness function consists of total system losses are formulated. Numerical example is provided to show the validity and effectiveness of the proposed algorithm. . 2. RADIAL DISTRIBUTION SYSTEM In the figure, load points where the transformers are tapped off from the primary circuit is marked by dots .. Also shown in the figure, there are two types of switches in the system. Normally closed switches connecting the line switches (CB1-CB6) and normally open switches on the tie lines connecting either two primary feeders (CB7) or two substations (CB8) or loop type laterals (CB9).

199

P. Subburaj et al: Distribution System Reconfiguration for Loss Reduction using Genetic Algorithm

Distribution systems are normally operated as radial networks. However, changing the state of some sectionalizing switches changes configuration.
SS1 CB1 CB2 CB3 CB5

CB7 SS2 CB9 CB8 CB6

CB4

Fig. 1. Schematic diagram of a primary circuit of a distribution system

The base network can be reconfigured by first closing a tie line switch (CB7). Since this switching will create a loop in the system, a branch in the loop containing a switch has to be opened (CB3) to restore the radial structure of the system. As a result of this, the loads between the branches 8 and 11 will be transferred from one feeder to the other. The load transfer between different substations can be obtained by branch exchange type switching. By closing the sectionalizing switch (CB8) and opening (CB6), the loads at branches 18, 19, 20 are transferred from SS2 to SS1. The network reconfiguration is made a) b) To reduce the system power loss. To relieve the overloads in the network.

The test system consists of three feeders , 16 buses and 16 switches fig 4.Out of the 16 switches, 13 are normally closed (sectionalizing) and 3 are normally opened (tie) switches. The operation of (s12, s15) makes transfer of load 5 from feeder 1 to feeder 2 by opening a sectionalizing switch s12 and closing the tie switch s15. Hence, closing a switch should always follow the opening of a switch. The load at bus 11 can be transferred to feeder 1 by closing the tie switch 15 and opening the sectionalizing switch 19. Similarly, other loads can be transferred from feeder by switching operations. 3. FORMULATION OF THE PROBLEM In this section, the network reconfiguration problem for loss minimization is discussed in detail. To simplify the presentation, we will represent the system on a per phase basis and the load along a feeder section as constant P, Q loads placed at the end of the lines. It is assumed that every switch is associated with a line in the system. The system of fig.1 can be translated to an equivalent network as shown in fig.2. In this figure, solid branches represent the lines that are in service and constitute the base radial configuration; dotted branches (branches 20, 21, 22) represent the lines with open switches. The base network can be reconfigured by closing an open branch, say branch 21 in the figure. Since this switching will create a loop in the system, (composed of branches 1, 2, 3, 21, 11, 10, 9, 8, 7 and 15), a branch in the loop containing a switch has to be opened, say branch 7 , to restore the radial structure of the system. As a result of this switching, the loads between the branches 7-11 will be transferred from one feeder to the other.

200

J. Electrical Systems 2-4 (2006): 198-207

The network reconfiguration problem for loss reduction involves the load transfer between the feeders or substations by changing the position of the switches. The radial configuration corresponds to a spanning tree of a graph representing the network topology. The minimal spanning tree problem can be stated as follows. Given a graph. Find a spanning tree such that the objective function is minimized while the following constraints are satisfied, (i) voltage constraints, (ii) radiality constraints. This is a combinatorial optimization problem since the solution involves the consideration of all possible spanning trees.
SS1 1 2 3 4 5 6 SS2 20 19 18 17

15 16 7 8 9 10

21 11

22 12 13 14

Fig. 2. One line diagram of a small distribution network

3.1 Mathematical Formulation The loss reduction in network reconfiguration problem is formulated as Min
n 1 i =0

Pi 2 + Qi2 Vi 2

(1)

Such that
Vi min < Vi < Vi max

(2)

Where n = number of buses ri = Resistance of the branch Pi = Real power flowing through the branch Qi = Reactive power flowing through the branch Vi = Voltage at the receiving end of the branch. The network reconfiguration has to obey the following rules: 1) 2) No feeder section can be left out of service Radial network structure must be retained.

