International Journal of Electrical and Computer Engineering (IJECE)

Vol. 6, No. 2, April 2016, pp. 439~446

ISSN: 2088-8708, DOI: 10.11591/ijece.v6i2.9182  439

Economic Valuation of Power and Energy Losses in

Distribution Networks

Smajo Bisanovic*, Mersiha Samardzic**, Damir Aganovic*

* Public Enterprise Elektroprivreda of Bosnia and Herzegovina d.d. – Sarajevo, Bosnia and Herzegovina
** Faculty of Electrical Engineering, University of Sarajevo, Bosnia and Herzegovina

Article Info ABSTRACT

Article history: This paper presents a framework for determining the price of power and
energy at each node in distribution network as well as the price of energy
Received Oct 11, 2015 losses in their elements. The proposed framework is based on the concept of
Revised Dec 6, 2015 the radial structure network and gives one approach to solving the pricing
Accepted Dec 26, 2015 problem that is based on purchase price of power and energy at the network
supply point. In this way it is possible to determine the economic value of
energy losses whether in the network as a whole or in particular voltage
Keyword: levels. The model has been successfully tested and results from test studies
are reported.
Distribution network
Power and energy losses
Price at the network node
Price of energy losses
Copyright © 2016 Institute of Advanced Engineering and Science.
All rights reserved.

Corresponding Author:
Smajo Bisanovic,
Public Enterprise Elektroprivreda of Bosnia and Herzegovina d.d. – Sarajevo,
Vilsonovo setaliste 15,
71000 Sarajevo, Bosnia and Herzegovina.
Email: [email protected]

1. INTRODUCTION
Researchers’ attention has been occupied by estimation of power and energy losses in distribution
networks long since. The greatest efforts when it comes to estimation of power and energy losses are
concentrated on energy losses assessment on a yearly basis and power losses assessment at maximum load in
the network [1]. The need for power and energy losses estimation arises from the following reasons:
optimization of expansion and further development of the distribution network, choice of optimal location
and sizing of distributed generators and compensation devices, dynamic network reconfiguration and voltage
optimization in distribution network, analysis of the network efficiency and performances, etc. In addition, in
a competitive and deregulated environment the quality of losses estimation is crucial for fair competition in
electricity markets. In today’s market, distribution utilities, suppliers, distribution network operators, as well
as consumers, expect estimation of losses with highest accuracy. Correct allocation of losses is necessary for
correct allocation of corresponding costs. The losses in distribution network must be fairly allocated among
all consumers and distributed generators. In recent literature, regarding losses costs allocation, several
methods have been proposed, such as postage stamp [1-3], MW-mile [4], circuit based and proportional
sharing [5]. Recently, there has been proposed a modified proportional sharing procedure [6] based on the
allocation of entire losses to consumers disregarding the influence of distributed generators using the basic
proportional sharing principle. Secondly, marginal procedures have been extensively proposed in order to
send efficient economical signals to the market agents. Marginal methods require a slack bus designation and
do not assign arbitrarily power losses between producers and consumers [1].
Allocation of losses costs in distribution networks is a complex problem whose importance
increased in competitive market and in networks with high penetration of distributed generators [7]. Pricing

Journal homepage: http://iaesjournal.com/online/index.php/IJECE

440  ISSN: 2088-8708

of distribution network includes the allocation of capital and operating costs to users (consumers, generators)
of the network in fair and equitable manner, taking into consideration that each user is charged for those costs
only for which they are responsible. Marginal cost pricing is the most widely accepted concept for achieving
this. By definition, the marginal cost of a good or service is the increase in the total cost of providing the
good or service as a result of a relatively small increase in the rate of output of the good or service [7]. In
order to allocate power losses in distribution networks with distributed generation, the concept of marginal
loss coefficients is introduced [8-10]. These coefficients measure the change in total active power losses
caused by marginal changes in consumption and/or generation of active and reactive power at each node in
the distribution network.
In this paper we propose a simple methodological framework that determines prices of power and
energy at each node of the distribution network, as well as the economic value of energy losses in the
network elements. Given its simplicity, proposed framework can be very useful for a quick losses costs
assessment, as a part of losses costs allocating procedures for network users and in other applications
regarding distribution network, such as optimization of development and operation.
This paper is organized as follows. Section 2 provides the mathematical formulation of the problem.
In Section 3 are presented the results from several case studies. The conclusions and point to future research
are outlined in Section 4.

