1

A Simplified Approach for Load Flow Analysis of Radial Distribution Network with Embedded Generation
K Vinoth Kumar, M P Selvan
Vn Sn I Ln Ib I cb
= Voltage at the node n = Complex load power at the node n = Load current injections at the node n = Branch current of the branch b = Line charging current of the branch b = Branch current in the l th lateral line
Abstract This paper presents a simple approach for load flow analysis of a radial distribution network with embedded generation. The proposed approach utilizes forward and backward sweep algorithm based on Kirchoffs current law (KCL) and Kirchoffs voltage law (KVL) for evaluating the node voltages iteratively. In this approach, computation of branch current depends only on the current injected at the neighbouring node and the current in the adjacent branch. This approach starts from the end nodes of sub lateral line, lateral line and main line and moves towards the root node during branch current computation. The node voltage evaluation begins from the root node and moves towards the nodes located at the far end of the main, lateral and sub lateral lines. The proposed approach has been tested using three radial distribution systems of different size and configuration and is found to be computationally efficient. The effectiveness of the proposed approach is further demonstrated by integrating the embedded generation into the load flow analysis of the radial distribution network.

I bl ,l

I bsl , sl = Branch current in the sl th sub lateral line I bm,m = Branch current in the mth minor line nMl nlsl nslm bMl blsl bslm snll
= Node number in the main line from which this l th lateral line begins = Node number in the lateral line from which this sl th sub lateral begins = Node number in the sub lateral line from which this

Index Termsembedded generation, forwardbackward

sweep, load flow analysis, radial distribution system List of symbols

mth minor begins = Branch number connecting the main line with the l th lateral line = Branch number connecting the lateral line with the sl th sub lateral line = Branch number connecting the sub lateral line with
the mth minor line = Starting node number of l th lateral line

= Total number of nodes in the given radial distribution network Nb = Total number of branches in the given radial distribution network Nl = Total number of lateral lines in the given radial distribution network N sl = Total number of sub lateral lines in the radial distribution network Nm = Total number of minor lines in the given radial distribution network enM = Ending node number in the main line

Nn

snslsl = Starting node number of sl th sub lateral line snmm = Starting node number of mth minor line enll
= Ending node number of l th lateral line

enslsl = Ending node number of sl th sub lateral line enmm = Ending node number of mth minor line l sl m A
= denotes a lateral line, l = 1,2,3, N l = denotes a sub lateral line, sl =1,2,3.... N sl = denotes a minor line, m = 1,2,3. N m = Swept area by the rotor = Density of air = Wind speed = Power coefficient = tip speed ratio = pitch angle = Turbine rotor speed in rpm Gear ratio

Zb
Rb Xb yb

= Impedance of the branch b = Resistance of the branch b = Reactance of the branch b = Line charging admittance of the branch b

uw Cp
K. Vinoth Kumar is a research scholar in the department of electrical and electronics engineering, National Institute Technology, Tiruchirapalli, Tamil nadu, India 620 015. (e-mail: k_vinothkumar75@yahoo.com, 407107003@nitt.edu) M.P.Selvan is with the department of electrical and electronics engineering, National Institute Technology, Tiruchirapalli, Tamilnadu, India 620 015. (corresponding author :phone: 91-431-2503262; fax: 91-4312500133; e-mail: selvanmp@ nitt.edu)

Nr =

R = Rotor radius at the turbine blades

2 I. INTRODUCTION Though the conventional load flow methods like Newtons method and fast-decoupled method are simple, due to the radial nature and high reserves and the installation and running costs of electricity generation units attracted the researchers to focus on the selection of resources used for embedded generation. Renewable energy sources were found to be the cost effective solution for embedded generation. Such embedded generations in the distribution system are also called as dispersed generations or distributed generations (DG). Renewable energy sources like solar, wind, geothermal, wave, tidal, biomass, etc., are the various possible solutions for distributed generation of electrical energy. Based on the technology adopted, the installation cost, running cost, green house gas emission and also the reactive power requirement of the distributed generations will differ. Wind being an emission free and viable technology, it is popular in India and most parts of the world. The operation of wind turbine generator system (WTGS) that commonly employs induction generator is completely different from other technologies employed for electric power conversion in distributed generation. The induction generators require reactive power support for their operation. This distinct behaviour of WTGS makes the problem complicated compared to other DG technologies, which are simply represented as negative loads. Therefore WTGS is considered in this paper as an embedded generation and are integrated into the load flow analysis of radial distribution networks as a function of their node voltage. This paper is organized as follows: Section-II presents the possible configuration of the radial distribution network. Section-III describes the proposed approach for load flow analysis of radial distribution network and provides the flow chart of the proposed approach. In Section-IV, modeling of WTGS for load flow study is presented. SectionV validates the proposed approach and provides the test results. II. RADIAL DISTRIBUTION NETWORK A typical radial distribution network consisting of root node, main line, lateral line, sub lateral line and minor line is shown in Fig.1.

