IEEE Transactions on Power Systems, Vol. 14, No.

2, May 1999 397

Teaching Distribution System Reliability Evaluation

Using Monte Carlo Simulation
Roy Billinton, Fellow, IEEE Peng Wang
Power Systems Research Group
Electrical engineeringDepartment
University of Saskatchewan
Saskatchewan,Canada

Abstract-Analytical techniques for distribution system reliability distribution system reliability evaluation. The mean values,
assessment can be effectively used to evaluate the mean values of however, do not provide any information on the variability of
a wide range of system reliability indices. This approach is the indices. The probability distribution, however, provides
usually used when teaching the basic concepts of distribution
system reliability evaluation. The mean or expected value,
both a pictorial representation of the way the indices vary and
however, does not provide any information on the inherent important information on significant outcomes, which,
variability of an index. Appreciation of this inherent variability is although they occur very infrequently, can have very serious
an important parameter in comprehending the actual reliability system effects. These effects, which can easily occur in
experienced by a customer and should be recognized when practise, may be neglected if only average values are
teaching distribution system reliability evaluation. This paper available. Probability distributions of the relevant reliability
presents a time sequential Monte Carlo simulation technique indices can also be important in reliability cost/worth analysis
which can be used in complex distribution system evaluation, and for industrial customers with critical processes or commercial
describes a computer program developed to implement this customers with nonlinear cost functions [ 13. The conventional
technique. General distribution system elements, operating approach to teaching distribution system reliability evaluation,
models and radial configurations are considered in the program.
The results obtained using both analytical and simulation in either a university or industry based setting, is to use the
methods are compared. The mean values and the probability basic analytical equations to calculate load point failure rates,
distributions for both load point and system indices are average outage durations and average annual outage times [ 11.
illustrated using a practical test system. This indices can then be combined with the customer
composition in the distribution system to evaluate the system
Keywords: Reliability, Power distribution system, Monte Carlo indices of SAIFI, SAIDI, CAIDI and ASAI ...[11. It is usual
simulation practise in teaching power system reliability evaluation to
INTRODUCTION focus on the calculation of predictive indices. In the
distribution area, the usual indices are the load point failure
The techniques used in power system reliability evaluation can
be divided into the two basic categories of analytical and rate (or frequency), the average outage duration and the
average annual outage time. Normal utility practise is to
simulation methods. The analytical techniques are highly
developed and have been used in practical applications for measure distribution system performance in terms of SAIFI,
several decades [ 11. Analytical techniques represent the SAIDI etc. Extensive involvement with Canadian utilities
system by mathematical models and evaluate the reliability created the realization that the predicted load point indices
indices from these models using mathematical solutions. The could and should be extended using the customer composition
exact mathematical equations can become quite involved and to predict the conventional system indices. This is now an
approximations may be required when the system is integral part of the basic power system reliability course at the
complicated. A range of approximate techniques therefore has University of Saskatchewan. This approach does not usually
been developed to simplify the required calculations. encompass any discussion or appreciation that the customer
Analytical techniques are generally used to evaluate the mean based indices can exhibit considerable natural variability on an
or expected values of the load point and system reliability annual basis, while still having the expected values determined
indices and this is the approach normally taught in university by the analytical approach. In order to obtain this appreciation
and industry based courses. The mean values are extremely of the annual variability, it is necessary to include probability
useful and are the primary indices of system adequacy in or frequency distribution concepts in the evaluation process.
The Monte Carlo simulation approach [ 1,2,3] described in this
PE-124-PWRS-0-07-1998 A paper recommended and approved by paper provides the opportunity to develop an appreciation of
the IEEE Power Engineering Education Committee of the IEEE Power the variability associated with the annual indices.
Engineering Society for publication in the IEEE Transactions on Power
Systems. Manuscript submitted November 12, 1997; made available Reference [4] presents an analytical technique to evaluate the
for printing June 22, 1998. probability distributions associated with the basic distribution
system reliability indices. This technique can be used to

