The Online Journal on Electronics and Electrical Engineering (OJEEE)

Vol. (1) No. (1)

A New Approach for Capacitor Allocation in Radial Distribution Feeders

Mohamed M. Hamada, Mohamed A. A. Wahab, Abou-Hashema M. El-Sayed, and Husam A. Ramadan Department of Electrical Engineering Department, Faculty of Engineering, Minia University, Minia, Egypt
Abstract-This paper presents a new approach based on the total reactive loss in a distribution system to decide the optimal locations and sizes of the compensation shunt capacitors. This approach is implemented in two stages. First, a capacitor rated at the reactive power needed at certain busbar is directly connected at this busbar. This is applied on all buses of the distribution system simultaneously. The second is to connect a single capacitor rated at the total reactive power losses needed by the distribution system at certain bus. Optimization technique is used to determine the candidate bus at which this capacitor should be connected. This technique is built to satisfy the voltage profile and sectional loss constraints. This approach is tested on two practical distribution systems with 9 and 34-buses respectively. Computational results showed that the proposed approach could quickly achieve an optimal or near-optimal solution. The results of the proposed approach are compared with the results of those in literature. The comparison showed that the proposed approach achieves the goals of solving reactive power compensation problem for distribution systems despite its simplicity. Keywords- Distribution systems, compensation, optimization technique. I. INTRODUCTION Distribution networks are well known for their high R /X ratio and significant voltage drop that could cause substantial power losses along the feeders. It is estimated that as much as 13% of the total power generation is lost in the distribution networks [1]. Portion of this loss is caused by the reactive current flowing in the network. Voltage profiles throughout the network have to be kept at acceptable levels to ensure service reliability among other issues. Capacitor banks are commonly used in various parts of the electric grid to maintain voltage levels within appropriate limits, minimize the power losses and release the line capacity. With regard to the power losses in the feeders, capacitor installations have demonstrated their effectiveness in reducing the overall current by canceling part of the reactive current supplied by the substation. Published literature describing capacitor allocation algorithms are abundant. Lee et al. [2] pioneered the analytical methods. In Ref. [2], fixed and switched capacitors are placed for optimizing the net monetary savings associated with the reduction of power and energy losses, both the capacitor locations and sizes are treated as reactive power continuous variables. A voltage dependent methodology for shunt capacitor compensation of primary distribution feeders is presented in Ref. [3]. Ponnavaikko and Rao [4] used a numerical method called the method of local variations and further expanded the problem to include the effects of load growth, and switched capacitors for varying load. Similarly, Baran and Wu [5] formulated the capacitor placement problem using mixed integer programming. The optimal selection and placement of capacitor banks using binary particle swarm optimization (PSO) is integrated with the estimation of harmonic levels in Ref. [6]. Methods based on heuristic search techniques are introduced for distribution system loss reduction by system reconfiguration [7], [8]. Abdel-Salam et al. [9] proposed a heuristic technique to identify a section in the distribution system with the highest losses due to reactive load currents and then pinpoint the sensitive node in that section having the greatest effect on the system loss reduction. Sizes of capacitors placed on the sensitive nodes are then determined by maximizing the power loss reduction from capacitor compensation. Chis et al. [10] improved the work of [9] by determining the sensitive nodes that have the greatest impact on loss reduction for the entire distribution system directly and by optimizing for the capacitor sizes based on maximizing the net economic savings from both energy and peak power loss reductions. In addition, the method in [10] also accounts for varying loads of the distribution system considered. Mekhamer et al. [11] present review and implementation of previous strategies and comparison between them and the so-called exact solution. The best results were offered by the exact solution. Most of previous studies consider the loss reduction for the capacitor allocation problem [2-11], but these studies take in consideration the reduction of total losses regardless the variation in the individual sectional losses. Hamada et al. [12] introduced a new strategy for capacitor allocation handling the reduction in the section losses by adding a new constraint to the well-known constraint (allowed voltage violation constraint). The new constraint has been the sectional ohmic losses in each branch of the feeder. This paper presents a new simple approach to determine the optimal sizes and locations of capacitor banks in radial distribution feeders. From the point of view of this approach, the amount of compensation required for any feeder is equal to total reactive loads in addition to the total reactive losses in all sections of this feeder. This approach first commends

