Sie sind auf Seite 1von 9

[INTERNATIONAL JOURNAL FOR RESEARCH &

DEVELOPMENT IN TECHNOLOGY]

Volume-4,Issue-3, Sep - 2015


ISSN (O) :- 2349-3585

Improved Distribution Loss Allocation for Radial


Systems Including DGs
SK. Chanbasha 1, Prof.shyam Mohan S.palli 2
1

12

PG Student, 2Professor& H.O.D


Dept of EEE, sir C R Reddy College of Engineering Eluru.

Abstract:-Distribution of system losses is an integral part of

their generation. In the final step, normalization is executed.

electricity pricing. This aspect can play an important role in

The application of the proposed method is illustrated on two

the operation of electricity markets. Market participants

distribution feeders, and the results are com-pared with other

restructure and privatise the network to maximize their

methods.

profits. However,

the new players pose certain problems

and challenges. One of it is the issue of power loss

Index TermsDistributed generation, loss allocation, radial

allocation. Evaluation and allocation of the system losses to

distribution systems.

suppliers and consumers are key issues to be addressed, in

I. INTRODUCTION

order to set up appropriate economic penalties or rewards for

The increase in deployment of distributed generation (DG)

suppliers and consumers this allocation process called loss

and the shift of distribution loads from customer mode to

allocation. To date, despite the existence of many loss

prosumers have altered distribution systems from passive to

allocation methods, no one method is commonly used in

active mode. As a result, some of the transmission net-works

established electricity markets. Furthermore, some markets

issues have been generalized to distribution systems as well

are still considering using different methods that will provide

[1]. One of these issues is loss allocation (LA), which

more efficient treatment of losses and aid in improving

specifies the fraction of total distribution loss that each load or

market operations and structures. The liberalization of the

DG is responsible for. Although there are many transmission

electricity market and the introduction of distributed

LA methods in the literature, distribution LA is still a new

generation (DG)

topic and most of the distribution system operators still do not

have

increased

the importance

of

distribution loss allocation (LA).

have a standard policy.

In this paper a new method for distribution power LA in


radial systems is presented. The proposed method, which is

Most of the methods implemented for distribution LA, have

based on the results of power flow and considers active and

been mainly proposed for transmission LA, a detailed survey

reactive power flows of lines for LA, is composed of three

of literature is made and listed below:

steps. In the first step the loss allocation to the loads that are
connected to the node is obtained after calculating the loss

[1]

pro rata [2], which allocates the distribution loss to

allocation to nodes starting from source node. In the next

DGs and loads based on their active power levels,

step, the total power loss is allocated to the nodes in order to

neglecting their location, which is not fair;

calculate the power loss allocated to the DGs based on the

[2]

marginal method, which calculates the marginal loss

results of the first step. In contrast to the previous step, in

coefficients (i.e., the changes in total loss due to a change

this step, allocating power losses to the nodes starts from

in active/reactive node injection), based on the results of

sink nodes, which are the nodes whose load is more than

power flow; these coefficients are then used to obtain the

www.ijrdt.org | copyright 2014, All Rights Reserved.

73

Volume-4,Issue-3, Sep-2015
ISSN (O) :- 2349-3585

Paper Title:- Improved Distribution Loss Allocation for Radial Systems Including DGs

share of DGs and loads in total loss; the results of this

meanwhile, the BCDM [8], categorized as circuit-

method needs reconciliation in order to compensate for

based methods, is a version of tracing methods used

over-recovery of loss [3][5];

for distribution LA; the method presented by Costa

[3]

direct loss coefficients method, presented in [3],

and Matos in [11] is another version of tracing

which finds a direct relation between the losses and nodal

methods, which implemented the quadratic LA

injections; both this method and marginal method are

technique; the power summation algorithm proposed

based on the results of NewtonRaphson power flow and,

in [6] is a tracing method, in which the active and

hence, have the flaws of application of this type of power

reactive power of the receiving end of each branch

flow in specific distribution systems where the number of

are decomposed to the nodal injection of the system

nodes is large, the lines resistance is negligible to their

nodes and the losses of downstream branches that are

reactance, or consists of very long or very short lines [6];

connected to the branch;