0885-8950/99/$10.00 0 1998 IEEE

Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on March 04,2023 at 10:05:10 UTC from IEEE Xplore. Restrictions apply.
 398

determine approximate probability distributions. When the using the relationships between the element states and system
system is complicated and the probability distributions of the states. The system reliability indices and their probability
component parameters are widely spread, it is difficult and &stributions are obtained from the artificial history of the
sometimes not possible to evaluate the probability system.
distributions of the reliability indices using the technique [4]
TIME SEQUENTIAL SIMULATION APPLIED TO
and the results obtained can have large errors.
DISTRIBUTION SYSTEM RELIABILITY
Increased availability of high speed computers has made EVALUATION
Monte Carlo simulation [1,2,3] a viable technique for many
Element Models and Parameters
power system reliability applications [2,5]. Many practicing
power system engineers who have had training in probability The essential requirement in time sequential sirrulation is to
and statistics have had little or no contact with Monte Carlo generate realistic artificial operatinghestoration histories of
simulation and therefore have not used it in reliability the relevant elements. These artificial histories depend on the
evaluation. This is also true for most engineering system operatinghestoration modes and tht: reliability
undergraduate or graduate students who follow a conventional parameters of the elements.
curriculum. The basic concepts of Monte Carlo simulation are Distribution system elements include transmission lines (main
covered in [3] and extended to electric power systems in [l, and lateral line sections), transformers, disconnc:ct switches,
21. A number of papers have been published [5, 8, 91 on the fuses, breakers and alternate supplies. Line r,ections and
application of Monte Carlo simulation in the reliability transformers can generally be represented by the two-state
evaluation of generating systems, transmission systems, model shown in Fig. 1 where the up state indicates that the
substations and switching stations. Distribution system element is in the operating state and the down :state implies
reliability assessment using analytical techniques is well that the element is inoperable due to failure.
established [8, 91 and new techniques [lo, 111 are being
developed. Relatively little work has been done to evaluate the r i
reliability of distribution systems using simulation techniques
[5, 8, 91. Simulation considering non-exponential down times
has been used in the distribution system reliability evaluation
in [12].
Fig. 1 State space diagram of element
The first section of this paper briefly illustrates the basic
concepts of time sequential Monte Carlo simulation. The The time during which the element remains in thi: up state is
development of the algorithm using this technique for called the time to failure (TTF) or failure time (FT). The time
distribution system reliability evaluation is described in the during whch the element is in the down state is called the
second section. A simulation program is then briefly restoration time and can be either the time to repair (TTR) or
introduced followed by a practical distribution system the time to replace (TTR).The process of transiting from the
reliability application. up state to the down state is the failure process. Transition
from an up state to a down state can be caused by the failure of
TIME SEQUENTIAL MONTE CARLO SIMULATION
an element or by the removal of elements for maintenance.
TECHNIQUE
Fig. 2 shows a simulated element operatinghestoration history.
The behavior pattern of n identical systems in real time will all
be different in varying degrees, including the number of TTR TTR
failures, times to failure, restoration times, etc. This is due to
d b T T F + 4 - -
the random nature of the processes involved. The behavior of
a particular system, could therefore follow any of these
behavior patterns. The time sequential simulation process can
u p i i r -
be used to examine and predict real behavior patterns in
Fig. 2 Element operatinghepair histoly
simulated time, to obtain the probability distributions of the
various reliability parameters and to estimate the expected or The parameters TTF, TTR are random variables and may have
average value of these parameters. different probability distributions. The probability
distributions most often used to simulate these tunes are the
In a time sequential simulation, an artificial history that shows
Exponential, Gamma, Normal, Lognormal and Poission. Fuses
the up and down times of the system elements is generated in
and breakers are used to automatically isolate failed elements
chronological order using random number generators and the
or failed areas from healthy areas. They can exist in either
probability distributions of the element failure and restoration
functioning or failed states which can be described in terms of
parameters [3]. A sequence of operating-repair cycles of the
their probabilities. Alternate supply situations can be
system is obtained fkom the generated component histories

Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on March 04,2023 at 10:05:10 UTC from IEEE Xplore. Restrictions apply.
 399