2.3 Power flow equations Power flow in a radial distribution network can be described by a set of recursive equations called Dist Flow branch equations that use the real power, reactive power and voltage at the sending end of a branch to express the same quantities at the receiving end of the branch as

201

P. Subburaj et al: Distribution System Reconfiguration for Loss Reduction using Genetic Algorithm

Pi +1 = Pi ri

Pi 2 + Qi2 PLi +1 Vi 2 Pi 2 + Qi2 QLi +1 Vi 2 Pi 2 + Qi2 Vi 2

i i+1
0

(3)
Qi +1 = Qi x i

(4)

Vi 2 1 = Vi 2 2 (ri Pi + x iQi ) + (ri2 + x i2 ) +

0 i-1

(5)
n

P0, Q0

Pi-1, Qi-1

Pi, Qi

Pi+1, Qi+1

Pn, Qn

PLi-1, QLi-1 PLi, QLi PLi+1, Q Li+1

PLn, QLn

Fig. 3. One line diagram of a radial network

Dist Flow branch equations can be written in backward, by using the real power, reactive power, voltage at the receiving end of a branch to express the same quantities at the sending end of the branch as
Pi 1 = Pi + ri Pi ' 2 + Qi' 2 + PLi Vi 2 Pi ' 2 + Qi' 2 + QLi Vi 2 Pi ' 2 + Qi' 2 Vi 2

(6)

Qi 1 = Qi + x i

(7)

2 Vi 1 = Vi 2 + 2 (ri Pi ' + x iQi' ) + (ri2 + x i2 )

(8)

where
Pi ' = Pi + PLi

(9)

and
Qi' = Qi + QLi

(10)

3.3 Simplified dist flow method The quadratic terms in the Dist Flow branch equations represent the losses on the branches and hence they are much smaller than the branch power terms Pi, Qi. By dropping the second order terms, the approximate power flow equations are of the form
Pi +1 = Pi PLi +1 Qi +1 = Qi QLi +1

(11) (12) (13)

Vi 2 1 = Vi 2 2 (ri Pi + x iQi ) +

202

J. Electrical Systems 2-4 (2006): 198-207

The power loss in a branch is expressed as

LPi = ri Pi 2 + Qi2 Vi 2

(14)

Where LPi = Power loss in ith branch 4. GA BASED ALGORITHM The coding scheme and the step by step algorithm of the proposed GA for loss minimum is given below. The coding scheme is illustrated in table I. 0 indicates an open switch and 1 indicates a closed switch. There are 13 sectionalizing switches and 3 tie switches for the 16-bus distribution system. Hence the GA encoding requires 16 bits for a chromosome string as given in table I. The basic configuration of 16-bus distribution system consists of three normally open tie-switches. Tie switches s15, s21, s26 are coded as 0. The remaining 13 switches are coded as 1.
Table I. Coding scheme for the distribution system. 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0

The proposed algorithm steps can be summarized as follows. 1. Read the bus data and line data for the distribution test system. 2. Generate number of random variables in 1 and 0s of 16 bit size as per the population size and initialize maximum number of iterations for genetic algorithm. 3. Check the radiality; if the distribution network has no closed loops and all the loads were connected, the network is radial, otherwise go to step2. 4. Choose the initial population that obeys the radiality and also check the sending end and receiving end connection for the branches in the network. 5. Calculate the fitness for the population. (Eqn-1) 6. Perform cross over operation by taking two chromosomes at a time from the population. Generate a random number for the crossover point. Perform crossover operation by interchanging the bits in the left side of the cross over point. 7. Perform mutation operation (after cross over operation) by generating two random numbers mutating the chromosomes and change the bit 1 to 0 or 0 to 1 for the two chromosomes. 8. If the off spring does not obey the radiality replace it with 1s. 9. Sort the population and off springs. 10. Choose the best individuals of population size according to their fitness. 11. Steps 6 to 10 are repeated until maximum iteration is reached. 12. The minimum loss, maximum loss, average loss and fitness values are displayed for the given numbers of iteration. 13. The switch status for the minimum loss configuration is also displayed. 14. Stop the process.