[4]

substitution method, where the responsibility of a

the method presented in [12], which applies graph

participant is calculated by subtracting the total loss when

theory for LA; however, the method is not applicable

the participant is not attached to the system from the loss

when DGs exist in the distribution systems;

when it is attached; this method is proved to give unfair

results [3];
[5]

Circuit-based methods, which contain a group of

methods as follows.

the exact method for real power LA [13], in which


the consumers have to pay for losses;

The method presented in [14], which allocates the


losses to the consumers of a radial distribution

1) Z-bus method [7], which is not applicable to distri-

system in a quadratic way, based on identifying the

bution systems containing only overhead lines, since

real and imaginary parts of the current of each branch

the Y-bus matrix is singular for such systems, due to the

current of each branch

fact that the shunt admittance of such lines is Negligible


[8].
2) The method based on modified bus admittance
matrix [9].
3) Succinct method [10], which considers active and re-

Reference [15] and [16] present a comparative study of


distribution LA methods. The following points should be
considered in distribution LA:
1) The slack node for distribution systems is always the

active flows for LA and is proved by Carpaneto et al. to

node

connecting

transmission

and

distribution

be incapable of providing reliable results under

systems; how-ever, in transmission LA, there are

particular circumstances [8].

many alternatives for the slack node.

4) Branch current decomposition method (BCDM) [8], in

2) Unlike the transmission LA methods, in which a

which the loss allocated to each node is calculated

fraction of loss may be allocated to the slack node, in

based on the current of its upward Branches

distribution LA methods, no loss is assigned to the

tracing methods, which are based on attributing the

slack node.

branch power flows to the nodes injection powers;

3) The methods used for transmission LA could be used

although a considerable amount of transmission LA

in distribution systems; however, the loss allocated to

literature is dedicated to these methods, the

the slack node in these methods should be

terminology seems to disappear in distribution LA;

redistributed among other nodes in proportion to the

this is due to the fact that in radial distribution

nodes currents [8]. This point is really important, as

systems, each branch current can be easily written in

the slack node current in distribution systems is

terms of the current of its downward nodes;

usually large and it allocates a high proportion of

www.ijrdt.org | copyright 2014, All Rights Reserved.

(74)

Volume-4,Issue-3, Sep-2015
ISSN (O) :- 2349-3585

Paper Title:- Improved Distribution Loss Allocation for Radial Systems Including DGs

total losses to itself, using the transmission LA


methods.
4) It is implicitly assumed that the loads and DGs have
bilateral contracts with the distribution company [17].

This paper proposes an LA method that can be applied to

Fig. 2. Sample feeder.

radial medium voltage distribution systems with DGs. The

In case (PG1+PG2+ ) > (PD1+PD2+), the

method starts by assigning zero power losses to a specific

allocates zero losses to the DGs connected to the node.

group of nodes. Then, the power loss allocated to other nodes

method

3) The proposed method does not allocate negative

is calculated based on the power loss of the lines connecting

losses to the loads and DGs.

the zero assigned nodes and these nodes. Since this method

4) The distribution system is assumed to be a radial

results in over-recovery of total loss, normalization is executed

system, in which the loads and DGs have private

at the end to compensate. The method is simple and is based

owners.

on the results of power flow.

5) Consider the circuit depicted in Fig. 2, which shows

This paper is organized as follows: the next section explains

two nodes of a system. The power loss of the line

the bases of the method, which is proceeded by the

connecting nodes 1 and 2 can be written as

formulation of the method presented in Section III. In Section


IV, the proposed method and five other methods are applied to
PLoss1, 2 = r1, 2

a rural distribution feeder in order to compare the results.