Reference Number: W09-0006

24

The Online Journal on Electronics and Electrical Engineering (OJEEE)

Vol. (1) No. (1)

allocating capacitor at each bus of the feeder equal to its reactive load to achieve unity power factor at each bus, then performing the load flow for the feeder after installing these capacitor to obtain the total reactive losses in this case. These reactive losses are compensated by a capacitor its size equals these losses. This capacitor is placed at a candidate bus. The candidate bus is chosen to achieve minimum total losses and satisfy the voltage constraint and the sectional losses constraint mentioned in [12]. This paper presents implementation of the new approach and the so-called exact solution in [11], and a comparison between them to show the validity of proposed approach. The compression is made at the same conditions and assumptions. II. PROBLEM FORMULATION A. Problem Formulation The purpose of placing compensating capacitors along the distribution feeders is to lower the total power loss and bring the bus voltages within specified limits while minimizing the total power cost. The total power loss PTloss is given by

Fig.1. One-line diagram of a radial distribution feeder. where, (k+1) is the case after the capacitor placement and (k) is the case before the capacitor placement. B. The Proposed Strategy The proposed approach is to connect a capacitor rated at the reactive power needed at certain bus directly at this bus. This is applied on all buses of the distribution system simultaneously. The total reactive power losses are calculated from the results of a load flow program. Then add a capacitor its value equal to the resulted Q total loss at the candidate bus. The candidate bus is chosen to minimize the cost function in eq. (3), and satisfies constrains of eq. (4) and eq. (5). The solution algorithm of this approach can be summarized as follows: 1- Perform the load flow program (LF) for the original feeder to get the Section losses, the voltage at each bus, the total active losses and the total reactive losses. 2- Put QC at all buses simultaneously equals the value QC = Q load (i) where i index is the bus number. 3- With QCs are placed run the LF program. If the voltage is less than Vmin and the sectional losses are less than that of case before adding QC, put QC = Q total losses at the bus which satisfies eq. (3), eq. (4) and eq. (5). 4- Repeat step 3 until satisfying all constraints. C. Exact Solution. [11] As it was stated at [11], the solution algorithm of this method can be summarized as follows: 1- Assume that we should put QC at bus number one. Try all of the 27 values of QC at this bus. For each value of QC, perform the load flow program and evaluate the cost function. Record QC that gives minimum cost and record this cost. 2- Repeat step 1 for all buses. Then select to put QC1 at the bus (K1) that has lowest minimum cost function. 3- With optimal QC placed at bus K1, repeat step 1and 2 to select the next candidate bus (K2) where the next optimal QC2 will be placed. 4- With QC1 and QC2 placed at buses K1 and K2, repeat steps 1 and 2 and so on until candidate buses are exhausted. 5- If the minimum voltage is still <0.9, try to increase QC of the candidate node that is very far from the substation. This will increase the voltage such that 0.9 Vi 1 p.u.

P Tloss

P loss(i,
i0

n -1

i 1)

, i = 1,..., n

(1)

where, i is the bus number and n is the total number of buses as shown in Fig. 1. Considering investment cost, there is a finite number of standard capacitor sizes that are integer multiples of the smallest size Qc. the cost per kVAr varies from one size to another. Generally, large capacitor sizes are cheaper than smaller ones. Let the maximum permissible capacitor size be limited to [11]:
c Q max c 0

H Q 0c

(2)

where, Q is the smallest capacitor size in Table (A-1) in Appendix (A) and H is an integer resulted by dividing the
c maximum allowed capacitor size for this feeder Q max on the

smallest value Q 0c ; then, at each selected location, there are

c c H sizes to choose from. Let K 1c , K 2 be the ,...., K H corresponding capital investment per kVAr. Assuming that only capacitor banks are used for voltage excursions, the objective (cost) function can be selected as

c Cost C P PTloss K c jQ j j 1

(3)

Where, Cp is the cost per power loss ($/kW) and j = 1, 2,, J represents the selected buses. The objective function eq. (3) is to be minimized subjected to two constraints: The first constraint: Vmin Vi Vmax The second constraint:
P (k 1) Sec.loss P

, i=1, 2,., n.