2. MATHEMATICAL FORMULATION OF THE PROBLEM

In order to present mathematical model for power and energy losses valuation in distribution
networks, it is suitable to start with simple distribution network with n nodes, as illustrated in Figure 1.

1 2 3 4 5 n-2 n-1 n
2/3 4/5 n-2/n-1

Figure 1. Simple concept of distribution network with n nodes

For valuation power and energy losses according to the methodological approach that follows, it is
necessary to determine economic value of 1 kW (P) and 1 kWh (W) at each node of the distribution
network (1,2, ..., n).

2.1 Calculation of Parameter P

Economic value of 1 kW (P) at each node of the network according to Figure 1 can be determined
starting from the purchase price of 1 kW at node 1, P1, (at the network supply point) and costs for transfer
power to particular node in the network. If the purchase price P1 at node 1 is known, then the economic
value of power at the end of section 1–2, at node 2, can be determined as follows:

 P 2   P1   P12 (1)

where P12 is increment of the economic value of power that is transmitted from node 1 to node 2.
Increment P12 represents the annual costs related to amortization, maintenance and other fixed
costs for section 1–2, iP12, plus costs for power losses in this section, pP12. The annual costs iP12 for
section 1–2 are given as follows:

12 12
iP12  (2)
100 P2

where:
 12 is fixed annual costs factor for section 1–2, [%];
 I12 is purchase value of the equipment for section 1–2 (capital costs), [$];
 P2 is power transferred to node 2 at maximum load, [kW].

IJECE Vol. 6, No. 2, April 2016 : 439 – 446

IJECE ISSN: 2088-8708  441

The costs related to power losses in section 1–2 are given as follows:

 P12  P1
 pP 12  (3)
P2

where P12 is power losses in section 1–2 at maximum load, [kW].

According to equations (2) and (3), equation (1) obtains the following form:

 12 12  P12  P1
 P 2   P1   (4)
100 P2 P2

Each kW which is transferred to node 2 is charged at the cost of transmission to that node.
Economic value of 1 kW at the end of the following section, at node 3, is found in a completely analogous
way.
In the general case, for the section (n–1)–n, or for node n, it can be written:

 ( n  1)n ( n  1)n  P( n  1)n  P ( n 1)

 P n   P ( n  1)   (5)
100 Pn Pn

n  ( k  1)k ( k  1)k n  P( k  1)k  P( k  1)

 Pn   P1   100 Pk
  Pk
(6)
k 2 k 2

where:
 (k–1)k is fixed annual costs factor for section (k–1)–k, [%];
 I(k–1)k is purchase value of the equipment for the section (k–1)–k (capital costs), [$];
 Pk is power transferred to node k at maximum load, [kW];
 P(k–1) is economic value of 1 kW at node (k–1), [$/kW].

2.2 Calculation of Parameter W

Economic value of 1 kWh (W) at each node of the network according to Figure 1 can be determined
starting from the purchase price of 1 kWh at the node 1, W1, (at the network supply point) and costs of
energy losses in appropriate sections of the distribution network. As the costs associated with the
amortization, maintenance and other fixed costs as well as power losses related to the economic value of 1
kW at each node in the distribution network, the economic value of 1 kWh of electrical energy at the
appropriate nodes will affect only energy losses. The initial assumption in determining parameter Wn is the
existence of equality between the economic value of the energy accepted at the beginning of the one section
and the economic value of energy delivered in the same period at the end of this section. If the accepted
amount of energy during this period at the beginning of section 1-2 is W1, with the price W1, the economic
value of 1 kWh at the end of this section, where the delivered energy is W2, with the price W2, can be
determined according to the equation:

 W 1 W1   W 2 W2 (7)

Since

W2  W1   W12 (8)

where W12 is energy losses in section 1–2, from equation (7) follows:

W1
W 2  W1 (9)
W1   W12

Economic Valuation of Power and Energy Losses in Distribution Networks (S. Bisanovic)
442  ISSN: 2088-8708

Economic value of 1 kWh at the end of the following section, the transformation 2/3, at node 3, is
found in a completely analogous way:

W2 W1 W2
W 3  W2  W1 (10)
W2   W23 W1   W12 W2   W23

In the general case, for node n, it can be written:

n W( k  1)
 Wn   W1 W   W( k  1)k
(11)
k 2 ( k  1)

All equations for determining parameters P and W are derived for case illustrated in Figure 1 with
a directional ‘transmission’ of energy containing lines and transformers, and which are connected to each
other from the higher to lower voltages. However, these equations can be used for appropriate voltage level,
considering that all their variables are related with this voltage level (variable costs, power, energy, losses).
In this way it is possible to determine the economic value of 1 kW and 1 kWh at each node of the distribution
network.