=k(

),

(1)

Finally, the last section presents the concluding remarks.


Where r1, 2 is the resistance of the line; P1,2 and Q1,2 are the
II.PROPOSED METHOD ASSUMPTIONS AND BASES

active and reactive power through the line; and V1 is the


voltage of node 1. As it is seen, P Loss1, 2 is composed of two

1) The distribution loss allocated to the node connecting

terms. The first term, which is k

, is due to the active flows

the distribution and transmission network is set to

through the line and the second term, which is k

zero.

the reactive flows. Let us denote these two terms, respectively,

2) Consider the node depicted in Fig. 1.

as

and

, is due to

. Consequently, (1) can be written as

In case (PG1+PG2+ ) > (PD1+PD2+), the


proposed method allocates zero losses to all loads
connected to this node, since it means that the loads
are locally fed by the DGs and, hence, do not result in

PLoss1, 2 =

(2)

As P1,2 = PD21+PD22+PLoss1,2,PLoss1,2 can be written as


=k

=k (PD21+PD22+PLoss1,2)2

(3)

any power loss.


Which is equal to k (PD21+ PD22)2, since PLoss1,2 is usually small
compared to PD21 and PD22 and , hence, can be ignored.
Based on the Shapley value, which is for calculating the
Fig. 1. Sample node of a distribution system.

contribution of a player in a game played by a number of


players, the contribution of load PD21 in
K(

www.ijrdt.org | copyright 2014, All Rights Reserved.

PD21PD22)

is equal to
(4)

(75)

Volume-4,Issue-3, Sep-2015
ISSN (O) :- 2349-3585
And that of load PD22 is K (

Paper Title:- Improved Distribution Loss Allocation for Radial Systems Including DGs

their active generation;

PD21PD22). A similar

formulation can be derived, in case the number of loads or

ii. Assigning zero loss to the active sink nodes;

lines connected to node 2 increases.

iii. Calculating the loss assigned to other nodes


due to active flows;

Moreover, the same approach is applicable if we want to


allocate

to QD21 and QD22.


iv.

Calculating the loss allocated to the DGs


due to active flows.

III.PROPOSED LA METHOD
b) Loss due to reactive flows:
i.

The method is composed of three steps as follows.

are the nodes whose reactive demand

1) Calculating the loss allocated to the loads:


a)

exceeds their reactive generation;

Loss due to active flows:


i.

ii.

Specifying the active source nodes, which

iii.

exceeds their active demand;

Calculating the loss assigned to other nodes


due to reactive flows;

Assigning zero active loss to the active


iv.

source nodes;
iii.

Assigning zero loss to the reactive sink


nodes;

are the nodes whose active generation

ii.

Determining the reactive sink nodes, which

Calculating the loss allocated to the DGs


due to reactive flows.

Calculating the loss assigned to nodes other


than the source nodes, due to active flows;

iv.

Calculating the loss allocated to the loads


due to active flows.

c)

Total loss: by summing up the loss allocated to DGs


due to active and reactive flows.

b) Loss due to reactive flows:


i.

Specifying the reactive source nodes, which


are the nodes whose reactive generation
exceeds their reactive demand;

ii.

3) Normalization so that the losses allocated to the loads


and DGs add up to the total active loss.

Assigning zero loss to the reactive source


nodes;

iii.

iv.

Calculating the loss allocated to other nodes

The loss allocated to each node when calculating the loss

due to reactive flows;

allocated to the loads is different from the values obtained

Calculating the loss allocated to the loads

when calculating the loss allocated to DGs. In any of the first

due to reactive flows.

and second steps, the loss allocated to each node is calculated


based on the loss allocated to its adjacent nodes.

c) Total loss: by summing up the loss allocated to loads


due to active and reactive flows.
2) Calculating the loss allocated to the DGs:
a)

Loss due to active flows:


i.