described by probabilities that alternate supplies are available. probability distributions of these indices can be calculated
A uniform distribution is used to simulate these probabilities. from the load point operatinghestoration histories.
Determination of Load Point Failures The three basic load point reliability indices usually used are
the average failure rate A, the average outage time r and the
The hnction of a distribution system is to supply electric
average annual unavailability or average annual outage time U
power to individual customers. Element failures may affect
one or more load points. The most difficult problem in the [I]. The average values of the three basic load point indices
for load pointj can be calculated from the load point up-down
simulation is to find the load points affected by the failure of
operating history using the following formulae:
an element and to determine their operatinghestoration
histories, which are dependent on the network configuration ,
h . =- Ni (1)
the system protection and the maintenance philosophy. In J
order to create a structured approach, the distribution system E Tuj
can be broken down into general segments. A complex radial =Tij
distribution system can be divided into the combination of r. =- (2)
main feeder (a feeder is connected to a switching station) and ' N .
J
subfeeders (a subfeeder is a branch connected to a main feeder
or to other subfeeders [6]. The direct search procedure for (3)
determining the failed load points and their operating-
restoration histories is as follows: Where ETuj and ET4 are the respective summations of all
(1) Determine the type and location of the failed element, the
the up times Tu and all the down times Td and N j is the
failed element number and the failed feeder number that the
failed element is connected to. number of failures during the total sampled years.
(2) Determine the affected load points connected to the failed
In order to determine the probability distributions of the load
feeder and the failure durations of these load points according
point failure frequency, the period values k of this index are
to the configuration and protection scheme of the failed
calculated for each sample year. The number of years m(k) in
feeder.
which the load point outage frequency equals k is counted.
(3) Determine the subfeeders which are the down stream
The probability distribution p(k) of the load point failure
feeders connected to the failed feeder and the effects of the
frequency can be calculated using
failed element on the load points connected to these
subfeeders. m(k)
P(k) = - k = 0, 1, 2 ...
(4)Repeat (2) and (3) for each failed subfeeder until all the M
subfeeders connected to the failed feeder are found and where A4 is the total sample years. The probability distribution
evaluated. of the load point unavailability can be calculated in a similar
(5) Determine the upfeeder which is the up stream feeder to manner. To calculate the probability distribution of outage
which the failed feeder is connected and the effects of the duration, the failure number n(i) with outage duration between
failed element on the load points in the upfeeder i-1 and i is counted. The probability distributionp(i) is
(6) Repeat ( 2 ) to (5) until the main feeder is reached and
evaluated. i = 1, 2, 3 ..
The operatinghestoration history of a load point is shown in
Fig. 3 and is conceptually similar to that of a component as where N is the total failures in the sampled years.
shown in Fig. 2. In this case, however, it is based on the The system indices can be calculated from the basic load point
operatinglrestoration histories of the pertinent elements, the indices as system indices are basically weighted averages of
system configuration and protection scheme. the individual load point values. Distributions of the system
indices therefore can also be obtained from the period load
TTR TTF TTR point indices.
Simulation Procedure
The process used to evaluate the distribution system reliability
indices using time sequential simulation consists of the
following steps:
Fig. 3 Load point operatinghestorationhistory
Step 1: Generate a random number for each element in the
System Reliability Indices and Their Distributions system and convert it into TTF corresponding to the
Distribution system reliability can be expressed in terms of probability distribution Of parameter.
load point and system indices. Both the average values and the 2: Determine the with minimum TTF,

Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on March 04,2023 at 10:05:10 UTC from IEEE Xplore. Restrictions apply.
 400

Step 3: Generate a random number and convert this number disconnect switches which can be installed in each end or both
into the repair time (RT) of the element with minimum TTF ends of the line. A lateral section usually consists of a line,
according to the probability distribution of the repair time. transformer, fuse or their combination. The disconnects, hses
Step 4: Generate another random number and convert this and alternate supplies can operate in the different modes
number into switchmg time (ST) according to the probability described earlier. The existing disconnect switckes, breakers
distribution of the switching time if this action is possible. and hses are shown in Fig. 4, but additional switches,
Step 5: ‘LJtilize the procedure described earlier under breakers and hses can be added at any location. The basic
“Determination of Load Point Failures” and record the outage data used in these studies is given in [7]. The failure rate of
duration for each failed load point. each element is assumed to be constant. The repair and
Step 6: Generate a new random number for the failed element switching times are assumed to be lognormally distributed. It
and convert it into a new TTF, and return to Step 2 if the is assumed that the standard deviations of the transmission line
simulation time is less than one year. If the simulation time is repair time, transformer replace time and switching time of all
greater than one year, go to Step 9. elements are one hour, 10 hours and 0.4 hours respectively.
Step 7: Calculate the number and duration of failures for each The simulation was performed for a period of 15000 years.
load point for each year.
33KV
Step 8: Calculate the average value of the load point failure
rate and failure duration for the sample years.
Step 9: Calculate the system indices and record these indices
for each year.
Step 10: Calculate the average values of these system indices.
Step 11: Return to Step 2 if the simulation time is less than the
specified total simulation years, otherwise output the results.
SIMULATION PROGRAM
A program has been developed to simulate the performance of
a complex radial distribution system using the time sequential
technique. Exponential, Normal, Lognormal and Gamma
distributions can be used to model the element probability
distributions associated with failure, repair and switching
times. The program can simulate main sections and lateral
sections which have different structures. The program can also
simulate the different operating models of fuses, breakers and
alternate supplies. The following load point and system
indices and their probability distributions can be calculated:
(1) Average load point failure frequency (failuredyear).
(2) Average load point failure duration (hourdyear).
(3) The probability distribution of failure durations. Fig. 4. Distribution system of the RBTS
(4) The probability distribution of the annual failure rate.
Average Value of the Load Point and System Indices
(5) The probability distribution of the annual failure duration.
(6) Average values of all system indices defined in [ 11. The average values of the load point and system indices were
(7) The probability distributions of the system indices calculated using both the analytical (A) and simulation (S)
Additional information in the form of the variation in the techniques. Table 1 shows representative load point indices.
annual failure rate and failure duration as a function of the Table 1 Comparison of the load point indices
number of samples is also developed. This is also available for Load Failure rate (occ/yr) Unavailabili
the system indices and shows the convergence of the (9 (A) I (S) I difference I (A) I (