2.3 Calculation of Parameters P and W in Radial Network with more Main Sections
Model presented by equations (6) and (11) can be generalized for any number of main sections (one
main section is illustrated by simplified scheme in Figure 1 to determine the value of the parameters P and
W at each node of the network). The simplified scheme with N main sections is illustrated in Figure 2. The
parameters P and W for any section (j = 1, ..., N) can be determined as follows:

n  (( kj )1)k (( kj )1)k n  P((kj ) 1)k (Pj()k -1)

(Pjn)   P1   100 Pk( j )
  Pk( j )
( j  1, ..., N ) (12)
k 2 k 2

n W((kj ) 1)
(Wj )n   W 1  W( j )   W((kj ) 1)k
( j  1, ..., N ) (13)
k 2 ( k  1)

21 31 41 51 (n-2)1 (n-1)1 n1
21/31 41/51 (n-2)1/(n-1)1

1 22 32 42 52 (n-2)2 (n-1)2 n2
22/32 42/52 (n-2)2/(n-1)2

2N 3N 4N 5N (n-2)N (n-1)N nN
2N/3N 4N/5N (n-2)N/(n -1)N

Figure 2. Simple concept of distribution network with N main sections

2.4 Economic Value of Losses

The economic value of power P and energy W at each node of the network can be used to
determine the economic value of the losses. If the amount of power losses at the time of maximum load in

IJECE Vol. 6, No. 2, April 2016 : 439 – 446

IJECE ISSN: 2088-8708  443

network for section (k –1)–k is P(k –1)–k, and the amount of energy losses in this section in the considered
period is W(k –1)–k, then costs of the losses in this section can be determined as follows:

Closs
( k  1)k
  P( k  1)k  P ( k  1)   W( k  1)k  W ( k 1) (14)

The amount of energy losses can be expressed by power losses at the time of maximum load and
equivalent time duration of the peak losses:

 W( k  1)k   P( k  1)k ( k  1) k (15)

where (k –1)–k is equivalent time duration of the peak losses, then equation (14) can be written in the form:

Closs
( k  1) k
  P( k  1)k  P ( k 1)   P( k 1)k ( k  1)k  W ( k  1) (16)

Economic value of the energy losses can be found from the equation (16) as follows:

Closs
( k  1)k
 P( k  1)
c( k  1)k     W ( k  1) (17)
 W( k  1)k ( k  1)k

and it depends on the equivalent time duration of the peak losses  and parameters P and W. For longer time
 (for which the load diagram is more uniform), economic value of the energy losses in the appropriate
section is lower. Equation (17) can be used for the actual tariff system. Network losses can be valorized by
actual tariff system in a way that parameter P is tariff element for power and parameter W is tariff element
for energy. The main difficulty in application of equation (17) is the unknowing the load diagram of network
elements. If the load diagram is not accessible, unlike values of delivered energy and maximum power, then
equivalent time duration of the peak losses  can be calculated according to empiric relation, for example:

T2
  0, 17T m  0, 83 m (18)
T

3. NUMERICAL RESULTS
The proposed mathematical model was successfully tested on three distribution networks with
simple configurations and in this section results are presented. In all test cases the annual costs related to
amortization, maintenance and other fixed costs of the network elements are neglected.

Test case 1.
Data for test case 1 is given in Figure 3. According to mathematical model given in Section 2, the
values of parameters P and W for each node, as well as economic value of energy losses for each network
section, are given in Table 1. Parameters for node 1 are: P1 = 0,10627 $/kW and W1 = 0,05983 $/kWh.

