16th NATIONAL POWER SYSTEMS CONFERENCE, 15th-17th DECEMBER, 2010

176

Topology based approach for Efficient Load flow Solution of Radial Distribution Networks
K.V.S. Ramachandra Murthy, M. Ramalinga Raju, G. Govinda Rao, and K. Narasimha Rao
Abstract: In this paper, load flow solution is obtained for radial distribution system using a new algorithm. Network topology is exploited to build two matrices for obtaining load flow solution of radial distribution system. Effective load at each node consists of two parts. Part one is the sum of all active and reactive powers connected to nodes beyond a particular node. And part two is the power loss in all the lines connected beyond the node under consideration. Two new network matrices are proposed to obtain the two parts of the effective load. bus injection to Node power matrix is used to identify first part of the effective load at each node and Line loss to node power matrix is used to calculate the second part of the effective load at each node. In this method, simple algebraic expression of voltage magnitude is used. This method is very efficient and requires very less computer memory. The proposed algorithm is applied to 15bus and 85 bus systems used in the literature and the results are compared. Results are accurate up to the fourth decimal.

topology based technique for three phase un-balanced systems is developed by J.H. Teng [3]. Equivalent current injection at each node is considered for the analysis in their work. Shirmohamadi et al [3] have presented a compensation based power flow for weakly meshed transmission and distribution systems. Baran and Wu [5] and Chiang [6] have obtained the load flow solution in a distribution system by iterative solution of three fundamental equations representing real power, reactive power and voltage magnitude. In many developing countries, the 11KV rural distribution feeders are radial and too long. The voltages at the far end are very low with very high regulation. In this paper, the proposed method involves only evaluation of a simple algebraic expression of voltage magnitude [2] and no trigonometric terms. Topology based approach is used for evaluating equivalent load at every node. This eliminates the complex process of identifying nodes connected beyond a particular node as described in [2]. The two developed matrices, bus injection to node power matrix and
line loss to node power matrix are very easy to form. The features of this method are robustness and computer economy. Convergence

I.

INTRODUCTION

Distribution load flow is important for distribution automation systems, and distribution management systems. Network optimization, VAR planning, switching, state estimation etc. need the support of robust and efficient load flow solution. Traditional transmission system load flow methods Gauss-Siedel and Newton Raphson techniques, cannot be used for distribution systems as R/X ratio is high. Kresting and Mendive [1] have presented a load flow technique based on the ladder network theory. Teng and Lin(1994) proposed solution for meshed topology In India, all the 11KV rural distribution feeders are radial and too long. The voltages at the far end of many such feeders are very low with very high voltage regulation. D. Das and D. P. Kothari have proposed a load flow solution method by writing a algebraic equation for bus voltages [2]. Network

is always guaranteed. The assumption is that shunt capacitance is negligible at the distribution voltage level. II. FORMULATION FOR THE MODEL

V11

R+ jX

V22

P + jX2 2

K.V.S. Ramachandra Murthy is working as Associate Professor, College of Engineering., Visakhapatnam. Mail: [email protected]

G.V.P.

M. Ramalinga Raju is Professor and Head of the Department of E.E.E. at University college of Engineering, JNTU, Kakinada, India. G.Govinda Rao and K. Narasimha Rao are working as Professors at G.V.P. College of Engineering., Visakhapatnam., India.

P2 and Q2 are effective active and reactive powers supplied beyond node 2. Formulae have been derived in ref[2]. This model of calculating voltages is termed in [2] as simple and efficient method of load flow by Das and Kothari. The same equations are used along with the matrices developed from network topology. This model also requires less memory and rapid. The load flow equations used are,

(1)

Department of Electrical Engineering, Univ. College of Engg., Osmania University, Hyderabad, A.P, INDIA.

