Optimization of Distribution System Reliability

L. Zemite, J. Gerhards, M. Gorobetz, A. Levchenkov

 Different types of losses calculations depend on different

Abstract— reliability analysis of distribution systems has been durations of interruptions, distribution of the customers’
attracting increasing attention. A special concern pertains to the groups, methods of results calculations, methods of data
distribution networks on which most failures occurs. The obtaining, etc. [2, 3].
optimization of distribution system of breakers and power
According to the calculation types the methods of losses
switches is a possible strategy to improve reliability. The paper
describes development procedure for modelling restoring after a calculation can be divided into three subgroups – analytical,
fault and calculating associated reliability indices and customers’ simulation and methods of customers’ interview [2, 3].
outage costs. The developed model of the network and reliability After choice of methods – analytical, simulation or interview,
and outage costs calculating algorithm is suitable for multi-criteria the direct and indirect losses resulted from interruptions should
analysis of the network. Proposed reliability and outage costs be analyzed and calculated.
calculation algorithm is based on Monte Carlo simulation and
The factors influencing the reliability can be divided into
genetic algorithm.
subgroups according to the customers of electric energy,
Index Terms— distribution network, simulation, power supply undelivered energy or power, duration of interruptions,
reliability. frequency as well as combining these subgroups in different
ways.
The customers groups are divided taking into account equal
I. INTRODUCTION electric energy consumption and equal interruption losses

W ith the development of economy and mankind the electric (Fig.1.) [4].
distribution networks and technical and technological Methodology of calculation of electric supply losses from
solutions of the equipment connected to them are also changing interruptions The values of the electric supply reliability and
resulting in the changes of its application opportunities and losses resulted from the interruptions are calculated with the
requirements to quality of the supplied energy. accounting of the following factors: structure of electric supply
Taking into account the conditions of the free market and Observance of the reliability influencing factors in methods for calculation of the
increasing demands of the customers in an uninterrupted losses resulted from the electric supply interruptions

electric power supply the effectiveness of capital investments

are expected to be determined as well as losses resulted from Group of
Society Seperate customers
the supply interruptions and electric supply reliability should be customers

calculated.
The basic task of an electric supply network operator is to Private Commercial Agricultural Public Industrial
customers customers customers customers customers
provide a customer with energy supply of a necessary level of
reliability and quality with as low financial expenses as
possible. Intensity of Unsupplied Unsupplied Duration of Frequency of Different
interruption energy power interruption interruption combinations
The evaluation of interruptions risks of an electric supply
system requires to know structure of the network, its load and
Fig. 1. Factors influencing the losses from electric supply interruptions.
customers data [1].
network, undelivered electric energy, distribution functions of
II. METHODS FOR CALCULATION OF THE LOSSES the interruptions durations probabilities, the duration of electric
RESULTED FROM THE ELECTRIC SUPPLY supply system interruptions elimination for reserved,
INTERRUPTIONS unreserved and auxiliary elements as well as the losses of the
The losses resulted from the electric supply interruptions that network, society and customers resulted from the interruption
have economic and social influence on the society can be [5, 6, 7].
divided into direct and indirect losses. The purpose of the calculations is to consider different
Direct losses are connected with undelivered electric energy. scenarios as well as to calculate the losses form the
Indirect losses are not connected with the interruptions interruptions.
themselves but with their consequences. For the calculation of the losses the following tasks are

TRANSACTIONS ON ENVIRONMENT AND ELECTRICAL ENGINEERING ISSN 2450-5730 Vol 1, No 3 (2016)

© L. Zemite, J. Gerhards, M. Gorobetz, A. Levchenkov
defined:
 to model the network and select the criteria for ECOST(t,N,d ) = N  f m (t)  f n (t)  f h (t)  C(d)  t d (1)
reliability and from interruptions resulted losses;
 to provide an opportunity to calculate the losses T7 T6

resulted from the electric supply interruptions; 39 40

T18
L17 L18


T5 51

to develop the methods for calculations of losses

L16
36 37 38

L23
T4 T17

resulted from the interruptions for different periods of QF1

L1 L2 34
35

L3
QF2
L4
49
48
50

L5 L6
QF5

time and models of network taking into account A1

12 30 32 17 5 7
52 9
55
10
A4

L7
1 2 3 4 6 8 11

L20

L24
13 QF3
54

L19

L25
27 T16
18

consumption of electric energy, loading factor, length

T2 33 53

L8
T1 T20

L10
14 T9 T19
15 31
19 20 T21

of the line, structure of the network, number of the

L11
L26 T8 T10
T3

L9
21
28 29
L21

customers, expenses for the interruptions elimination

T11

L12
16
41 44

A2 43
42

and capital investments, etc. and analyze the obtained

22
T12 T13

L13
results.
T14
45
23
L22

L14
T15

In the calculations of total losses for different scenarios for

46
24 47

L15
the analysis of capital investments scenario the several factors 25

QF4

considered in the multi-criteria analysis are taken into account A3

26

(Fig.2.) [4, 8].