Determining the active sink nodes, which


are the nodes whose active demand exceeds

www.ijrdt.org | copyright 2014, All Rights Reserved.

(76)

Volume-4,Issue-3, Sep-2015
ISSN (O) :- 2349-3585

Paper Title:- Improved Distribution Loss Allocation for Radial Systems Including DGs

Fig. 5. Steps of the proposed LA method.


Load

node no.

(kW)

(KVAR)

LA results in kW
Pro rata

Marginal

z- bus

method

method

method

BCDM

Succinct

proposed

method

method

89

50

0.16

0.34

0.22

0.33

0.22

0.032

111

63

0.20

0.49

0.29

0.43

0.29

0.11

140

80

0.25

0.56

0.43

0.62

0.44

0.31

141

80

0.25

0.52

0.78

0.96

0.79

0.56

338

192

0.60

2.08

2.12

2.56

2.15

3.56

89

50

0.16

0.49

0.30

0.41

0.30

1.1

11

152

86

0.27

0.37

0.77

0.97

0.79

0.29

12

266

151

0.48

-0.30

1.36

1.70

1.42

0.46

13

10

0.02

-0.03

0.05

0.06

0.05

0.023

15

205

116

0.37

-1.98

-0.25

-0.25

-0.15

16

72

41

0.13

-0.80

-0.09

-0.09

-0.06

17

241

137

0.43

-2.14

-0.51

-0.51

-0.18

Subtotal

3.31

-0.41

5.47

10.20

6.05

6.4407

AND LA RESULT
IV. CASE STUDY

www.ijrdt.org | copyright 2014, All Rights Reserved.

(77)

Volume-4,Issue-3, Sep-2015
ISSN (O) :- 2349-3585

Paper Title:- Improved Distribution Loss Allocation for Radial Systems Including DGs

In this section, the proposed LA method is implemented on a


sample rural distribution system, whose single-line diagram is
shown in Fig. 6. This system comprises 17 nodes, 12 loads, 3
DGs, and 16 distribution lines. Nonzero shunt parameter of
lines makes it possible to use the Z -bus LA method, since the
Y -bus matrix is not singular in this case. The loads and DGs
data and the results of the proposed LA method as well as pro
rata [2], marginal [3], Z-bus [7], BCDM [8], and succinct
method [10] are provided in Table II. As Table II shows, three
DGs are located at nodes 1517. The DGs are considered as
negative loads for power-flow calculation

Fig 6: Test Distribution Feeder

The proposed method has been applied to a larger system


comprising 30 nodes, whose data can be found in [13], except
the loads and DGs output power, which are given in Table II.
The results of applying the proposed method to this system
are shown in Table II as well
Generation
node no.

P
(kW)

Q
(kVAR)

Pro rata
method

Marginal
method

Z bus
method

BCDM

Succinct
method

Proposed
method

15
16
17

300
200
260

145.29
96.86
125.92

1.31
0.87
1.13

2.79
2.10
2.15

0.36
0.24
0.55

-1.53
-1.08
-0.78

0.22
0.17
0.19

0.035
0.017
0.0017

3.31

7.04

1.15

-3.40

0.58

0.0537

Subtotal

www.ijrdt.org | copyright 2014, All Rights Reserved.

(78)

Volume-4,Issue-3, Sep-2015
ISSN (O) :- 2349-3585

Paper Title:- Improved Distribution Loss Allocation for Radial Systems Including DGs

the users demand/generation. This can be seen as the loss


allocated to node 8 is large.
The marginal method involves calculating the Jacobian
matrix, which is an extra calculation burden needed for
The following comments could be mentioned about
different LA methods as follows.

this LA method. Likewise, the Z-bus method is based on


calculating the Z-bus matrix, which is time consuming

As (7) shows, the loss allocated to each node due to active

for huge distribution systems. In contrast, the proposed

flows is the summation of two terms: 1) a fraction of the loss

method does not need any further calculation except the

allocated to the nodes that send active power to the node, and

results of power flow.

