POWER SYSTEM RELIABILITY EVALUATION AND QUALITY

ASSESSMENT BY FUZZY LOGIC TECHNIQUE

M. A. Farahat ‘I) and 3. M. AI-Shammari “’

(1) Zagazig University, Zagazig, Egypt
( 2 ) College of Technology at Hail, Hail, Saudi Arabia

ABSTRACT

The basic function of power system is to supply electrical energy to both large and small customers as economically as
possible, and with an acceptable degree of reliability and quality. Reliability and quality are two essential measures and
important components of all power system planning and operation procedures. Reliability is one of the key design
factors when designing complex, critical and expensive systems. This paper presents a fuzzy logic technique for
evaluating the reliability of complex and large power system to deal with their imprecise variables. In the proposed
technique, fuzzy set theory was used to evaluate the coefficients of different terms in reliability expression. System with
hundreds of buses and tens of complex stations can be analyzed using advanced and numerically effective large-scale
algorithms for reliability and quality assessment. An advance computer program for large-scale reliability and quality
assessment is studied and applied to full-scale model of Saudi Arabia electricity system. The proposed algorithm is very
useful to the on-line security assessment of large and complex power system. The results obtained are claimed to have
far-reaching implications on various planning and operation aspects of the Saudi power system.

I INTRODUCTION supply interruptions. For this reason, the new generation

facilities (capacity and location), and the expansion of
Reliability is an essential measure and an important the transmission and distribution systems must be
component of all power system planning and operation planned, built and operated so that the likelihood of
procedures. Reliability of various power system service interruptions is reduced, A modern power
functional zones is determined using either analytical system must be satisfying the system load at a
probability techniques or stochastic simulation methods. reasonable cost and with a reasonable assurance of
The basic function of a power system is to supply continuity and quality. The recognition of reasonable
electrical energy to both large and small customers as assurance is the basis for a wide range of studies
economically as possible, and with an acceptable degree generally designated as reliability assessment. The term
of reliability and quality. Reliability is the ability of a reliability has an extremely wide range of meaning [3].
power system to provide service to customers while In the power system context it can be divided into two
maintaining the quality and price of electricity at an basic aspects. System adequacy and system security.
acceptable level. The power system should also remain Adequacy relates the existence of sufficient facilities
within a set of operational constraints. These constraints within the system to satisfy the consumer load demand
relate directly to the quality of supply such as bus bars or system operational constraints, Reliability evaluation
voltage variations and frequency variations. The ability of interconnected system i s an important problem and
of the power system to operate within the desired limits several methods have been proposed for its solution [4 -
set by the operator upon the event of a contingency is 61. Power systems can be divided into the three
referred to as the secure operation of the power system functional zones of generation, transmission and
[l-21. Though it is essential that the service should be distribution. These functional zones can be combined to
continuously available on demand, it is not possible to create three hierarchical levels (HL) which provide a
do so due to random system failures which are generally practical framework for power system reliability
outside the control of the power system engineers and evaluation. HL1 assessment, usually termed as
operators. The probability of the customer service “generating capacity adequacy evaluation”, is mainly
interruption however can be reduced by increasing concerned with assessing the amount of generating
monetary investment during the planning phase. Also, capacity that must be installed in order to satisfy the
the basic function of the bulk generation and perceived system load and to perform necessary
transmission system is to supply the power energy corrective or preventive maintenance with an acceptable
required by the customers with a certain degree of level of risk. HL II analysis, termed as “CO positive
reliability and quality of service, at the minimum cost. system reliability evaluation” or “bulk power system
Power systems are subjected to contingencies resulting reliability evaluation”, are concerned with the
in the disconnection of generation units, transmission composite problem of assessing the generation and
lines, transformers, etc. which may ultimately result in transmission facilities in regard to their ability to supply

478
adequate, dependable and suitable electrical energy at represent the system loss of load probability for any
the bulk power load points. HL 111 analysis includes all assumed contingency m in the power grid,
three functional zones. The objective of an HL 111 is to
obtain suitable adequacy indices at actual consumers
LULP,'"' = Atmj-, (31
load points. In addition to the basic three hierarchical represent the loss of load probability at bus !for
levels, assessment can be performed separately on any contingency m,,
system subset such as sub transmission, substation,
switching stations and generating stations [ 7 ] . Many
research and development studies were performed to
and Ptodenotes the scheduled demand at load bus e,
develop methods and criteria for reliability evaluation
and quality assessment of system of generation, M, denotes the number of contingencies considered
transmission and distribution [8 - IO]. and Y, is a 0 or 1 factor to indicate subsystem (if
desired).
The objective of the research described in this paper 2. System-Wide Expected Value of Demand Not
was to develop a reliability evaluation technique. An Served:
introduction to fuzzy logic is introduced. In this paper, a "I