1 2 3 Pmax = 6,2 MW
Wa = 18750 MWh
5 km; 0,45 /km Wr = 9100 MVArh
35 kV Sn = 8 MVA
mn,T = 35/10,5 kV
PCu,n = 50 kW
PFe,n = 10 kW

Figure 3. Simple distribution network for test case 1

Economic Valuation of Power and Energy Losses in Distribution Networks (S. Bisanovic)
444  ISSN: 2088-8708

Table 1. Results for test case 1

node node P(k–1)k Pk Tm,(k–1)k (k–1)k W(k–1)k Wk Pk Wk c(k–1)k

k–1 k kW kW h h kWh kWh $/kW $/kWh $/kWh

1 2 87,17 6.332,46 3.024,2 1.380,7 120.345,40 18.886.323,6 0,10773 0,06021 0,05991

2 3 45,29 6.200,00 3.024,2 1.380,7 136.323,56 18.750.000,0 0,10852 0,06065 0,06029

Test case 2.
The second test case is illustrated in Figure 4. Parameters for node 1 are: P1 = 0,11439 $/kW and
W1 = 0,07126 $/kWh. The values of parameters P and W for each node, as well as economic value of
energy losses for each network section, are given in Table 2. Data needed for the calculation is specified in
Figure 4.

3 S3,max = 2+j3 MVA

km
/ T3,m = 2350 h
0,3
km;
1 2 4

3 km; 0,2 /km

10 kV
5k
m;
0,2
/
km

4 S4,max = 4+j5 MVA

T4,m = 3500 h

Figure 4. Simple distribution network for test case 2

Table 2. Results for test case 2

node node P(k–1)k Pk Tm,(k–1)k (k–1)k W(k–1)k Wk Pk Wk c(k–1)k

k–1 k kW kW h h kWh kWh $/kW $/kWh $/kWh

1 2 48,98 6.046,21 2.450,0 985,2 48.256,57 41.242.518,2 0,11532 0,07134 0,07138

2 3 12,74 2.000,00 2.350,0 922,8 11.755,84 11.037.600,0 0,11605 0,07142 0,07147

2 4 33,47 4.000,00 3.500,0 1.755,7 58.762,39 30.134.400,0 0,11702 0,07148 0,07141

Test case 3.
The third test case is illustrated in Figure 5. Parameters for node 1 are: P1 = 0,12437 $/kW and W1
= 0,09351 $/kWh. The values of parameters P and W for each node, as well as economic value of energy
losses for each network section, are given in Table 3. Data needed for the calculation is specified in Figure 5.

IJECE Vol. 6, No. 2, April 2016 : 439 – 446

IJECE ISSN: 2088-8708  445

1 21 31 41 S = 2,6+j0,9 MVA
Z = 0,385+j0,316 pu Z = 0,249+j0,221 pu
W = 16626,48 MWh
Tm = 2950 h
110/20 kV 20/10 kV
Z = 0,124+j0,257

0,332+j0,361 pu
S = 1,9+j0,25 MVA
W = 12815,88 MWh
Tm = 3430 h

42 S = 1,7+j0,25 MVA
22 Z = 0,266+j0,208 pu Z = 0,297+j0,213 pu
W = 11764,68 MWh
32 Tm = 3120 h
S = 2,5+j0,45 MVA 20/10 kV
W = 14454,00 MWh Z = 0,124+j0,257
Tm = 2815 h 52
S = 2,3+j0,2 MVA
W = 13499,16 MWh
Tm = 3620 h

Figure 5. Simple distribution network for test case 3

Table 3. Results for test case 3

node node P(k–1)k Pk Tm,(k–1)k (k–1)k W(k–1)k Wk Pk Wk c(k–1)k

k–1 k kW kW h h kWh kWh $/kW $/kWh $/kWh

1 21 83,02 4.529,35 3.152,7 1.477,7 122.680,06 16.665.399,6 0,12665 0,09390 0,09359

21 31 19,24 2.610,11 2.950,0 1.326,1 25.513,24 16.639.886,4 0,12758 0,09404 0,09399

31 41 10,11 2.600,00 2.950,0 1.326,1 13.406,39 16.626.480,0 0,12808 0,09412 0,09414

1 22 142,00 6.633,71 3.192,2 1.508,1 214.156,21 25.500.307,7 0,12703 0,09401 0,09359

22 32 47,04 4.086,67 3.420,3 1.689,8 79.489,37 25.420.818,3 0,12849 0,09469 0,09409

32 42 10,41 1.700,00 3.150,0 1.475,6 15.361,47 11.764.680,0 0,12928 0,09482 0,09478