203

P. Subburaj et al: Distribution System Reconfiguration for Loss Reduction using Genetic Algorithm

5. RESULTS AND DISCUSSION To illustrate the efficiency of the proposed idea for network reconfiguration, 16 bus distribution systems is used as test system. The numerical data for 16 bus distribution system is tabulated in appendix. The simulation studies are carried out on Intel Pentium IV, 3.0 GHz system in MATLAB environment. The voltages for this system are calculated by the power flow equations. The real power loss for the basic configuration is 0.006157 p.u. The switches that are opened are tieswitches are s15, s21, s26 as shown in fig.6. The minimum voltage level is 0.970051 p.u. The cross over rate and mutation rate are chosen as 0.8 and 0.05. The population size is varied from 10 to 20. The no. of iterations is chosen as 200. The real power loss for the loss minimum configuration is 0.005629 p.u. The minimum voltage level improves to 0.972244 p.u. The switches that are opened for the minimum loss configuration are s17, s19, s26 as shown in fig. 4 and the status of the switch is also given in table IV. The fitness values for various population sizes are compared and tabulated in table III. The relative average fitness decreases gradually when the number of iterations increases and reaches a constant value as shown in fig. 5. It was observed that for low population size(10), the solution may not converge to minimum loss configuration, while a population size of 20 gives minimum loss configuration.
FEEDER-I FEEDER-II FEEDER-III

1 s11 8

s16 s17 s18

2 s24 14

3 s22 s21 13

S
4 s12 s15

s19

9 10 s20

5 s13 s14 6

11 12

s23

s26 7 16

s25
15

Fig 4. Minimum loss configuration system

7.4 7.2 7 Average fitness value 6.8 6.6 6.4 6.2 6 5.8 x 10
-3

20

40

60

80 100 120 No.of iterations

140

160

180

200

Fig 5. Iteration performance for minimum loss configuration.

204

J. Electrical Systems 2-4 (2006): 198-207

Table II: Loss analysis of the test system Minimum Loss Loss Minimum / Voltage Improvement Loss Voltage Basic Configuration (p.u) 0.006157 0.970051 Configuration (p.u) 0.005629 0.972244 Percentage % 8.58 0.227

Table III: Fitness value results Minimum Population size Fitness Value 10 20 0.005887 0.005629 Fitness Value 0.006032 0.005887 Fitness Value 0.006129 0.005817 Average Maximum

Table IV: Switch status for the minimum loss configuration 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 0

6. CONCLUSION In this paper, distribution systems loss minimum re-configuration methodology using genetic algorithm was proposed. The solution methodology employs a search over different radial configurations by considering branch exchange type switching. GA was able to produce a near optimal solution by adopting the adaptive nature of natural genetics. From the numerical example, it was seen that the estimation method is computationally efficient and the loss-reduction of 8.58% is achieved by this algorithm. This result demonstrates the validity and effectiveness of the proposed methodology. REFERENCES
[1] [2] [3] [4] [5] S. Civanlar, J. J. Grainger, H. Yin, S. S. H. Lee, Distribution feeder reconfiguration for loss reduction, IEEE Trans.Power Del., Vol.3,No.3,pp.1217-1223, July1998. M. E. Baran and F. F. Wu, Network reconfiguration in distribution systems for loss reduction and load balancing, IEEE Trans. Power Del., Vol.4, No.2, pp1401-1409, April1989. D. Shirmohammadi and H. Wayne Hong, Reconfiguration of electric distribution networks for resistive line losses reduction, IEEE Trans.Power Del., Vol.4, No.2, pp1492-1498, April1989. T. Taylor and D. Lubkeman, Implementation of heuristic search strategies for distribution feeder reconfiguration, IEEE Trans.Power Del., Vol.5, No.1, pp239-246, Jan1990. H. D. Chiang and R. J. Jumeau, Optimal network reconfigurations in distribution systems: part2: solution algorithms and numerical results, IEEE Trans.Power Del., Vol.5, No.3, pp15681574, July1990. H. D. Chiang and R. J. Jumeau, Optimal network reconfigurations in distribution systems: part1: a new formulation and a solution methodology, IEEE Trans.Power Del., Vol.5, No.4, pp1902-1908, Nov 1990. S. K. Goswami and S. K. Basu,A new algorithm for the reconfiguration of distribution feeders for loss minimization, IEEE Trans.Power Del., Vol.7, No.3, pp1484-1491, July1992.