simulation process.
SYSTEM ANALYSIS
The developed program has been used to evaluate a range of
distribution systems. The following illustrates an application
to the practical test system [7] shown in Fig. 4. This
distribution system contains 4 feeders, 3 subfeeders, 42 main
sections, 42 lateral sections and 40 load points. Each system
The average values of the system indices are shown in Table 2
segments consists of a mixture of components. A main section
for the two approaches.
can be a transmission line, a combination of line and

Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on March 04,2023 at 10:05:10 UTC from IEEE Xplore. Restrictions apply.
 401

per year is less than 0.1 and the probability of two failures per
year is less than 0.01. It can be seen in Fig. 6 that the
probability of zero failures per year is about 0.02 at load point
30 and the probability of having 6 or more outages per year is
very small. The additional information provided by a
probability distribution can be very important for those
customers which have special reliability requirements.
The results shown in Tables 1 and 2 for both the simulation The probability distributions of failure duration for load points
and analytical approaches are very close. The maximum 1 and 30 are shown in Figs. 7 and 8. A class interval width of
difference in the load point indices is 7.95 percent at load one hour has been used in this example. It can be seen in Fig.
point 20. The maximum error in the system indices is 3.18 7 that failure durations between 0 and 1 hour at load point 1
percent for SAIFI. The analytical approach provides a direct have the largest probability. Failure durations between 1 and 2
and practical technique for radial distribution system hours have the second largest probability and durations with
evaluation and is quite adequate if only the average values ofthe third largest probability are between 4 and 5 hours. Failure
durations in excess of 12 hours have a very small possibility.
the load point and system indices are required. The variability
For load point 30, outage durations between 4 and 5 hours
in the indices can be obtained from the basic data generated in
the Monte Carlo simulation. No additional computations are have the largest probability (0.38). Failure durations are
required other than collating the generated annual values intomainly distributed between 0 and 12 hours. The longest
class intervals and portraying the indices in the form of duration is about 12 hours. The information provided by these
probability or frequency distributions. probability distributions are very useful in reliability
worth/cost analysis for customers with nonlinear customer
Probability Distributions of the Load Point Indices damage functions. The average outage duration of 2.488 hours
The probability distributions of the annual failure frequency obtained using the analytical technique does not provide any
and failure duration for each load point in the distribution information on the dispersion of the actual outage durations. A
system were evaluated. Figs. 5 and 6 show the failure one hour class interval is used in Figs. 7 and 8 for illustrative
frequency histograms for load points 1 and 30. purposes. Any class interval can, however, be used in
Probability describing the distribution.
I T

Probability
n 45
04
0 35

0.3
0.25
0.2
0.15
0.1

0 I 2 t 4
0.05
0 t-9

Failure Frequency (FailureslYear) 0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4

Fig. 5 Failure frequency histogram, Load point 1 Failure Duration (Hours/Occ.)

Probability Fig. 7 Failure duration histogram, Load point 1

Probabilihi

0 1 2 3 4 5 6 7 8 9 0 I 2 3 4 5 6 7 8 9 1 0 l l 1 2 1 3 I I
Failure Duntion (HwdOee.)
Failure Frequency(Failures/Year)
Fig. 6 Failure frequency histogram, Load point 30 Fig. 8 Failure duration histogram, Load point 30

The probability distribution of failure frequency clearly shows Probability Distributions of System Indices
the probability of having a different number of load point The probability distributions of all seven system indices for
failures in each year for each load point. It can be seen in Fig. each feeder were also evaluated. Fig. 9-15 show the
5 that the probability of having zero failures per year at load probability distributions of SALFI, SAIDI, CAIDI, ASAI,
point 1 is more than 0.9. The probability of having one failure ASUI, ENS and AENS [ 11 for Feeder 4.

Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on March 04,2023 at 10:05:10 UTC from IEEE Xplore. Restrictions apply.
 402

Probability Probability

--L
014 T 0.9 T
I
012 t 08

O I t
0 08
006
0 04
1 03

---
0 02 02
0 - i 01
0 0 02 04 06 08 I I 2 14 16 I S 2 2 2 24 26 28 3 32 34 3 6 3 8 0 __?- -T

SAlFI 0 0 1 2 3 4 5 6 7 8 9 IO 11 12 13 14 15 16 17 18 19
Fig. 9 Histogram of SAIFI, Feeder 4 ENS
Probability Fig. 14 Histogram of ENS, Feeder 4
04 T
075 - Probability

0 25 06

01
0 05
0 EJL-=+LIKL 01
0 0 05 I 1 5 2 25 1 7 5 4 45 5 5 .*

0 0 0.05 0.1 0.15 0.2

SAIDI
AENS
Fig. 10 Histogram of SAIDI, Feeder 4
Fig. 15 Histogram of AENS, Feeder 4
Probability
The probability distribution of SAIFI is a combination of the
failure frequency distribution weighted by the percentage of
customers connected to the corresponding load points. The
distribution shows the variability in the average stnnual system
customer interruption frequency. The distribution of SAIDI is
the summation of the unavailability distribution weighted by
the percentage of customers connected to corresponding load
0 0 I 2 3 4 5 6 7 8 9 1 0 1 1 1 2 I 3
points. The distribution shows the probabilities of different
CAIDI average annual system customer failure duration!;. The CADI
Fig. 11 Histogram of CAIDI, Feeder 4 distribution shows the probability of different failure durations
Probability
for each customer interruption in each year. The probability
035 1 distribution of ASUI is linked to the distribution of SAIDI and
03 + provides the probability of different percentages of
unavailable customer hours in each simulation year. The
distribution of ENS is a summation of the load point
unavailability distributions weighted by the corresponding
load level and shows the probability of the different total
0 05
.--+--
energy not supplied in each year. The distribution of AENS is
0.9955 0.996 0.9965 0.997 0.9975 0.998 0.9985 0.999 0.9995 1 1 the distribution of ENS per customer. The index distributions
ASAI
supplement the average values obtained either by simulation
Fig. 12 Histogram of ASAI, Feeder 4 or analytical evaluation. They provide a pictorial
Probabilitv
representation of the distribution system reliability profile
which adds considerably to a student’s appreciation.
CONCLUSION
The appreciation of the average indices conventionally
computed and discussed when teaching distritiution system
reliability evaluation can be considerably enhanced by relating
them to the annuaI dispersion associated with these indices.
This can be done using sequential Monte Carlo simulation.
0 0 O.ooO5 0.001 0.0015 0.032 0.0025 0.003 0.0035 0.004 0.0045 This paper illustrates a time sequential Simulation technique to
ASUI evaluate complex radial distribution systems. A simulation
Fig. 13 Histogram of ASUI, Feeder 4 program has been described. The analysis of a practical test

Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on March 04,2023 at 10:05:10 UTC from IEEE Xplore. Restrictions apply.
 403