16th NATIONAL POWER SYSTEMS CONFERENCE, 15th-17th DECEMBER, 2010

177

(2) --- (9) (3) B. Line Loss to Node Power matrix: (LLNP) The relation between the line losses and node power can be obtained by following equations. N2 = Sl1 + Sl2 +Sl3+ Sl4 +Sl5 similarly, N3 = Sl4 N4 = Sl5

(4) (5) A. Bus-Injection to Node Power matrix: ( BINP) For distribution systems, the models which are based on the effective load at every node is very convenient. At each bus i the complex power S is specified by, --- (6) Pi and Qi are the real power and imaginary powers at ith node. The set of equations can be written by applying Kirchhoffs current law (KCL) to the distribution network. The KCL equations will equally hold good for powers also. For example, consider a simple 6 bus system. N2, N3, N4 , N5 and N6 are the equivalent powers at each Node N2= S3+ S4 + S5 + S6 N3 = S3 + S5 N4 = S4+ S6

----- (10)

Complex power losses associated with lines 1,2,3,4,5 are given by Sl1, Sl2, Sl3, Sl4 and Sl5 respectively. Sum of the line losses appearing at Node 2, 3, 4, 5 and 6 are given by N2, N3, N4, N5, N6. Total Line loss supplied by node2, N2 End nodes will not have to supply and line loss component. So, they will have all zeros in their corresponding rows. Node 2 will have to supply all the line losses except branch -1 loss. SL is the column matrix containing all line losses. [N] = [LLNP] [SL]

-----

(7)

--- (11)

Where S3, S4, S5 and S6 are complex load powers respectively at buses 3,4,5 and 6. Effective load at Node [N] can be obtained by, [N] = [BINP] [S] ---- (8) The constant BIBP matrix has entries of 1 and 0 only. For a distribution system with m-branch sections and n-load buses, the dimension of the BIBC is m X n. The bus where there is no load need not enter into matrix [S]

Effective load at each node = N + N

--- (12)

After calculating the effective load at each node, recalculate the receiving end voltages using the above equations(1) (3). And calculate power losses using equations (4) and (5). The voltage values at all nodes change after adding losses. Algorithm for the proposed Method : 1. Read the system data, V1= 1.0 pu. Line losses are assumed to be zero in the first iteration. 2. Build BINP matrix and LLNP matrix. 3. Obtain Peffective + j*Qeffective, at each node using equation(8). N represents the part of load powers in the effective load at various nodes. [N] = [BINP] [S] where, S is the column matrix of all loads. Initialize iteration count =1. Obtain receiving end voltages using simple formulae mentioned above.

4. 5. Fig. 1 Sample distribution system.

Department of Electrical Engineering, Univ. College of Engg., Osmania University, Hyderabad, A.P, INDIA.

16th NATIONAL POWER SYSTEMS CONFERENCE, 15th-17th DECEMBER, 2010

178

TABLE II BUS VOLTAGES IN PU OBTAINED IN DIFFERENT ITERATIONS..

6.

Calculate power loss on all lines using the formulae,

Bus voltages in pu. in different iterations Iteration1 0.971806 0.957311 0.951559 0.950573 0.958778 0.956556 0.957502 0.968497 0.967424 0.950614 0.946496 0.945185 0.949264 0.949096 Iteration2 0.971289 0.956676 0.950912 0.949926 0.958238 0.956014 0.95696 0.967976 0.966903 0.94996 0.945836 0.944525 0.948616 0.948447 Iteration3 0.971284 0.956671 0.950907 0.94992 0.958232 0.956009 0.956955 0.967971 0.966898 0.949954 0.945831 0.944519 0.94861 0.948441 Iteration4 0.971284 0.956671 0.950907 0.94992 0.958232 0.956009 0.956955 0.967971 0.966898 0.949954 0.94583 0.944519 0.94861 0.948441

Multiply the power loss column matrix, SL with LLNP matrix to get N matrix. N represents the part of line losses in the effective load at various nodes. [N] = [LLNP] [SL] 8. Calculate the total effective load at various nodes by adding the N and N matrix. Effective load at each node = N + N 9. Increment the iteration count Repeat the steps from step (5) using new effective loads at every node. 10. If the difference in the voltages between present iteration and previous iteration is greater than 0.001 pu, then increment the iteration count and repeat from step(5), otherwise, print the result. III. RESULTS For fifteen bus system, number of iterations taken is 3. The algorithm is efficient, robust and rapid. It requires less memory and convergence is faster. For the first time, topology is fully exploited with the modeling equations suggested in [2]. Voltages obtained in different iterations are presented in Table II. The results are promising and compared with [2] results. Data for the15, 85 bus system is given in [2]. 85 bus system result is compared with the result of [6] in Table III.
TABLE. I COMPARISON OF BUS VOLTAGES OF 15 BUS SYSTEM Bus voltages in p.u. Bus No 1 2 3 4 5 6 7 8 9 10 11 12 13 14
15

7.

