Reliability Engineeringand System Safety 46 (1994) 3-13

© 1994 Elsevier Science Limited

Printed in Northern Ireland. All rights reserved
ELSEVIER 0951-8320/94/$7.00

Power system reliability assessment---A

conceptual and historical review
Ron Allan
Manchester Centre for Electrical Energy, University of Manchester Institute of Science and Technology, Manchester, UK,
Mro 1QD

This paper provides a review of the conceptual aspects and historical

developments relating to power system reliability, of the factors affecting it, of
the underlying concepts concerning its assessment, and of the various criteria
used in the different functional zones. It considers all areas of a power system
including generation, transmission and distribution.

1 BACKGROUND complete system including distribution. These are

discussed separately in this paper.

Reliability is an inherent characteristic and a specific

measure of any component, device or system which 2 HLI STUDIES
describes its ability to perform its intended function.
The reliability measures used in a power system 2.1 Requirements at HLI
indicate how welt the system performs its basic
function of supplying electrical energy to its During planning, it is necessary to determine how
customers. Reliability levels are interdependent with much generating capacity needs to be installed in
economics 1 since increased investment is necessary to order to satisfy the expected demand in the future and
achieve increased reliability or even to maintain to provide sufficient reserve to perform corrective and
reliability at current and acceptable levels. The preventive maintenance. The ability to move the
concepts of reliability economics are not the subject of energy to bulk supply points (BSPs) or customers is
this paper although it is important to recognize that not considered at this stage. Historically, the reserve
reliability and economics must be treated together in capacity has been set equal to a percentage of the
order to perform objective cost-benefit studies. These expected load, or equal to one or more largest units,
concepts of reliability are not new and a continuous or a combination of both. These deterministic criteria
stream of relevant papers have been published since are now being largely replaced by probabilistic
the 1930s. z-6 A selection of the most distinctive papers methods that respond to the stochastic factors
appear in Ref. 7. influencing the reliability of the system.
Historically, power system reliability has been During operation, the actual amount available may
assessed using deterministic criteria, techniques and be less than the total installed capacity due to
indices, and it is only relatively recently that maintenance or other operational problems. Probabil-
techniques, data and computational resources have istic techniques exist for assessing how much capacity
reached the stage where probability methods can be to commit to actual service. However, few utilities use
applied to practical planning and operational decision these techniques at the present time.
making.
The functional zones of a power system (generation, 2.2 Probabilistic criteria and indices
transmission and distribution) can be divided8 into the
three hierarchical levels shown in Fig. 1. The first level The probabilistic criteria and indices that can be used
(HLI) relates to generation facilities, the second level in HLI planning studies include the following.
(HLII) refers to the integration of generation and Loss of load probability (LOLP) is the oldest and
transmission, and the third level (HLIII) refers to the most basic probabilistic index. It is the probability
4 Ron Allan

utilities. It is again directly related to LOEE which is

normalised by peak demand instead of by total energy
gen~ranon hierarchical level I demanded. The weakness is that it introduces an
facilities HLI index having time as the units. This is not real time
but would have been the annual unavailability if all
energy interruptions only occurred at peak load. In
u'ansmission hierarchical level II reality, the annual unavailability is greater than the
facilities HLII values given by SM.
I It must be stressed that all the above measures are
expectations, i.e. they are not deterministic values but
distribution hierarchical level I11 only the average value of a probability distribution.
facilities ] IIL Ill They provide very valuable indicators of the adequacy
of a system taking into account the stochastic and
deterministic characteristics of the generation system
Fig. 1. Hierarchical levels. and customer demands.