[6]

[7]

205

P. Subburaj et al: Distribution System Reconfiguration for Loss Reduction using Genetic Algorithm

[8]

K. Nara, A. Shiose, M. Kitagawa, T. Ishihara, Implementation of genetic algorithm for distribution systems loss minimum reconfiguration, IEEE Trans.Power Syst., Vol.7, No.3, pp1044-1051, August 1992. H. kim, Y. Ko, K. H. Jung, Artificial neural-network based feeder reconfiguration for loss reduction in distribution systems, IEEE Trans.Power Del., Vol.8, No.3, pp1356-1366, July1993.

[9]

[10] V. Borozan, D. Rajicic, R. Ackovski, Improved method for loss minimization in distribution networks, IEEE Power Syst., Vol.10, No.3, pp1420-1425, August 1995. [11] R. Taleski and D. Rajicic, Distribution network reconfiguration for energy loss reduction, IEEE Trans.Power Syst., Vol.12, No.1, pp398-406, Feb1997. [12] Y. J. Jeon and J. C. Kim, Network reconfiguration in radial distribution system using simulated annealing and tabu search, in Proc.IEEE Power Eng.soc.Winter Meeting, Jan 2000, pp23-27. [13] H. C. Chin and K. Y. Huang, A simple distribution reconfiguration algorithm for loss minimization, IEEE Trans.Power Systems., Vol.2, No.2, pp607-611, May2000. [14] A. B. Morton and I. M. Mareels, An efficient brute-force solution to the network reconfiguration problem, IEEE Trans.Power Del., Vol.15, No.3, pp996-1000, July2000. [15] W. M. Lin, F.-S. Cheng, M. T. Tsay, Distribution feeder reconfiguration with redefined genetic algorithm, in IEEE Proc.Gener., Transm., Distrib., Vol.147, No.6, pp349-354.Nov 2000. [16] Y. Y. Hong and S. Y. Ho, Genetic algorithm based network reconfiguration for loss minimization in distribution systems, proc., pp486-490, in IEEE Proc., 2003 [17] V. C. Veerareddy, N. Perumal, Y. Rajasekharareddy, Loss reduction in distribution networks by network reconfiguration: a two stage solution approach, in National power &Energy Conference 2004 proc, Malaysia, pp241-246., 2004. [18] K. Prasad, R. Ranjan, N. C. Sahoo, A. Chaturvedi, Optimal reconfiguration of radial distribution systems using a fuzzy mutated genetic algorithm, IEEE Trans.Power Del., Vol.20.No.2, pp1211-1213, April 2005. [19] Y. Y. Hong and S. Y. Ho, Determination of network configuration considering multiobjective in distribution systems using genetic algorithms, IEEE Trans.Power Syst., Vol.20, No.2, pp1062-1069, May 2005. [20] S. Sivanagaraju, N. Visali, V. Sankar, T. Ramana, Enhancing voltage stability of radial distribution systems by network reconfiguration, Elec.Power.Comp.Syst.,Vol.33,pp539550,May 2005.

APPENDIX
FEEDER I
1 s11 8 s18 4 s12 s15 s20 5 s13 s14 6 7 s26 16 11 12 s23 s19 s17 s21 9 10 14 s24

FEEDER - II
s16 2

FEEDER - III
3 s22 13

s25
15

Fig 6. Basic configuration of test system

206

J. Electrical Systems 2-4 (2006): 198-207

Table A: Switch Status

Table B: Network data for the 16-bus distribution system

Switch S. No & ID Bus to Bus 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. S11 S12 S13 S14 S16 S18 S17 S19 S20 S22 S24 S23 S25 S15 S21 S26 1-4 4-5 4-6 6-7 2-8 8-9 8-10 9-11 9-12 3-13 13-14 13-15 15-16 5-11 10-14 7-16 Section Resistance (p.u) 0.075 0.080 0.090 0.040 0.110 0.080 0.110 0.110 0.080 0.110 0.090 0.080 0.040 0.040 0.040 0.090 Section Reactance (p.u) 0.10 0.11 0.18 0.04 0.11 0.11 0.11 0.11 0.11 0.11 0.12 0.11 0.04 0.04 0.04 0.04 2.0 3.0 2.0 1.5 4.0 5.0 1.0 0.6 4.5 1.0 1.0 1.0 2.1 1.6 1.5 0.8 1.2 2.7 3.0 0.9 0.1 2.0 0.9 0.7 0.9 1.0 1.8 1.8 0.6 3.7 1.2 1.1 1.2 End Bus Load (MW) End Bus Load MVAR End BusCapacitor MVAR

207