TABLE III COMPARISON OF VOLTAGES OF 85 BUS SYSTEM Voltages obtained by proposed method 1 0.995676 0.989259 0.981067 0.97708
0.962749

Bu sno 1 2 3 4 5
6

Voltages given in ref[6] 1 0.995678 0.989264 0.981075 0.97709

Bu sno 44 45 46 47
48

Voltages obtained by proposed method 0.878084 0.87705 0.876448 0.876345

Voltages given in ref[6] 0.87814 0.877107 0.876504 0.876402

7 8 9 10
11

0.953938 0.915787 0.913975 0.910456

0.953959 0.915825 0.914013 0.910494

49 50 51 52
53

0.874778 0.874326 0.873983 0.872321

0.874854 0.874407 0.874064 0.872395

12 13 14 15 By proposed method 1 0.971284 0.956671 0.950907 0.94992 0.958232 0.956009 0.956955 0.967971 0.966898 0.949954 0.94583 0.944519 0.94861
0.948441 16

0.905465 0.904482 0.904184 0.904006

0.905503 0.90452 0.904222 0.904044

54 55 56 57
58

0.871348 0.871909 0.874675 0.911481

0.871422 0.871983 0.874751 0.91152

Given by Reference[2] 1.0000 0.97128 0.95667 0.95090 0.94992 0.95823 0.95601 0.95695 0.96797 0.96690 0.94995 0.94583 0.94452 0.94861
0.94844

17 18 19 20 21 22 23 24 25 26 27 28 29
30

0.988653 0.974363 0.972684 0.972101 0.971401 0.970739 0.972561 0.953542 0.910232 0.905959 0.900353 0.897752 0.89299
0.888598

0.988657 0.974374 0.972695 0.972112 0.971412 0.97075 0.972572 0.953562 0.910275 0.906007 0.900408 0.89781 0.893019
0.888635

59 60 61 62 63 64 65 66 67 68 69 70 71
72

0.904459 0.900404 0.899337 0.898603 0.899415 0.895566 0.895399 0.895265 0.893579 0.890702 0.888571 0.88802 0.887765
0.893446

0.904499 0.900444 0.899378 0.898643 0.899451 0.895602 0.895435 0.895301 0.893616 0.890738 0.888607 0.888057 0.887801
0.893482

31 32

0.886531 0.885241

0.886571 0.885283

73 74

0.889334 0.889132

0.889309 0.889106

Department of Electrical Engineering, Univ. College of Engg., Osmania University, Hyderabad, A.P, INDIA.

16th NATIONAL POWER SYSTEMS CONFERENCE, 15th-17th DECEMBER, 2010

179

33

0.884206

0.884251

75

0.888866

0.888841

34 35 36 37
38

0.879696 0.877215 0.877129 0.905694

0.879752 0.87728 0.877194 0.905742

76 77 78 79
80

0.887615 0.895233 0.909995 0.893326

0.887652 0.895418 0.910033 0.893362

5.

6.