R ratio of the distribution lines, they X

cannot be effectively used for the load flow analysis of radial distribution systems. Few researchers have modified the Newton method and fast decoupled method to suit the nature of the distribution network [1][3]. They are neither computationally efficient nor convergent for ill-conditioned systems. A compensation-based technique has been proposed in [4]. This technique requires adoption of a methodology for numbering every branch and also the current in any branch is computed as the sum of load current injections at all nodes located beyond the branch under consideration. This is true for all branches irrespective of their location in the network. Hence it is quite evident that repetitive mathematical operations are required and thus the compensation-based technique needs longer computational time. Moreover, a number of attempts have been made using ladder network theory for load flow analysis of a radial network [5],[6]. In [7], Stevens et al have shown that this ladder network theory is computationally fast but did not converge in five out of twelve cases studied. In [8], [9] a different approach for solving load flow problem involving the following two major steps has been proposed: i) Identification of all nodes located beyond each branch. ii) Calculation of branch currents and node voltages. Identification of nodes in a large system with multiple branches is tedious and takes a longer duration to determine. In addition, the shortcomings associated with [4] hold for this approach also. In this context, a novel simple approach for the forwardbackward sweep algorithm is proposed in this paper, which overcomes all the above drawbacks for balanced radial distribution network. In the proposed approach, load flow analysis of a radial network is performed by treating every lateral and sub lateral line as an individual main line. The branch current evaluation starts from the far end of each of the sub lateral, lateral and main lines and moves towards the root node. Computation of branch current depends only on the current injected at the neighbouring node and the current in the adjacent branch. This avoids repetitive computations at each branch and thus makes the approach computationally simple and efficient. Once the branch currents are determined, the node voltage evaluation begins from the root node and moves towards the nodes located at the far end of the main, lateral and sub lateral lines. This simplified approach has been used for the load flow analysis of the radial distribution network in the presence of embedded generation to bring about its merit. Due to the increasing demand of electricity, diminishing fossil fuel

Fig.1 Single Line Diagram of a Radial Distribution Network

Root node:

The node connected to the voltage regulating station/substation in the radial distribution network. Main line: Line emanating from the root node. Lateral line: Line emanating from the main line. Sub lateral line Line emanating from the lateral line Minor line: Line emanating from the sub lateral line

3 The approach proposed in this paper, assuming balanced load condition, is presented in the following section.
III.

I b = I b +1 + I L (b +1) + I cb + I bl ,l
l =1

Nl

- (9)

PROPOSED APPROACH

Where

b = 1, 2,........................(enM 1) I bl ,l = 0 if ( b + 1) nMl l l = 1, 2,....................Nl I b +1 = 0 if ( b + 1) = enM

i) Current injections at any node n can be written as,

I Ln =

S V

* n * n

- (1)

Where n = 1,2,3,. N n ii) Line charging current in any branch b can be written as,

iv) Voltage of any node n is given by, Vn = Vn 1 Ib Z b Where Vn 1 = voltage at (n-1)th node.

- (10)

1 1 I cb = ybVb + ybVb +1 2 2
Where b = 1,2,3, N b

b = ( n 1)

- (2)

Vb ,Vb +1 = node voltage of node b and (b + 1) respectively. yb = line charging admittance of branch b iii) a) Branch current in any branch b in the minor line can be written as, I b = I b +1 + I L(b +1) + I cb - (3)

Where

b = ( snmm 1) to (enmm 1) m , m = 1, 2,....................N m I b +1 = 0, if (b + 1) = enmm

I b = Current in the branch b Zb = Impedance of the branch b The approach begins with the assumption of flat voltage start at all nodes. The node current injections, line-charging current and all branch currents are evaluated using (1) to (9). The node voltages are evaluated using (10). The node voltages evaluated are compared with the previous values of node voltages. If the differences in the node voltages between successive iterations are not within the specified tolerance then the above procedure is repeated until convergence in node voltages.
The real and reactive power loss in the network is given by,
Nb