2) the loss of the lines connecting the node to the nodes that

The methods, such as Z-bus and succinct, which were

send active power to it. As a result, one can conclude that the

originally derived for transmission LA, allocate a fraction

loads located at the end of long feeders usually allocate high

of total loss to the slack bus. Since the loss allocated to

losses. In other words, one of the

the slack bus should be zero in distribution LA, the


results of such methods need to be modified, in order to

advantages of the proposed method are considering system

distribute the loss allocated to the slack bus among other

topology in LAThe results of the proposed method not only

nodes [8].In contrast, since the proposed method was

depend on the users location, but the method also considers

originally for radial distribution systems, it does not need


such modifications.

www.ijrdt.org | copyright 2014, All Rights Reserved.

(79)

Volume-4,Issue-3, Sep-2015
ISSN (O) :- 2349-3585

Paper Title:- Improved Distribution Loss Allocation for Radial Systems Including DGs

Gen., Transm. Distrib., Jan. 2000, vol. 147, no. 1, pp.

CONCLUSION
This paper presents a novel LA method for radial
distribution systems, in which the loss allocated to each

714.
4.

E. Carpaneto, G. Chicco, and J. S. Akilimali, Loss

node is dependent on the loss allocated to its adjacent

partitioning and loss allocation in three-phase radial

nodes and the loss of the lines connected to the node. The

distribution systems with distributed generation,

proposed method has the following properties, which are

IEEE Trans. Power Syst., vol. 23, no. 3, pp.1039

explained in [19] to be the desirable properties of every

1049, Aug. 2008.

LA method:

5.

E. Carpaneto, G. Chicco, and J. S. Akilimali,

The method is consistent with the results of power

Computational

flow.

allocation methods for distribution systems with

The losses allocated to the loads/DGs depend on the

distributed generation, in Proc. IEEE Melecon,

amount of energy they consume/produce.

Benalmdena (Mlaga), Spain, 2006, pp. 10281031.

The location of each load and DG is a key factor in

6.

aspects

of

the

marginal

loss

M. Atanasovski and R. Taleski, Power summation

the amount of loss allocated to them.

method for loss allocation in radial distribution

The method is easy to understand.

networks with DG, IEEE Trans. Power Syst., vol.

The implementation of the method is straightforward

26, no. 4, pp. 24912499, Nov. 2011.

and does not need complicated programming or

7.

A. J. Conejo, F. D. Galiana, and I. Kochar, Z-bus


loss allocation,IEEE Trans. Power Syst., vol. 16, no.

extensive computational effort.

1, pp. 105110, Feb. 2001.


In order to allocate energy losses throughout a day, the

8.

E. Carpaneto, G. Chicco, and J. S. Akilimali,

method must be executed separately for each hour, which is

Branch current decomposition method for loss

time-consuming. Hence, the authors are working on a

allocation

stochastic method, which could find equivalent loads based on

distributed generation, IEEE Trans. Power Syst.,

their variation during a particular time span with an equal

vol. 21, no. 3, pp. 11701179, Aug. 2006

energy loss effect to replace the value of loads.

9.

in

radial

distribution

systems

with

R. S. Daniel, R. S. Salgado, and M. R. Irving,


Transmission loss allocation through a modified

REFERENCES

Ybus, in Proc. Inst. Elect. Eng., Gen., Transm.


Distrib., Mar. 2005, vol. 152, no. 2, pp. 208214.

1.

2.

3.

P. M. Sotkiewicz and J. M. Vignolo, Nodal pricing

10. W. L. Fang and H. W. Ngan, Succinct method for

for distribution networks: Efficient pricing for

allocation of networklosses, in Proc. Inst. Elect.

efficiency enhancing DG, IEEE Trans.Power Syst.,

Eng., Gen., Transm. Distrib., Mar.2002, vol. 149, no.

vol. 21, no. 2, pp. 10131014, May 2006.