39 40 41 42
43

0.892587 0.884599 0.883658 0.88353

0.892617 0.884642 0.8837 0.883572

81 82 83 84
85

0.903436 0.902776 0.902735 0.902552

0.903974 0.903408 0.902773 0.90259

and transmission netowkrs. IEEE Trnaas. On Power Systes, Vol 3., pp 753-762. Baran, M.E. and Wu, F.F. Optimal Sizing of capacitors placed on a radial distribution system, IEEE Transactions, Vol PWRD-2, 1989, pp735-743. V.V.K. Reddy and M. Sydulu, Index and GA based optimal location and sizing of distribution system capacitors, Proceedings of IEEE Winter Meeting, 2007. BIOGRAPHIES

IV. CONCLUSIONS In this paper, load flow solution is obtained for radial distribution system using a new method. Network topology is used to build two matrices for obtaining load flow solution of radial distribution system. Two new network matrices are developed to obtain the two parts of the effective load. bus injection to Node power matrix is used to identify major part of the effective load at each node and Line loss to node power matrix is used to calculate the other part of the effective load at each node. This method is very efficient and requires very less computer memory. The proposed algorithm is applied to 15bus, and 85 bus systems used in the literature and the results are compared with those given in the research papers. Results are accurate up to the fourth decimal. REFERENCES
1. Kresting, W.H and Mendive, D L An application of ladder network theory to the solution of three phase radial load flow problem, IEEE PES Winter meeting, New York, January 1976, paper No. A 76 044-8. D. Das, D.P. Kothari, A Kalam Simple and effcient method for load flow Solution of radial distribution networks., Journal of Electrical Power and Energy Systems, Vol 17, No.5, pp335 346,1995. Jen-Hao Teng, A Network- Topology based Three-Phase Load flow for Distribution systems, Proceedings of the National Science Council, ROC(A), Vol.24, No.4, 2000, Pp. 259-264. Shirmohammadi D., H.W. Hong, A. Semlyen, and G.X. Luo 1988, A compensation based power flow method for weakly meshed distribution

K. V. S .Ramachandra Murthy did his graduation in Electrical Engineering and M. Tech in Power Systems from R.I.T., Jamshedpur in 1994 and 2002 respectively. He is pursuing his Ph.D. from J.N.T.U., Hyderabad, India. He had four years of industrial experience and eight years of Teaching Experience. He is currently employed as Associate Professor in Electrical Engineering Department in G. V. P. College of Engineering, Visakhapatnam, India. His research interests are in the Management of Electrical Energy. M. Ramalinga Raju did his B. Tech in Electrical Engineering from J.N.T.U. College of Engineering, Kakinada in 1986 and M. Tech in Electrical Machines and Industrial Drives from Regional Engineering College, Warangal, India in 1989. He did his Ph.D. from J.N.T.U., Hyderabad, India in the year 2004. Presently he is Professor & Head ' Department of Electrical and Electronics Engineering, University College of Engineering, Kakinada ,JNTU Kakinada, India. His areas of interest are Energy Management, Power Electronics. G. Govinda Rao was born in Srikakulam, Andhra Pradesh, India on January 19th, 1939. He graduated in Electrical Engineering from the College of Engineering, Kakinada, affiliated to Andhra University in 1959. He obtained his M.E. Degree from Andhra University in 1966 and Ph.D. from I.I.Sc, Bangalore in 1981. He had his teaching career at Andhra University college of Engineering for four decades and is currently employed in Gayatri Vidya Parishad College of Engineering, Visakhapatnam, India. Govinda Rao received the Best Teacher Award from the A.P. State Govt. in 1996 and is a Felllow of Institute of Engineers, India and a Member of the Indian Society of Technical Education. His research interests are in the areas of Fundamentals of Circuit Analysis and Power System Analysis and Stability. K. Narasimha Rao has obtained degree in Electrical Engineering from the Institute of Engineers (INDIA), Kolkata in summer 1996 and Post Graduation M. Tech. in IT Power Engineering from Jawaharlal Nehru Technological University, Hyderabad in December, 2002 and Ph. D. from JNT University, Hyderabad, India in the year 2009. He is having 14 years of experience in teaching and presently working as Professor with GVP College of Engineering, Visakhapatnam, AP, India. He is a Life member of ISTE, New Delhi and Member of Institute of Engineers (India), Kolkata. His research of interests includes Power System Security, Application of Neural Networks & Fuzzy in power system.

2.

3.

4.

Department of Electrical Engineering, Univ. College of Engg., Osmania University, Hyderabad, A.P, INDIA.