fuzzy system for evaluating the reliability of a iarge E (DNS)= Y,E (DNS,) (5)
power systems is presented. Fuzzy set theory is used to I=I
evaluate the coefficients of different terms in reliability where nLis the number of load buses in the system,
expression. The aim of this paper is an application of Mr
the fuzzy concept for the reliability assessment of large & ( D N S t )= C & ( D N p ) (6)
and complex power system. Also, to study and assess m=l
the main features of composite system reliability (with represent the expected value of demand not served at
an emphasis on the I32 kV and 380 KV transmission bus e,
portions) and overall system quality of the power grid of
the Saudi Electricity Company. E(DNS!(" = fmDNSl(m) (7)
represent the expected value of demand not served at
bus t for the contingency m and DNS, (m) = Demand
11 RELIABILITY AND QUALITY EVALUATION not served at bus I for contingency m.
OF POWER SYSTEM 3. System-Wide Expected Value of Energy Not Served:
The reliability indices of the power system represent its
ability to achieves its main objective of maintaining as
much as possible a continuous and sufficient power
supply to its customers at a reasonable cost and rate of
return. The power system consists of numerous m=l
elements, which are characterized by forced outage rate represent the expected value of energy not served at a
representing their tendency to be off-service due to bus I ,
malfunctions. The forced outage rates are calculated
based on historical outage data of the elements E (ENS,'"') = f,ENS,'"'
themselves or of similar in-service elements. The represent the expected value of energy not served at
reliability of the system depends on the reliability of its bus l for contingency m,
individual elements, operating practices of the system as
well as the system size and structure. Other factors can ENS,(") = T(m)",(m) (1 1)
be considered as well, such as the environmental effects represent the energy not served at bus C for
and load shedding properties. contingency m and T"'denotes the time duration of
contingency m.
11.1 Power System Reliability Indices
11.2 Station Reliability Measures:
In general, a set of system-wide outage-based reliability
indices can be defined. These reliability indices are The following are some key indicates and measures,
sufficient to describe a range of practical reliability which govern the station reliability assessment
measures in large-scale power system. The following procedures:
system-wide contingency-based reliability indices may
be defined: 1. Maximum station flow (MVA), which can be
1. System-Wide Loss of Load Probability (LOLP): sustained in a given station subject to a specified
MC single or multiple contingencies (outages).
LOLP = LOLP"' (1) 2. Station supplied load loss (MVA), which can be
m=l caused in a given station subject to a specified single
or multiple contingencies.
Where LOLP'"' = Maxl fy( LoL<'"' } (2)

479
3. Station bottled generation (for power plants), which MLD-LOST = Minimum load lost
can be caused in a given generating station subject to = Max {0,L-F}
a specified single or multiple contingencies. XLD-LOST = Maximum load lost
4. Non-utilized station capacity, which can be caused in = Max(0, L+O-F} & 5 L
a given station subject to a specified single or MGN-BTLD = Minimum Generation Bottled
multiple contingencies. This is a measure of the = Max{O, G-F)
amount of incoming power to the station, which can XGN-BTLD = Maximum Generation Bottled
not be delivered to the grid because of limitations on = M={O, G+I-Fj & I G
the station flow patterns. MCP-NUTZ = Minimum Capacity Un-utilized
5 . Surplus station capacity, which can be caused in a = Max{ 0, F-G-I}
given station subject to a specified single or multiple XCP-NuTZ = Maximum Capacity Un-utilized
contingencies. This is a measure of the amount of = MaxiO, F-G}
station flow that can be sustained by the station, but MCP-SPLS = Minimum Capacity Surplus
cannot be delivered to the grid because of limitation
= Max{O, F-L-0)
on the station output feeder capacities. XCP-SPLS = Maximum Capacity Surplus
= Max{0,F-L}
11.3 System Reliability Measures:

The following are some key indices and measures, System-wide quality indices are evaluated using similar
formulas as (12) to (19), which in this case are applied
which govern the system reliability assessment
to system areas and zones of interest. The system
procedures :
connectivity structure is used in rather complex
1. Loss of load probability (LOLP), which indicates the
algorithms to interconnect various stations among a
probability that a system load would be lost, fully or
given area (or zone) and between different areas (or
partially, due to randomly occurring single or multiple
zones) in the system.
contingencies (outages) in the system.
2.Expected demand not served (EDNS) and expected
load not served (ELNS), which indicate the amount of 11.5 System Quality Assessment:
MW (or MVA) that is likely not to be supplied to a
The following are some key indices and measures,
system load center due to randomly occurring.
which govern the quality assessment procedures:
3. Expected energy not served (EENS), which indicates
1 . Area and zone loss of load probability (LOLP), which
the amount of TWh of energy ‘per year that is likely
indicates the probability that an area or zone load
not to be supplied to a system load center due to
randomly occurring in the system. would be lost, fully or partially, due to randomly
occurring single or multiple contingencies in the
system.
11.4 Station and System Quality Indices:
2. Area and zone expected demand not served (EDNS)
and expected load not served (ELNS), which indicate
Fig. (1) demonstrates the basic model structure for
the amount of MW (or MVA) that is likely not to be
evaluating various quality indices. The following
supplied to an area or zone load randomly occurring
symbols and code-names are used
L = TOT-LOAD (MVA) = Total local load supplied single or multiple contingencies in the system.
3. Area and zone expected energy not served (EENS),
by the station
F = “FLOW (MVA) = Maximum flow through which indicates the amount of TWh of energy per
year that is likely not to be supplied to an area or zone
the station
G = TOT-GENX (MVA) = Total generation
load due to randomly occurring singIe or multiple
contingencies in the system.
available at the station
4. Probability of generating power and energy bottled,
I = TOT-TMP (MVA) = Total transmission input
to the station which can be caused in a given area or zone subject to
0 = TOT-TOUT (MVA) = Total transmission a specific single or multiple contingencies.
5. Probability of non-utilized power and energy, which
output from the station
can be caused in a given area or zone subject to a
specific single or multiple contingencies.
r 6. Probability of surplus power and energy, which can
be caused in a given area or zone subject to a specific
single or multiple contingencies.

I11 FUZZY SETS AND MEMBERSHIP

FUNCTIONS
Fig. (1) Basic model for quality evaluation
A complex expression of the power system reliability
Then, the following quality indices are defined: can be consider as [lo, 121:

480
 By applying some of rules used as the guiding rules for
fuzzy reasoning to find the value of reliability
where m is the number of system components coefficients.
(transmission lines, transformers, generators, etc.); p, q
is the probabilities of system components to success and
fail respectively; and a, i s the ith coefficient taking IV CASE STUDY
values from the range [0, 11. The system reliability can
be fully determined by a certain set of values assigned The system under study is a distribution system of the
Saudi Consolidated Electric Company - Central Power
to the ai coefficients. These coefficients are depending
on the system configurations and on the component System (SCECO-C). The 2000 system peak base-case is
used in this work. The power system under study is
rating or system redundancy. Fuzzy set theory is based
on a recognition that certain sets have imprecise divided into five areas, namely Riyadh City (Cl),
boundaries. The imprecision of fuzzy sets deals with the Qassim (C2), AI-Khaj (C3), Riyadh Rural fC4) and
shades membership of an object in a set with imprecise Dawadmi (C5). The first area is divided into four zones.
boundaries, whereas the randomness of probability Each of the other four areas consists of one zone.
theory deals with the uncertainty regarding the
occurrence or nonoccurrence of some event. A fuzzy set
is a class of objects with a continuum of grades of V RESULTS
membership. The core concept of fuzzy-set theory is a
membership function which represents numerically the To evaluate the reliability of the distribution system, the
following data are needed:
degree to which an element belongs to a set. The
membership function takes on values between 0 and 1.
It is assessed subjectively in any instance. Fuzzy
1. The failure rates of all types of components, such
numerical data can be represented by means of fuzzy as lines, transformers, switches, fuses, ... and
subsets of the real line, known as fuzzy numbers [ 10 ~
breakers.
2. The restoration time of each component.
121. Consider a fuzzy number a of real d whose 3. The switching times of switches and breakers.
membership function: 4. System configuration.
5 . Load and number of customers of each toad point.

Any fuzzy approach depends essentially on historical

I x-a,
U-).
a-a,
,
otherwise data of the applied item. The historic data of the
distribution system from 1996 - 2000 are provided by
SCECO-C power system. Using these data to evaluate
where the subsets U and 1 imply the upper and lower
side of x, respectively. Generally R and L are called L-R the reliability indices. Based on the results obtained
fuzzy numbers, and typical types of membership throughout the present study, the following key findings
function. If F,, represents the fuzzy service continuity concerning system reliability levels are shown in Table
set, and Fd represents the fuzzy service discontinuity set (1). Table (2) and Table (3) summarize some key
and the fuzzy term D as the intersection of both, then D quality results for the SCECO-C power system under
is defined as: study. Table (4) shows brief definitions of various
D = F,, nFrd terms, which appear in the system reliability and quality
results.
And its membership function as shown in Fig. (2), was
characterized by:
All 380 kV stations in SCECO-C enjoy high station
PD (x>= P c rl P s d (4 reliability performance with zero loss of load
(23) probability. On the other hand, there are fifty of the 132
= mWP&), Psd(41 kV stations in the system with zero loss of load
probability, twenty with loss of load probability less
than 0.004, and seventeen with loss of load probability
less than 0.011. These are all considered as high-
reliability stations and, perhaps, do not need special
attention from the reliability point of view. However,
there are twenty of the 132 kV stations in the system
with unity loss of load probability.