(4)

(k) Sec.loss

(5)

Reference Number: W09-0006

25

The Online Journal on Electronics and Electrical Engineering (OJEEE)

Vol. (1) No. (1)

III. NUMERICAL RESULTS In order to test the proposed approach and validate its results, two distribution systems with 9 and 34-buses respectively have been taken as examples. The constants and conditions are as follows [11]: Power losses cost Cp = 168 U.S.$168/kW Voltage limits: Vmin= 0.9 p.u. Vmax=1.1 p.u. Fixed capacitors are only considered. The possible capacitor sizes have been shown in Table (A-2) in Appendix (A) with corresponding cost/kVAr [13]. The values of the choices are derived from Table (A-1) by assuming a life expectancy of ten years (the operating costs are neglected) [11]. A. The First Test Feeder The 9-bus radial distribution feeder in [3] is taken as the test feeder. The rated voltage is 23 kV. The single line diagram, the load data and feeder-line parameters for the system are shown in Appendix (B). By applying the load flow program on this feeder before compensation, the value cost and the total power losses are U.S. $ 131675 and 783.8 kW, respectively. The maximum and the minimum bus voltage magnitudes were 0.9929 and 0.8375 p.u respectively, where the voltage of the substation (bus number 0) is assumed to be 1 p.u. The results for this feeder are summarized in Table I. B. The Second Test Feeder Another 34-bus radial distribution test system has been used. This test system has a main feeder and four laterals (sub feeders). The data of the feeder is presented in [12]. The single line diagram, the load data and feeder line parameters for the system are shown in Appendix (B). The system voltage is 11kV. Before compensation, the cost is U.S. $327212, this is based on the previously defined cost function, the active and reactive losses are 221.5kW and 65.04kVAr, respectively, and the voltage limits in p.u. are 0.9417 Vi 1.The results for this feeder are summarized in Table II. C. Discussion on results Table I shows the results of the two methods for the first feeder. It is observed that the total active loss for the proposed approach is less than that of the exact solution. The fixed cost (the cost of the installed capacitors only), and the running cost (the cost of the total power losses only) for the proposed approach are less than that of the exact solution; so the proposed approach offers more cost saving than the exact solution. The voltage profile for the proposed approach is more close to that of the exact solution as shown in Fig.2. The active loss in each section of this feeder is shown in Fig. 3. The sectional loss resulting by using the proposed approach is always less than that of the original case (i.e. case without QC s), but the sectional loss resulting from the exact solution offers a violation in section (2-3) (i.e. the loss in this section

TABLE I RESULTS FOR THE TWO METHODS APPLIED TO THE 9-BUS FEEDER INCLUDING THE ORIGINAL CASE. Method No. Without Qc Qc in MVAr 0 0 0 0 0 0 0 0 0 0.783 131675 0 131675 1 0.8375 0 0 The Proposed approach Qc in MVAr 0.46 1.227 0.446 1.84 1.495 0.996 0.06 0.13 0.2 0.684 114912 1436.5 116348.5 1 0.891 99 15326.5 Exact solution in Ref. [11] Qc in MVAr 0 3.6 0 4.05 1.65 0 0 0.6 0 0.686 115248 1787.4 117035 1.007 0.9003 97 14640

Bus No 1 2 3 4 5 6 7 8 9 Total power loss in MW Running cost $ Fixed cost $ Total cost $ Max voltage in p.u. Min voltage in p.u. Loss reduction in kW Cost saving in $

TABLE II RESULTS FOR THE TWO METHODS APPLIED TO THE 34-BUS FEEDER INCLUDING THE ORIGINAL CASE Method No. Total power loss in MW Total cost $ Max voltage Min voltage Loss reduction in kW Cost saving in $ Without Qc 0.2215 37212 1 0.9417 0 0 The Proposed approach 0.1578 27980.9 1 0.951 63.7 9231.1 Exact solution in Ref. [11] 0.1655 28250 1 0.951 56 8962

is more than that of the original case). The results for the second feeder are shown in Table II. It is observed that the proposed approach offers less total power loss, less cost, hence more cost saving and more loss reduction than the exact solution. The voltage profile for the proposed approach for 34-bus system is the best case as shown from Fig. 4. The sectional losses for the two methods and the original case for this feeder are shown Fig. 5. The sectional loss resulting by using the proposed approach is always less than that of the original case, also is less than that resulting by the exact solution.