I bm, m = I b , if

b = bslm
Nm

- (4)

Real power loss, P = I b Rb

b =1

- (11)
2

b) Branch current in any branch b in the sub lateral line can be written as,

I b = I b +1 + I L (b +1) + I cb + I bm, m
m =1

Reactive power loss, Q = I b X b

b =1

Nb

- (5)

The flow chart shown in Fig.2 depicts the step-by-step procedure of the proposed approach..
START

Where

b = ( snslsl 1) to (enslsl 1) sl , sl = 1, 2,....................N sl ; I bm,m = 0 if ( b + 1) nslm m , m = 1, 2,....................N m I b +1 = 0 if ( b + 1) = enslsl

Read input data for the given radial distribution network

Assume flat voltage start at all nodes

Compute node currents and line charging currents using (1) and (2)

Compute branch currents in all the minor lines using (3) and (4)

Compute branch currents in all the sub lateral lines using (5) and (6)

I bsl , sl = Ib if b = blsl

- (6)

Compute branch currents in all the lateral lines using (7) and (8)

c) Branch current in any branch b in the lateral line can be written as, I b = I b +1 + I L (b +1) + I cb + I bsl , sl
sl =1 N sl

Compute branc h currents in the main line using (9)

Compute node voltages using (10) at all nodes except the root node

- (7)
Is difference in node voltages of succ essive iterations greater than the specified tolerance ? Yes

Where

b = ( snll 1) to (enll 1) l l = 1, 2,....................Nl ; I bsl , sl = 0 if ( b + 1) nlsl sl sl = 1, 2,....................N sl I b +1 = 0 if ( b + 1) = enll

No Compute line losses - real and reactive power losses using (11)

Print the results

STOP

I bl ,l = I b if b = bMl

- (8)

Fig.2 Flow Chart for the Proposed Approach

d) Branch current in any branch b in the main line can be written as,

4 IV. MODELING OF WTGS FOR LOAD FLOW ANALYSIS OF RADIAL DISTRIBUTION SYSTEM The WTGS is integrated in the load flow study of RDS based on the equations derived from the equivalent circuit of induction generator [10].
Mechanical power input equations:

I 2 = I1 + I m and I m =
Eqn. (5) can be written as,

V + I1 ( R1 + jX l1 ) jX m

- (15)

( R + jX l1 ) V I 2 = I1 1 + 1 + jX m jX m
I1 = R2 s + jX l 2 ( jX m ) + ( R1 + jX l1 ) R 2 + j(X + X ) l2 m s

- (16)

The mechanical power output of the wind turbine (Pm ) is a function of turbine speed ( N r ) and wind speed ( uw ) .
Pm (u w , N r ) = 1 3 C (( , Au w p 2

From the equivalent impedance of the induction machine, - (17) V

- (12)

The tip speed ratio given in eqn. (3) depends on the wind speed (u w ) and turbine speed ( N r )
2 N r R = 60 u w

From eqn. (16) and (17)

R V 2 + j ( Xl 2 + X m ) R1 + j ( Xl1 + X m ) V s I2 = + R2 R2 jX m jX m + jXl 2 ( jX m ) + ( R1 + jXl1 ) + j ( Xl 2 + X m ) s s

- (13)

- (18)

Empirically, the power coefficient is given as,

1 - (14) C p (uw , Nr , ) = c1 c2 c3 c4 x c5 e
1 1 0.035 and Where = + 0.08 1 + 3
c6

The air gap power is obtained as,

2 1 s Pg (V , r ) = I 2 R2 s

c1 to c6 , x are

- (19)

constants.

Electrical power output equations The equivalent circuit of the induction machine is shown in Fig. 3. Xl2 Xl1 R1
I1 I2 Im Xm R 2 /s

For a given wind speed and terminal voltage, the generator rotor speed is determined by equating the mechanical power input and the developed electrical power. Once the generator rotor speed is determined, the electrical power output can be computed as described in the following steps: For a known wind speed uw , rated slip of the induction machine 1) If

srated and the terminal voltage V

uw uwcutin uw < uw no min al , s = srated uw no min al uwcutin Where uw no min al and uwcutin are the nominal and cut-in s = srated and r = s (1 s )
Compute the accelerating power

Fig.3 Induction Machine Equivalent Circuit Where I1 - Stator current

wind speeds respectively. Else

I2 - Rotor current I m - Magnetising current V - Terminal voltage R1 - Stator resistance X l1 - Stator leakage reactance R2 - Rotor resistance X l 2 - Rotor leakage reactance X m - Magnetizing reactance s r
- Slip =

2)

Pa = Pg (V , r ) Pm (u w , N r ) using (12) and (19)

3) If Pa

= 0 , go to next step Pa = 0 to

Else Iteratively solve the equation in step-2 by setting obtain the rotor speed r .