VI CONCLUSIONS
X
Lower service reliability levels jeopardize energy
Fig. (2) Fuzzy membership supply continuity and increase the likelihood of

481
additional maintenance and the restoration costs due to Table (1) Some key findings conceming system
the resulting higher rate of system outages. System
quality indicates the desired balance between generation
facilities, transmission capabilities and consumer
demand levels in various zones o f the electric power
system. Poor system quality levels often imply either
deficiency or excess in the overall system capabilities as
designed by its planners. The costs associated with low
service reliability or poor system quality are enormous,
and can be largely avoided if enhances system planning
simulation models and appropriate computer-sided
solution tools are developed and used to detect and
correct potential problems.

Conventional techniques of power system reliability

evaluation of power system are more complicated and
required a large number of computational efforts. So in
this work, a new fuzzy technique for evaluating
reliability is presented. The proposed fuzzy algorithm is
simpler and the computational efforts are very small
compared with that of the conventional methods.

Table (2) Summary of some system quality results.

Table (3) Summary of some system quality results.

482
Table (4) Description of fields in output quality system [ 5 ] Y . Li and C . Singh, “A New Conceptual Framework
for multi-area Power System Reliability Evaluation”,
IEEE Power Engineering Society Summer Meeting, 18
- 22 July 1999, Vol. 2, pp. 676 - 68 1.
[6] N. Gubbala and C. Singh, “A Fast and Efficient
Method for Reliability Evaluation of Interconnected
Power System - Preferential Decomposition method”,
IEEE Trans. On Power Systems, Vol. 9, No. 2,May
1994, pp. 644 - 652.
[7] R. Billinton, H. Chen and J. Zho, “Individual
Generating Station Reliability Assessment”, IEEE
Trans. On Power Systems, Vol. 14, No. 4, November
1 9 9 9 , ~1238
~ . - 1244.
[8] R. N. Allan, R. Billinton, A. M. Breipohl and C. H.
Gigg, “Bibliography on the Application of Probability
Methods in Power System Reliability Evaluation”,
IEEE Trans. On Power Systems, Vol. 14, Part 1,
February 1999, pp. 5 1 - 57.
191 M. A. El-Gammal, E. N. Abdallah, A. R. Abdelaziz
and A. Abdelmordi, “A Fuzzy - Based g-Reliability
Assessment of Large Power System Critical Links
~

Connection”, Seventh International Middle ~ East

Power System Conference MEPCON 2000, Ain Shams
University, Cairo, Egypt, March 28 - 30, 2000, pp. 652
- 656.
[ 101 A. R. Abdelaziz, “A Fuzzy - Based Power System
VI1 REFERENCES
Reliability Evaluation”, Electric Machines and Power
System Journal, Vol. 27, No. 3, 1999, pp. 271 -278.
[ I ] H. Solehfar and S. Trinadi, “Application of [ 1 1 1 W. Pedrycz, “Fuzzy Sets Enginnering”, CRC Press
Perhubation Analysis to Sensitivity Computations of
Boca Ratom Ann Arbor London Tokyo, 1995.
Generating Units and System Reliability”, IEEE Trans. [I21 A. Kanagawa and H. Ohta, “Fuzzy Design for
On Power Systems, Vol. 13, No. 1, February 1998, pp.
Fixed-Number Life Tests”, IEEE Transactions On
152- 158. Reliability, Vol. 39, No. 3, August 1990, pp. 394 - 398.
[2] K. N. Tinnium, P. Restgoufard and P. F. Duvoisin,
“Cost - Benefit Analysis of Electric Power System
Reliability”; Proceeding of 261h Southeastem AUTHORS
Symposium On System Theory, 1994, pp. 468 472. ~

Dr. Mohamed A. Farahat, Electric Power Eng.

[3] M. Sebastian, P. V. Vidal and L. K. David, “BUS- Dept., Faculty of Engineering, Zagazig University,
Based Reliability Indices and Associated Costs in the Zagazig Egypt.
~

Bulk Power System”, IEEE Trans. On Power Systems, E - mail: farahat 707@hotmail.com
VOI,13, NO, 3, August 1998, pp. 719 - 724.
On leave: Electrical Technology Dept., College of
[4] R. Billinton, S. Kumar, N. Chowdhury, K. Chu, L. Technology at Hail, Kingdom of Saudi Arabia.
Goel, E. Khan, P. Kos, G. Nourbakhst and J. Oteng- B. M. AI-Shammri, Electrical Technology Dept.,
Adjci, “A Reliability Test System For Educational College of Technology at Hail, Kingdom of Saudi
Purposes-3asic Results”, IEEE Trans. On Power
Arabia.
system, Vol. 5, No. 1, February 1990, pp. 319 - 325. E-mail: Badr MS($hotmail.com

483