Reference Number: W09-0006

26

The Online Journal on Electronics and Electrical Engineering (OJEEE)

Vol. (1) No. (1)

IV. CONCLUSION This paper proposes an efficient and simple approach for the optimal location and sizing of fixed shunt capacitor in radial distribution feeders based on the total reactive powerrequired for the distribution feeder. The following conclusions may be deduced: The power losses of distribution feeders can be effectively reduced by proper reactive compensation equal to its total reactive loads in addition to the total reactive losses in the feeder sections.

By this approach, the losses in each section of the feeder are not allowed to increase than that of the original case (without capacitor placement) during the capacitor allocation and sizing scenario. In addition to power-loss reduction, the voltage profile can be improved as well by the proposed approach V. REFERENCES

[1] H. N. Ng, M. M. A. Salama, and A. Y. Chikhani, Classification of capacitor allocation techniques, IEEE Trans. Power Delivery, vol. 15, no. 1, pp. 387392, Jan. 2000.

Reference Number: W09-0006

27

The Online Journal on Electronics and Electrical Engineering (OJEEE)

Vol. (1) No. (1)

[2] S. H. Lee, J. J. Grainger, Optimum placement of fixed and switched capacitors on primary distribution feeders, IEEE Trans. Power Apparatus and Systems, vol. PAS100, no. 1, pp. 345-352, Jan. 1981. [3] J. J. Grainger, S. H. Lee, Capacity release by shunt capacitor placement on distribution feeders: a new voltage- dependent model, IEEE Trans. Power Apparatus and Systems, vol. PAS- 101 no. 5, pp. 12361244. May 1982. [4] M. Ponnavaikko and K. S. Prakasa Rao, Optimal choice of fixed and switched shunt capacitors on radial distributors by the method of local variations, IEEE Trans. Power Apparatus and Systems, vol. 102, no. 6, pp. 16071615, June 1983. [5] M. E. Baran and F. F.Wu, Optimal capacitor placement on radial distribution systems, IEEE Trans. Power Delivery, vol. 4, no. 1, pp. 725734, Jan. 1989. [6] T. M. Khalil, H. K. M. Yousssef, and M. M. Abdel Aziz, Optimal capacitor placement on radial distribution feeders in presence of nonlinear loads using binary particle swarm optimization, in Conf. 19th International Conference on Electricity Distribution Conf. ,Vienna, 21-24 may 2007. [7] S. Civanlar, J. J. Grainger, H. Yin, and S. H. Lee, Distribution feeder reconfiguration for loss reduction, IEEE Trans. Power Delivery, vol. 3, no. 3, pp. 1217 1223, July 1988. [8] T. Taylor and D. Lubkeman, Implementation of heuristic search strategies for distribution feeder reconfiguration, IEEE Trans. Power Delivery, vol. 5, no. 1, pp. 239246, Jan. 1990. [9] T. S. Abdel-Salam, A. Y. Chikhani, and R. Hackam, A new technique for loss reduction using compensating capacitors applied to distribution systems with varying load condition, IEEE Trans. Power Delivery, vol. 9, no. 2, pp. 819827, Apr. 1994. [10] M. Chis, M. M. A. Salama, and S. Jayaram, Capacitor placement in distribution systems using heuristic search strategies, IEE Proceedings Generation, Transmission and Distribution, vol. 144, no. 2, pp. 225230, May 1997. [11] S. F. Mekhamer, S. A. Soliman, M. A. Moustafa, and M. E. El- Hawary, Application of fuzzy logic for reactivepower compensation of radial distribution feeders, IEEE Trans. Power Systems, vol. 18, no. 1, pp. 206213, Feb. 2003. [12] M. M. Hamada, M. A. A. Wahab, A. M. El-Sayed, H. A. Ramadan A proposed strategy for capacitor allocation in radial distribution feeders, in Conf. 12th Middle East Power Systems Conf., MEPCON, Aswan, Egypt, 12-15 March 2008. [13] Y. Baghzouz and S. Ertem, Shunt capacitor sizing for radial distribution feeders with distorted substation voltages, IEEE Trans. Power Delivery, vol. 5, pp. 650 657, Apr. 1990.