main line, connecting lines, branch line, A – supply source, A1, A4 – supply
source of main line, A2, A3 – supply source of connecting lines, L1 – L26 – lines, QF1 – QF5 – power
Start switch, T1–T21 – transformer substations (TS), T16 – transformer substation connected to main line,
T2, T10 – transformer substation connected to connecting line, 1, 2, 3,…, 55 – breakers, QF4, QF5, 16
– places of distributing
Choose of
network Fig. 3. Model of distribution system.
distribution
Operational model
costs Rural area Urban area Average
For the determining of the efficiency of the planned capital
Depreciation of
elements
Elimination of
interruption
investments in the network with several possible solutions that
Location of
failure
Group of users can change the model of the network, the economic calculations
are required.
Distance to
failure Users’
TS load factor disposable
Interruption
prevention Calculations Capital costs
incomes While calculating the level of the losses from interruptions
Labor costs
and capital investments the capital investments expenses,
costs
TS power Daytime factor

Costs of
breakers Interruption Number of
ageing of the elements and the expenses for the elimination of
the supply system interruptions.
Calculation frequency users
losses resulting
from Costs of power
interruptions switches Interruption
duration
Weekday factor The expenses are formed from (2), where C – losses from
Life time of
elements
interruptions during the calculation period, €/year, Cki – capital
investments expenses, €/year, CANi – interruptions elimination
Lenght of lines Month factor

Loan period Interruption

duration
Electrical
energy market expenses, €/year, CEUi – direct and indirect losses of the
hypthesis price
Results
Inflation customers, €/year, n – number of new elements, m – number of
Lending interest Distribution
customers [8]
Society level
rate network level
Decision

C = i 1 Cki i 1 CEUi  C AN  min . (2)

n m
Direct losses
resulting of
interruptions
Indirect losses
resulting from Users level
End interruptions

Fig. 2. Multi–criteria analysis of calculation of electric supply reliability and