distribution system is illustrated using both analytical and (1) system average interruption frequency index, SAIFI
simulation techniques. An analytical approach evaluates the total number of customer interruptions -
average reliability indices using a set of mathematical SAIFI = total number of customers served
-- hjNi c
equations and therefore the procedure is relatively simple and C Ni
requires a relatively small amount of computer time. where hi is the failure rate and Ni is the number of customers
Simulation techniques evaluate the reliability indices by a at load points i.
series of trials and therefore the procedure is more (2) system average interruption duration index, SAIDI
complicated and requires longer computer times. A simulation sum of customer interruption durations - ui Ni
approach can, however, provide information on the load point SAIDI = -
and system indices that the analytical techniques can not total number of customers served CN
i
provide. where Vi is the annual outage time and Ni is the number of
REFERENCES customers at load point i.
1 R. Billinton and R. N. Alan, “Reliability Evaluation of Power (3) customer average interruption duration index, CAIDI
Systems”, 2nd Edition, Plenum Press, New York, 1995. sum of customer interrruption durations --c u i N i
2. R. Billinton and W. Li, “Reliability Assessment of Electric Power CAIDI =
Systems Using Monte Carlo Methods”, Plenum Press, New York, total number of customer interruptions
NiLj
1994.
3. R. Billinton and R. N. Alan, “Reliability Evaluation of where hi is the failure rate, Ui is the annual outage time and Ni
Engineering Systems”, 2nd Edition, Plenum Press, New York, 1992. is the number of customer of load point i.
4 R. Billinton and R. Goel, “An Analytical Approach to Evaluate (4) Average service availability (unavailability) index ASAZ
Probability Distribution Associated with the Reliability Indices of ASAI = customer hours of availableservice
Electrical Distribution Systems”, IEEE Trans. on Power Delivery, customer hours demanded
PWRD-I NO. 3 pp. 245-251 1986.
5 R. Billinton, R. Allan, L. Salvaderi, “Applied Reliability - Ni x 8760 - NiUi
Assessment in Electric Power Systems”, The Institute of Electrical
E N i x8760
and Electronics Engineers, Inc., New York. ASUI =I-ASAI
6. R. Billinton, P. Wang, “ A Generalized Method for Distribution where hi is the failure rate, Vi is the annual outage time, Ni is
System Reliability Evaluation”, IEEE WESCANEX, Conference the number of customers at load point i and 8760 is the
Proceedings, pp. 349-354, 1995.
number of hours in a calendar year.
7. R.N. Allan, R. Billinton, I. Sjamef, L. Geol, and K. S. So, “ A
Reliability Test System for Educational Purpose-Basic Distribution ( 5 ) Energy not supplied indices, ENS
System Data and Results”, IEEE Trans. on Power Systems Vol. 6, ENS = total energy not supplied by the system = La(ilUi
No. 2, May, 1991.
8. R. N. Allan, R. Billinton, A. M. Breipohl, and C. H. Grigg, where La(i) is the average load connected to load point i.
“Bibliography on the Application of Probability Methods in Power (6) Average energy not supplied indices, AENS
System Reliability Evaluation”, IEEE Trans. on Power Syst., Vol. 9,
- *

No. 1, pp. 41-49, Feb. 1994. AENS = total energy not supplied - zLa(i)ui
9. R. N. Allan, R. Billinton, A. M. Breipohl, and C. H. Grigg, total number of customers served - C Ni
“Bibliography on the Application of Probability Methods in Power where L . is the average load connected t o load point i.
System Reliability Evaluation”, to be published on IEEE Trans. on 4)
Power Systems.
IO. R.E. Brown, S. Gupta, R.D. Christie, S.S. Venkata, R. Fletcher, BIOGRAPHIES
“Distribution System Reliability Assessment Using Hierarchical Roy Billinton (Fellow, 1978) came to Canada from England in 1952.
Markov Modeling”, IEEE Trans. On Power Systems Vol. 11, No. 4, Obtained B.Sc. and M.Sc. Degrees from the University of Manitoba
May 1996. and Ph.D. and D.Sc. Degrees from the University of Saskatchewan.
1 I . C. Chen, J. Chen, “A Neural Network Approach for Evaluating
Joined the University of Saskatchewan in 1964. Presently Associate
Distribution System Reliabilty”, Electric Power System Research,Dean, Graduate Studies, Research and Extension of the College of
1993. Engineering. Author of papers on power system analysis, stability,
12. S. Asgarpoor, M.J. Mathine, “ Reliability Evaluation of economic system operation and reliability. Author or co-author of
Distribution System with Non-Exponential Down Time”, IEEE eight books on reliability. Fellow of the IEEE, the EIC and the Royal
Trans. On Power Systems Vol. 12, No. 2, May 1997. Society of Canada and a Professional Engineer in the Province of
Saskatchewan.
APPENDIX Peng Wang was bom in P. R. China. Received his BSc. Degree
The three basic load point indices and system indices are from Xian Jiaotong University in 1978. Obtained M.Sc. from
defined in [l]. The system indices have been used by Taiyuan University of Technology 1987, MSc. from the University
Canadian electric power utilities for many years. These indices of Saskatchewan in 1995 and is now working toward a Ph.D. degree.
are as the follows.

Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on March 04,2023 at 10:05:10 UTC from IEEE Xplore. Restrictions apply.