VI. APPENDICES Appendix (A)

Table (A-1) Available three- phase capacitor size and cost [13]. Size (kVAr) Cost ($) 150 750 300 975 450 1140 600 1320 900 1650 1200 2040

Table (A-2) Possible choice of capacitor sizes and cost / kVAr . J 1 2 3 4 5 Qc (kVAr) 150 300 450 600 750 $/kVAr 0.5 0.35 0.253 0.22 0.276 J Qc(kVAr) $/kVAr J Qc(kVAr) $/kVAr J Qc(kVAr) $/kVAr J Qc(kVAr) $/kVAr 7 1050 0.228 13 1950 0.211 19 2850 0.183 25 3750 0.183 8 1200 0.17 14 2100 0.176 20 3000 0.18 26 3900 0.182 9 1350 0.207 15 2250 0.197 21 3150 0.195 27 4050 0.179 10 1500 0.201 16 2400 0.17 22 3300 0.174 11 1650 0.193 17 2550 0.189 23 3450 0.188

6 900 0.183 12 1800 0.187 18 2700 0.187 24 3600 0.17

Appendix (B) The 9-bus test system

Fig.(B-1) The single line diagram of the 9-bus test system [3].

Table (B-1) The load data and the feeder data of the 9-bus test system [3]. Bus N o. 1 2 3 4 5 6 7 8 9 Load P (kW) 1840 980 1790 1598 1610 780 1150 980 1640 Q (kVAr) 460 340 446 1840 600 110 60 130 200 Bus No. From 0 1 2 3 4 5 6 7 8 To 1 2 3 4 5 6 7 8 9 0.123 0.014 0.746 0.698 1.983 0.905 2.055 4.795 5.343 0.4127 0.6051 1.205 0.6084 1.7276 0.7886 1.164 2.716 3.0264 Sectional parameters Ri,i+1(p.u.) Xi,i+1(p.u.)

Reference Number: W09-0006

28

The Online Journal on Electronics and Electrical Engineering (OJEEE)

Vol. (1) No. (1)

The 34-bus test system

Table (B-2) The load data and the feeder data of the 34-bus test system [11]. Bus N o. Load P (kW) Q (kVAr) Sectional parameters Bus No From 0 1 2 3 4 5 6 7 8 9 10 2 2_1 2_2 2_3 5 5_1 5_2 5_3 5_4 5_5 5_6 5_7 5_8 5_9 5_10 6 6_1 6_2 9 9_1 9_2 9_3 To 1 2 3 4 5 6 7 8 9 10 11 2_1 2_2 2_3 2_4 5_1 5_2 5_3 5_4 5_5 5_6 5_7 5_8 5_9 5_10 5_11 6_1 6_2 6_3 9_1 9_2 9_3 9_4 Ri,i+1 (/km) 0.195 0.195 0.299 0.299 0.299 0.524 0.524 0.524 0.524 0.524 0.524 0.524 0.524 0.524 0.524 0.299 0.299 0.378 0.378 0.378 0.524 0.524 0.524 0.524 0.524 0.524 0.524 0.524 0.524 0.524 0.524 0.524 0.524 Xi,i+1 (/km) 0.08 0.08 0.083 0.083 0.083 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.083 0.083 0.086 0.086 0.086 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.6 0.55 0.55 0.5 0.5 0.6 0.4 0.6 0.4 0.25 0.2 0.3 0.4 0.2 0.1 0.6 0.55 0.55 0.5 0.5 0.5 0.5 0.6 0.4 0.25 0.2 0.3 0.3 0.3 0.3 0.4 0.3 0.2 Length km

Fig.(B-2) The single line diagram of the 34-bus test system [11]. .

1 2 3 4 5 6 7 8 9 10 11 2_1 2_2 2_3 2_4 5_1 5_2 5_3 5_4 5_5 5_6 5_7 5_8 5_9 5_10 5_11 6_1 6_2 6_3 9_1 9_2 9_3 9_4

320 0 230 230 0 0 230 320 0 230 137 72 72 72 13.5 230 230 230 230 230 230 230 230 230 230 137 75 75 75 57 57 57 57

142.5 0 142.5 142.5 0 0 142.5 142.5 0 142.5 84 45 45 45 7.5 142.5 142.5 142.5 142.5 142.5 142.5 142.5 142.5 142.5 142.5 85 48 48 48 34.5 34.5 34.5 34.5

Reference Number: W09-0006

29