losses resulted from interruptions. III. DESCRIPTIONS OF CALCULATION OF ELECTRIC
SUPPLY RELIABILITY AND LOSSES FROM
20kV distribution network of Latvia is analyzed in details
INTERRUPTIONS
and calculated. It resulted in a developed model of network the
data determined below (Fig.3.). A – Star algorithm is a heuristic method for the way search
The reservation of electric supply is possible along the in given graph. The algorithm detects whether there way from
connecting line for supply sources A2, A3 and A4. the starting point to the end point.
In normal regime the power switches QF1, QF5, QF3 are in There are developed algorithm modifications, which are
on condition. intended for checking whether the customer is connected to a
In normal regime power switches QF4 and QF2 are in off power source. In the algorithm modifications there are taken
condition. Points 16, QF4 and QF5 are distribution places. into account in additional restrictions – the reserve source
In general case for a particular group of customers the losses searching in the case of network element interruption in
resulted from the interruptions depend on the number of accordance with network node positions [9].
customers (N), on the month fm(t), on the day of a week fn(t), If to take into account the features of the network structure
and time of a day fh(t). Thus the average incomes of the according to a particular customer it is possible to obtain a more
correspondent customers C(d) (€/year) and td – duration of accurate duration of the interruptions and frequency for each
interruptions, h (1) [3] customer.
 Reservation the elements of electric supply can be divided
into 3 basic groups – unreserved, reserved and auxiliary.
 [ x]  aN[ x]  tN  aR[ x]  tR  aP[ x]  tP . (6)
Reservation the elements of electric supply can be divided
into 3 basic groups – unreserved, reserved and auxiliary.
Unreserved elements are those which in case of interruptions
A. Calculations of interrupted energy assessment rate
customers cannot be provided with electric supply along the
lines or from other sides. The statistical data of several countries were compared, the
Reserved elements are those which in the case of calculated interrupted energy assessment rate (IEAR) was
interruptions customers can be supplied from other sides. determined for all groups of customers; the values of maximum
Auxiliary elements are the elements that in case of and minimum undelivered electric energy were calculated [10].
interruptions for customers can restore electric supply for the B. Calculations of losses from interruptions by Monte Carlo
element under consideration, as soon as interrupted element is method
interrupted [9].
The electric supply reliability and the losses resulted from the
The frequency of the elements interruptions regarding to the
electric supply interruption for the models with additional
considered transformer substations (TS) can be found as (3),
where N[x] – frequency of failures of unreserved elements, manually indicated localizations of power switches are
calculated on the basis of Monte Carlo modelling method.
TA [x] – frequency of failures of unreserved TS, N lin [x] – The calculations take into account the analysis of the factors
frequency of failures of unreserved lines, N AT [x] – frequency influencing the reliability for the urban and rural areas electric
of failures of unreserved breakers, N JS [x] – frequency of supply network models as well as for the average model with
failures of unreserved power switches, R[x] – frequency of the purpose to calculate the expected duration of the
failures of reserved elements, Rlin [x] – frequency of failures of interruptions without the analysis of previous power switches
localizations [8, 11].
reserved lines, RAT [x] – frequency of failures of reserved
breakers, RJS [x] – frequency of failures of reserved power C. Calculations of losses from interruptions by genetic
switches, P[x] – frequency of failures of auxiliary elements, algorithm with optimized numbers and localizations of power
switches
Plin [x] – frequency of failures of auxiliary lines, PAT [x] –
The optimization of the number and localization of the power
frequency of failures of auxiliary breakers, PJS [x] – frequency
switches is realized with the help of genetic algorithm (GA),
of failures of auxiliary power switches: taking into account capital expenses and the expenses of
elimination of supply system interruption [8, 10, 12].
The calculation of losses from interruptions according to the
N [ x]  TA[ x]  N lin [ x]  N AT [ x]  N JS [ x], number and localization of additionally placed power switches
 (3)
R[ x]  Rlin [ x]  RAT [ x]  RJS [ x], in main lines and connection lines is realized with the aim to
P[ x]  P [ x]  P [ x]  P [ x]. define an optimal number and localization of power switches
 lin AT JS
taking into account the capital investments for power switches
[10].
The total frequency of failures of the elements interruptions is D. Calculations of losses from interruptions by genetic
calculated with (4) algorithm with optimized numbers and localizations of
breakers
 [ x]  N[ x]  R[ x]  P[ x]. (4) The criteria of the optimization of the power switches resulted
in the decision to optimize the breakers in the main lines and
connecting lines, taking into account the expenses of the capital
The total frequency of the elements interruptions a[x] can be investments and interruptions elimination.
calculated as (5), where aN[x] – number unreserved elements, The optimization of the number and localization of breakers
in main and connecting lines is realized with the help of genetic
aR[x] – number reserved elements, aP[x] – number of auxiliary
algorithm for the electric supply network models of rural area,
elements
urban area and average type.
Capital investments and expenses for the interruptions
a[ x]  aN[ x]  aR[ x]  aP[ x] . elimination without loan, for the loan period of 10 years and for
(5) the loan period of 25 years.

IV. RESULTS OF CALCULATIONS OF ELECTRIC

The total duration of the elements interruptions  [x] is (6), SUPPLY LOSSES RESULTED FROM INTERRUPTIONS
where tN – duration of interruption of unreserved elements,
A. A. Results of interrupted energy assessment rate
tR – duration of interruption of reserved elements, tP –
duration of interruption of auxiliary elements IEAR calculated for an average electric supply network
model is demonstrated in Fig. 4.
 IEAR calculated for an electric supply network model in rural For an average electric supply model with two additional
areas is demonstrated in Fig. 5. manually indicated power switches the total losses resulted
IEAR calculated for an urban electric supply network model from the interruptions are reduced for 19 %, for the models of
is demonstrated in Fig. 6. rural areas – for 3.2 %, urban areas – for 19 %.
The summarized data show that the losses from the For an average electric supply model with three additional
manually indicated power switches the total losses resulted
from the interruptions are reduced for 22 %, for the models of
rural areas – for 21 %, urban areas – for 22.8 % (Fig. 7).
The optimal localization of the power switches in the network
is a significant factor for improving of electric supply reliability
and, therefore, decreasing of the losses from interruptions.
The absence of the optimal localization of sectioning

Fig. 4. Average IEAR.

interruptions in 20 kV electric supply network in Latvia are
equivalent to those in Finland, Norway, the Netherlands, USA

Fig. 7. The total losses resulting from interruptions.