32 52 76,26 2.300,00 3.620,0 1.857,0 141.616,87 13.499.160,0 0,13275 0,09569 0,09476

4. CONCLUSIONS AND FUTURE RESEARCH

This paper proposed a simple methodological framework that determines prices of power and
energy at each node of the distribution network, as well as the economic value of energy losses in the
network elements. Given its simplicity, proposed framework can be very useful for a quick losses costs
assessment, as a part of losses costs allocation procedures for network users. It is transparent and could be
practical for implementation. Applications on the test cases presented in Section 3 put these objectives into
perspective. In order to ensure the application of the model on real-life distribution network with distributed
generation, presented methodological framework should be modified, which is future research challenge.
Economically efficient network prices should be computed by considering the marginal impact of each user
on network costs: loads and generators. The type of user (load or generator) and the pattern of network use
are key determinants of individual user’s impact on the network costs.

REFERENCES
[1] P.M. de Oliveira-de Jesus, et al., "Cost loss allocation in distribution networks with high penetration of distributed
renewable generation – a comparative study", in Proceedings of the International Conference on Renewable
Energies and Power, Zaragoza, 2005.
[2] D. Shirmohammadi, and P.R. Gribik, "Evaluation of network capacity use for wheeling transactions", IEEE
Transactions on Power Systems, vol. 4, no. 4, pp. 1405-1413, 1989.

Economic Valuation of Power and Energy Losses in Distribution Networks (S. Bisanovic)
446  ISSN: 2088-8708

[3] H.H. Happ, "Cost of wheeling methodologies", IEEE Transactions on Power Systems, vol. 9, no. 1, pp. 147-155,
1994.
[4] A.J. Conejo, et al., "Z-bus loss allocation", IEEE Transactions Power Systems, vol. 16, no. 1, pp. 105-110, 2001.
[5] J.W. Bialek, "Tracing the flow of electricity", in IEE Proceedings Generation, Transmission and Distribution, vol.
143, pp. 310-320, 1996.
[6] P.M. Costa, and M. Matos, "Loss allocation in distribution network with embedded generation", IEEE Transactions
on Power Systems, vol. 19, no. 1, pp. 384-389, 2004.
[7] K.A. Papadogiannis, et al., "Cost allocation of losses in autonomous power systems with high penetration of RES",
WSEAS Transactions on power systems, vol. 4, no. 6, pp. 210-220, 2009.
[8] J. Mutale, et al., "Allocation of losses in distribution systems with embedded generation", in IEE Proceedings
Generation, Transmission and Distribution, vol. 147, no. 1, pp. 7-14, 2000.
[9] P.M. De Oliveira-de Jesus, and M.T. Ponce de Leao, "Cost loss allocation in distribution networks with embedded
generation: A fuzzy approach", in Proceedings of IEE MedPower 2004, Limassol, 2004.
[10] J.M. Vignolo, and P.M. Sotkiewicz, "Allocation of loss costs in distribution networks: the nodal factor pricing
method", International Journal of Power Systems, vol. 10, no. 10, pp. 1-10, 2004.

BIOGRAPHIES OF AUTHORS

Smajo Bisanovic received the degree of Electrical Engineer in 1991, MSc degree in 1994 and
PhD degree in 2009 from the Faculty of Electrical Engineering, University of Sarajevo, Bosnia
and Herzegovina. He is associate professor at the Faculty of Electrical Engineering, University
of Sarajevo, Bosnia and Herzegovina. His areas of interest include operation, planning and
economics of power systems and application of reliability theory to power systems.

Mersiha Samardzic received a BSc degree in power electrical engineering for the Faculty of
Electrical Engineering, University of Sarajevo, Bosnia and Herzegovina in 2014. She is now
pursuing her MSc degree in the same Faculty. Her research interests include computer
simulations and design analysis applied to power systems.

Damir Aganovic is received a MSc degree in power electrical engineering from the Faculty of
Electrical Engineering, University of Sarajevo, Bosnia and Herzegovina in 2010. He is currently
pursuing his PhD degree in the same field at the same University. He is an Expert associate for
Power System Operation Management at Public Enterprise Elektroprivreda of Bosnia and
Herzegovina. His areas of interest include operation and planning of power systems.

IJECE Vol. 6, No. 2, April 2016 : 439 – 446