4) Compute real and reactive power output of the WTGS using the following equations

- Synchronous speed of induction generator in rpm - Rotor speed of the induction generator in rpm

Pe = Re (VI1 ) , Qe = Im (VI1 )

- (20)

The data and the characteristics of the WTGS given in [11] are used in this paper.

The rotor current I 2 is given as,

5
V.

RESULTS

The proposed approach has been implemented using MATLAB and tested on a P-IV, 3.20GHz 1MB RAM computer. The computational efficiency of the present approach has been tested using Indian 25, 33, and 69 node radial distribution networks. The data for 25-node system is given in the appendix. The data for 33 and 69 node systems are given in [9].
TABLE I NETWORK SIZE AND CONFIGURATION OF THE SYSTEM UNDER STUDY 25 node 33 node 69 node Indian system system system 14 18 27 24 3 1 32 3 0 68 7 0

compared and the details are furnished in Table III with and without the presence of embedded generation in each of the radial distribution network. Fig.1 3 shows the voltage profiles of the three sets of radial distribution network with and without the presence of wind turbines. It can be seen that the voltage profiles of these networks are improved to a greater extent with embedded generation.
1.0100

BUS VOLTAGE MAGNITUDES IN P.U.

0.9900 0.9700 0.9500 0.9300 0.9100 0.8900 0.8700 0.8500 1 5 9

Number of nodes in main lines Total number of branches Number of lateral lines Number of sub lateral lines

NODES

13

17

21

25

Original RDS

RDS with WTGS

TABLE II RATINGS OF WTGS IN MW WITH LOCATION DETAILS FOR VARIOUS RADIAL DISTRIBUTION NETWORKS 25 node Indian system Bus 14 - 19 Bus 21 - 18 Bus 25 - 15 33 node system Bus 14 - 1.00 Bus 18 - 0.75 Bus 33 - 2.50 69 node system Bus 27 - 0.50 Bus 65 - 2.50 Bus 69 - 1.25

Fig. 1 Voltage profile of 25 node Indian radial distribution network

BUS VOLTAGE MAGNITUDES IN P.U.

1.0100 1.0000 0.9900 0.9800 0.9700 0.9600 0.9500 0.9400 0.9300 0.9200 0.9100
1 5 9 13 17 NODES 21 25 29 33

TABLE - III COMPARISON OF PROPOSED APPROACH WITH [8] System and its parameter
25 node Indian system

Without WTGS Proposed approach Existing method [8] Proposed approach Existing method [8] Proposed approach Existing method [8] Proposed approach Existing method [8] Proposed approach Existing method [8] Proposed approach Existing method [8] 0.0780 0.1100 2 2 0.0470 0.1100 2 2 0.0470 0.1100 2 2

With WTGS 0.0780 0.1100 2 2

BUS VOLTAGE MAGNITUDES IN P.U.
1.0200 1.0100 1.0000 0.9900 0.9800 0.9700 0.9600 0.9500 0.9400 0.9300 0.9200 0.9100 1 5

Original RDS

RDS with WTGS

CPU time in seconds Number of iterations CPU time in seconds Number of iterations CPU time in seconds Number of iterations

Fig. 2 Voltage profile of 33 node radial distribution network

0.0470 0.1100 2 2 0.0780 0.1250 2 2

69 node

33 node

13

17

21

25

29

33

37

41

45

49

53

57

61

65

69

The size and configuration of the systems under study are presented in Table-I. The ratings of the WTGS connected to the nodes in each of the radial distribution network considered in this paper are depicted in Table-II. The rate of convergence of the proposed approach and the existing method [8] are

NODES

Original RDS

RDS with WTGS

Fig. 3 Voltage profile of 69 node radial distribution network

6
VI.

CONCLUSION

A novel approach for load flow analysis of a radial distribution network, which is simple to implement and efficient in computation has been proposed and described in detail in this paper. The computational efficiency and speed of the proposed method has been tested using Indian 25, 33 and 69 node radial distribution networks. The comparison between the proposed and existing method ensures the speed and accuracy of the proposed approach in terms of CPU time for both original radial distribution network and for the radial distribution network with embedded generation for systems of different sizes and configurations. It can be concluded that the simplification made in the branch current computation of the proposed approach has resulted in improved computational speed of load flow analysis of radial distribution network.
VII.