2, pp. 171174.

M. Ilic, F. Galiana, and L. Fink, Power Systems

11. P.M. Costa andM. A.Matos, Loss allocation in

Restructuring: Engineering and Economics. Norwell,

distribution networks with embedded generation,

MA: Kluwer, 1998.

IEEE Trans. Power Syst., vol. 19, no. 1, pp. 384389,

Mutale, G. Strbac, S. Curcic, and N. Jenkins,

Feb. 2004.

Allocation of losses in distribution systems with

12. C. N. Macqueen and M. R. Irving, An algorithm for

embedded generation, in Proc. Inst. Elect. Eng.,

the allocation of distribution system demand and


energy losses, IEEE Trans. Power Syst., vol. 11, no.
1, pp. 338343, Feb. 1996.

www.ijrdt.org | copyright 2014, All Rights Reserved.

(80)

Volume-4,Issue-3, Sep-2015
ISSN (O) :- 2349-3585

Paper Title:- Improved Distribution Loss Allocation for Radial Systems Including DGs

13. J. S. Savier and D. Das, An exact method for loss


allocation in radial distribution systems, Int. J. Elect.
Power Energy Syst., vol. 36, no. 1, pp. 100106,
2012.
14. J. S. Savier and D. Das, Loss allocation to
consumers before and after reconfiguration of radial
distribution networks, Int. J. Elect. Power Energy

15. P. M. De Oliveira-De Jess, M. T. Ponce de Leo, J.


M. Yusta-Loyo, and H. M. Khodr, Cost loss
in

distribution

networks

electronics engineering; A.K.R.G College of engineering,


nallajerla affiliated to JNTUK Kakinada in 2013.he is

Syst., vol. 33, no. 3, pp. 540549, Mar. 2011.

allocation

SK.Chanbasha received his BTECH degree in electrical and

with

currently pursuing M.E degree in power systems automation


from the Sir C.R.Reddy college of engineering, Eluru.

high

penetration of distributed renewable generationA


comparative study, presented at the Int. Conf.
Renew. Energy Power Qual., Zaragoza, Spain, Mar.
2005.
16. J. S. Savier and D. Das, Energy loss allocation in
radial distribution systems: A comparison of practical
algorithms, IEEE Trans. Power Del., vol. 24, no. 1,

20. Prof Shyam Mohan S Palli received his B.Tech degree in in


Electrical and Electronics Engineering from JNTU Kakinada

pp. 260267, Jan. 2009.


17. E. Carpaneto, G. Chicco, and J. S. Akilimali,

in 1978; M.Tech degree in power systems from same institute

Characterization of the loss allocation techniques for

in 1980. He joined in teaching profession in 1981. He has put

radial systems with distributed generation,Elect.

up an experience of 35 years in teaching as lecturer, Asst.

Power Syst. Res., vol. 78, no. 8, pp. 13961406, Aug.

Professor, Professor. Presently he is working as Professor and

2008.

HOD, EEE of Sir C.R. Reddy college of Engineering, Eluru.

18. J. S. Savier and D. Das, Impact of network

He is author of Circuits and Networks, Electrical Machines

radial

Published by TMG New Delhi. He published many papers in

distribution systems, IEEE Trans. Power Del., vol.

referred journals and conferences. His areas of interests

22, no. 4, pp. 24732480, Oct. 2007.

include Machines and power system

reconfiguration

on

loss

allocation

of

19. A. J.Conejo, J. M.Arroyo, N. Alguacil, andA.


L.Guijarro,

Transmission

loss

allocation:

comparison of different practical algorithms,IEEE


Trans. Power Syst., vol. 17, no. 3, pp. 571576, Aug.
2002

www.ijrdt.org | copyright 2014, All Rights Reserved.

(81)