elements in the network can cause an increasing or insignificant
decreasing of the losses resulted from the interruptions as well
as the manually indicated localizations of the sectioning
elements do not give an opportunity to calculate the efficiency
Fig. 5. IEAR in a rural area.
of capital investments for the reliability level increasing [8].
and Sweden [4, 11, 13]. The reducing of the losses from interruptions are directly
proportional to TS power, number of customers, consumption
of electric energy and incomes of the customers.
C. Results of calculations of losses from interruptions by
genetic algorithm with optimized numbers and localizations of
power switches
The optimization of the number and localization of power
switches in main and connecting lines is realized with the help
of genetic algorithm for the electric supply network models of
rural area, urban area and average type.
The total losses of the distribution network, society and
customers resulted from the interruptions after the optimization
Fig. 6. IEAR in a urban area. of the number and localization of power switches and for
different loan periods are given in Fig.8.
B. Results of calculations of losses from interruptions by After the optimization of the number and localization of
Monte Carlo method power switches for the case of no loan the losses resulted from
the interruptions for the average network model are reduced for
The purpose is to calculate the parameters of reliability and
14 %, for the model of rural area electric supply network – for
losses resulted from the interruptions. Application of Monte
12.6 %, and for the urban area model – for 33.6 %.
Carlo method gives an opportunity to obtained more accurate
Total losses from interruptions for the loan period of 10 years
results using distribution of probabilities and reducing the
in the model of average network are reduced for 8 %, in the
number of assumptions. For an average electric supply model
model of rural area network – for 7 %, and in the model of urban
with one additional manually indicated power switch the total
area – for 30 %.
losses resulted from the interruptions are reduced for 14 %, for
the models of rural areas – for 1.6 %, urban areas – for 19 %.
 Total losses from interruptions for the loan period of 25 years model of rural area network – for 3 %, and in the model of urban
in the model of average network are reduced for 5.8 %, in the area – for 7 % (Fig. 9).
model of rural area network – for 11 %, and in the model of Taking into account the total losses from the interruptions, the
urban area – for 20 %. additional connection of breakers can be considered from the
Summarizing the results we can conclude that for the models perspective of smart grid development and easy maintenance.
of average and rural areas electric supply networks the capital
investments for the purchasing of power switches are
proportionally decreased.
The total losses resulted from the interruptions are
significantly decreased as a result of sectioning for the model

Fig. 9. Total losses resulted from the interruption for the case of breakers
optimization.

V. CONCLUSIONS
Fig. 8. Total losses resulted from the interruption for the case of power
switches optimization.
The proposed methods can be applied for the calculations of
the interruption risks of the customers and compensation of the
losses resulted from the interruptions.
of urban area electric supply network.
The possible future modifications of the proposed methods
This is connected with higher consumption of electric energy
are applicable in the solutions of different problems related to
in the urban regions and higher incomes of the customers [8].
the analysis of electric supply interruptions.
Taking into account the total losses from the interruptions the
The overview, analysis and systematization of the methods of
additional connection of power switches can be considered
distribution networks reliability calculation, optimization
from the smart grids development and maintaining easy point
methods of reliability improvement, the alternative of the
of view.
reliability improvement and the losses resulted from the electric
supply interruptions give an opportunity to search for the
D. Results of calculations of losses from interruptions by solutions of capital investments on the level of distribution
genetic algorithm with optimized numbers and localizations of networks, customers and society taking into account the
breakers tendencies of development and perspective technologies.
Note that the optimization of the number and localization of Optimization of the number of power switches and
the breakers does not give an opportunity to achieve a minimum localization completed with the help of genetic algorithm, in
level of the losses resulting from the interruptions. accordance with the loan term and properties of the network
Besides, the correspondence of the optimal number of the model, allows reducing the losses from interruptions from 5.8
breakers to that of the power switches, the breakers cannot % to 33.6 %.
provide the operation of the network without the interruptions Optimization of the number of breakers and localization
that is why the minimized expenses include the customers’ completed with the help of genetic algorithm, in accordance
losses from the interruptions. with the loan term and properties of the network model, allows
Fig.9. represents the total losses of the distribution network, reducing the losses from interruptions from 1 % to 9 %.
customers and society for an optimal number of breakers for the The risks, losses, planning and advantages of the electric
loan period of 10years, 25 years and without the loan. supply network should be analyzed in accordance with
After the optimization of the number and localization of economic, environment, electric supply quality, probability of
breakers for the case of no loan the losses resulted from the interruptions, the risks of the changes in legislation, etc.
interruptions for the average network model are reduced for 2 Estimating the solutions of the electric supply reliability
%, for the model of rural area electric supply network – for 1%, improvement most of the attention should be turned to the
and for the urban area model – for 9%. selection of optimal number and localization of the sectioning
Total losses from interruptions for the loan period of 10 years elements that provides an immediate improvement of the
in the model of average network are reduced for 6 %, in the electric supply reliability.
model of rural area network – for 3 %, and in the model of urban Taking into account the total losses from the interruptions the
area – for 9 %. additional connection of power switches and /or breakers can
Total losses from interruptions for the loan period of 25 years be considered from the smart grids development and easy
in the model of average network are reduced for 8 %, in the maintenance point of view, but not from the losses decreasing
point of view. reliability, automatization, power system development,
The volume of the possible losses resulted from interruptions planning and control. He was leading various national and
can be variable with the increasing of electric energy international projects, author of many patents, books and
consumption and/or changing of the customer properties. publications.