REFERENCES

[1] D. Rajicic, Y. Tamura, A modification to fast decoupled power flow for networks with high R/X ratios, IEEE Trans. on Power System, Vol. 3, No. 2, pp.743-746, 1988. [2] S. Iwamoto, Y. Tamura, A load flow calculation method for ill-conditioned power systems, IEEE Transactions on Power apparatus and Systems, Vol. 100, No. 4, pp.17361743, 1981. [3] S. C. Tripathy, G. Durgaprasad, O.P.Malik, G.S.Hope, Load flow solutions for ill-conditioned power systems by a Newton like method, IEEE Transactions on Power apparatus and Systems, Vol. 101, No. 10, pp.3648-3657, 1982. [4] D. Shirmohammadi, H. W. Hong, A. Semlyen, G. X. Luo, A compensation based power flow method for weakly

meshed distribution and transmission networks, IEEE Transactions on Power Systems, Vol. 3, No. 2, pp.753762, 1988. [5] W. H. Kersting, D. L. Mendive, An application of ladder network theory to the solution of three phase radial load flow problem IEEE PES winter meeting, New York, Jan. 1976. [6] W. H. Kersting, A method to teach the design and operation of a distribution system, IEEE Transactions on Power apparatus and Systems, Vol. 103, No. 7, pp.1945-1952, 1984. [7] R. A. Stevens, D. T. Rizy, S. L. Purucker, Performance of conventional power flow routines for real time distribution automation application Proceedings of 18th southeastern symposium on system theory, pp.196-200, April 1986. [8] S. Ghosh., D. Das, Method for load flow solution of radial distribution network, IEE Proc.- Generation Transmission and Distribution Vol. 146, No. 6, pp.641648, 1999. [9] Rakesh Ranjan, D. Das, Simple and efficient computer algorithm to solve radial distribution networks, Electric power components and systems, Vol. 31, No.1, pp.95107, 2003. [10] K.C. Divya., P.S. Nagendra Rao, Models for wind turbine generating systems and their application in load flow studies, Electric power systems research, Vol. 76, pp. 844 856, 2006. [11] Lubosny, Z.: Wind turbine operation in electric power systems, in Springer-verlag Berlin Heidelberg, New York, 1st edition, pp.192, 2003.

VIII APPENDIX DATA FOR 25 NODE INDIAN SYSTEM

Branch Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Branch Data From Bus To Bus Number Number 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 11 15 15 16 16 17 8 18 18 19 19 20 20 21 4 22 22 23 20 24 24 25 Line Resistance in p.u. 0.0225 0.0375 0.0337 0.0487 0.0187 0.0187 0.0375 0.0075 0.0187 0.0750 0.0112 0.0750 0.0562 0.0375 0.0375 0.0375 0.0562 0.0281 0.0056 0.0262 0.0037 0.0225 0.0938 0.0131 Line Reactance in p.u. 0.0182 0.0304 0.0274 0.0396 0.0152 0.0152 0.0304 0.0060 0.0152 0.0608 0.0091 0.0608 0.0457 0.0304 0.0304 0.0304 0.0457 0.0228 0.0045 0.0212 0.0030 0.0182 0.0762 0.0106 Bus Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Bus Data Real Power Demand in p.u. 0.0000 0.0180 0.0031 0.0000 0.0095 0.0302 0.0116 0.0000 0.0235 0.0105 0.0000 0.0280 0.0086 0.0152 0.0081 0.0126 0.0041 0.0117 0.0306 0.0000 0.0498 0.0136 0.0185 0.0111 0.0324 Reactive Power Demand in p.u. 0.0000 0.0066 0.0011 0.0000 0.0034 0.0032 0.0034 0.0000 0.0063 0.0045 0.0000 0.0091 0.0012 0.0016 0.0017 0.0037 0.0005 0.0045 0.0098 0.0000 0.0101 0.0043 0.0032 0.0027 0.0076

On a base of 100 MVA and 11kV