ACKNOWLEDGEMENT Mikhail Gorobetz is graduated PhD. in 2008, assistant

professor and leading researcher of Riga Technical University
This research work has been supported by Latvian Council of Institute of Industrial Electronics and Electrical Engineering.
Science (Project No. 673/2014). L. Zemite, J. Gerhards, M. Results of research activity are published in various
Gorobetz, A. Levchenkov, Riga Technical University Faculty international scientific proceedings and journals in fields of
of Electrical and Power engineering, Riga, Latvia. adaptive control, neural networks, genetic algorithms,
modelling and simulation of dynamic processes. He is leader of
various national projects and international projects. He is author
REFERENCES of many study books and patented inventions.
[1] Brown R. E. Electric Power Distribution Reliability. – Marcel Dekker,
New York, 2002. – p. 365. Аnatoly Levchenkov, PhD professor, Riga Technical
[2] Alevehag K. Impact of Dependencies in Risk Assessments of Power University, Institute of Industrial Electronics and Electrical
Distribution Systems. – Licentiate Thesis, Royal Institute of Technology,
School of Electrical Engineering, Electric Power Systems, Stockholm:
Engineering, Institute of Railway Transport. He got engineer
Sweden, 2008. – p. 155. diploma in electrical engineering in 1969, PhD degree in 1978.
[3] Helseth A. Modeling Reliability of Supply and Infrastructural His fields of interests are optimization theory, group decision
Dependency in Energy Distribution Systems. – Thesis for the degree of support systems, negotiation support systems, scheduling,
philosophy doctor, Trondheim, Norwegian University of Science and
Technology, 2008. – p. 132. logistics, intelligent transport systems, evolutionary algorithms
[4] Chayakulkheeree K. Economy of Power System Reliability. – A Training for embedded systems. He was leading various national and
Workshop on Power System Economics and Planning, Asian Institute of international projects, author of many patents, books and
Technology, 2006. – p. 68. publications.
[5] Kjølle G. H., Samdal K., Singh B., Kvitastein O. A. Customer Costs
Related to Interruptions and Voltage Problems: Methodology and Results.
– Power Systems, IEEE Transactions on, vol.23, no.3, Aug. 2008. – pp.
1030. – 1038.
[6] Billinton R. Reliability of power supplies. In Electronics and Power ,
vol.24, no.4, , April 1978. – pp. 307. – 310.
[7] Theil A., Theil M. Medium Voltage network reliability: efficiency
oriented supply restoration strategies. – 15th PSCC, Session 8, Paper 5,
Liege, 2005. – pp. 1–6.
[8] Distribuzione C. N. Optimal Placement of Automation Devices in Enel
Distribution Network. – in Electricity Distribution – Part 1, 2009. CIRED
2009. 20th International Conference and Exhibition, 8–11 June 2009. –
pp. 1. – 4.
[9] Haghifam M. R. Optimal allocation of tie points in radial distribution
systems using a genetic algorithm.– Eur. Trans. Elect. Power, vol. 14, no.
2, 2004. – pp. 85. – 96.
[10] Zemīte L., Gerhards J., Gorobecs M., Ļevčenkovs A. Optimization of
Switch Allocation in Power Distribution Systems – International
Workshop on Deregulated Electricity Market Issues (DEMSEE 2015),
Ungārija, Budapešta, 2014.. – pp.166. – 173.
[11] Sljivac D., Nikolovski S., Kovac Z. Distribution Network Restoration
Using Sequential Monte Carlo Approch. – The 9th International
Conference on Probabilistic Methods Applied to Power Systems, June
11–15, 2006. – p. 6.

Laila Zemite is graduated PhD in 2016, assistant professor of

Riga Technical University Institute of Power and Electrical
Engineering. Results of research activity are published in
various international scientific proceedings and journals in
fields of distribution system reliability, power system
development, planning and control. She is leader of various
national projects and international projects.

Janis Gerhards is PhD professor, Riga Technical University,

Institute of Institute of Power and Electrical Engineering. His
fields of interests are optimization theory, reliability research in
free electricity market conditions, distribution system