that the load will exceed the available generation. Its 2.3 Evaluation techniques
weakness is that it defines the likelihood of
encountering trouble but not the severity. Therefore it The above indices can be evaluated using either
cannot recognise the degree of capacity or energy analytical or simulation approaches. T M Analytical
shortage. techniques evaluate the reliability indices from a
Loss of load expectation (LOLE) is now the most mathematical model using mathematical solutions T M
widely used probabilistic index in deciding future whereas simulation techniques, often known as Monte
generation capacity. It is the average number of days Carlo simulation, estimate the indices by simulating
on which the daily peak load is expected to exceed the the actual process and random behaviour of the
available generating capacity. Alternatively it may be system. 9'11 These techniques can themselves be
the average number of hours for which the load is divided into non-sequential and sequential. Non-
expected to exceed the available capacity. This sequential simulation considers each time interval
concept implies a physical significance not forthcom- independently and therefore can not model time
ing from the LOLP, although the two values are correlations or sequential events. The sequential
directly related. It has the same weaknesses that exist approach, however, takes each interval (usually 1 h) in
in LOLP. chronological order.
Loss of energy expectation (LOEE) or expected Simulation is not needed generally to analyse
energy not supplied (EENS) are defined as the thermal systems although sequential simulation is very
expected energy that will not be supplied due to those useful in assessing systems having a time-dependent
occasions when the load exceeds the available history such as hydro-systems containing reservoirs
generation. It is presently less used than LOLE but is and pumped storage. 12 Both sequential and non-
a more relevant index because it encompasses severity sequential simulation are very useful in modelling
of the deficiencies as well as their likelihood. It is more complex systems, such as at the HLII level.
likely to grow in popularity with increasing energy
limitations and increased environmental controls. 2.4 Historical developments at HLI
Frequency and duration ( F & D ) is an extension
of LOLE and identifies expected frequency of The area of HLI assessment is the oldest and most
encountering deficiencies and their expected duration. extensively studied. Hence the greatest number of
It therefore contains additional physical characteristics papers have been published in this area. The first set
but, although widely documented, 2-6 is not used in of papers 13-17 envisaging the application of probability
practice. This is due mainly to the need for additional techniques appeared in the 1930s. These can be
data and greatly increased complexity of the analysis considered pioneering papers and it should be noted
without having any significant effect on the planning that they preceded the use of reliability techniques in
decisions. other applications such as military, which are often
Energy index of reliability (EIR) or unreliability given the credit of pioneer status.
(EIU) are directly related to LOEE which is A significant set of papers that added impetus to the
normalised by dividing by the total energy demanded. application of probability theory appeared in 1947
This ensures that large and small systems can be including major contributions by Calabrese and
compared on an equal basis and chronological changes others. 18-21 The 'Calabrese' method forms the basis of
in a system can be tracked. the loss of load approach which is still the most widely
System minutes (SM) is used by a number of used probability technique at HLI. This basic method
 Power system reliability assessment--a review 5

has been extended to include the loss of energy proposed peaking unit model is described in Ref. 42.
approach. Another significant technique is the This and other derivatives are now widely used.
frequency and duration method, the most important Models and evaluation techniques are still being
development being the application of the recursive developed for HLI studies. The IEEE reliability test
techniques. 22-26 This method is still little used in systems (IEEE RTS) were developed 43 to test such
practical HLI assessments. techniques. This RTS is extended in Ref. 44, which
All the above techniques are based on the analytical also includes important benchmark results against
approach. The alternative is (Monte Carlo) simula- which other methods can be compared.
tion. There has been a tendency for North America to There are relatively few papers dealing with the
use the former and for Europe and South America to operating reserve problem, mainly because few
use the latter. This distinction is now less clear but utilities have felt the need for using probabilistic
originally reflected the type of systems being studied techniques in this time domain. This is likely to
and their requirements. An early example of change with time, although not necessarily in the near
simulation is presented in Ref. 27. More recent papers future. Two basic solution techniques exist, the PJM
include Ref. 28, in which it was proposed that load method and the security function approach. The basic
and generation should not be treated independently PJM method, presented in Ref. 45, evaluates the
but in a related manner. Other pioneers of simulation probability of the committed generation just satisfying
include ENEL (Italy) and EdF (France). Papers from or failing to satisfy the expected demand during the
these organisations include Refs 29 and 30 and many period of time that generation cannot be replaced,
others in various CIGRE publications. known as the lead time. An application of this method
Initially, most applications related to single systems is given in Ref. 46. The security function approach is
and their capacity requirements, as illustrated in Ref. described in Ref. 47. This evaluates the probability of
31. However, utilities recognised the significant breaches of security including inadequate spinning
economic benefits that can be derived by sharing reserve using the concepts of conditional probability.
reserves, and this encouraged the development of The PJM method has been extended in several
analyses in interconnected systems, though still at respects. The inclusion of rapid start and hot reserve
HLI. Examples include Refs 32-34, which address the units is described in Ref. 48. These developments
loss of load approach, and Ref. 35 which addresses the were modified and extended in Ref. 49. The PJM
frequency and duration approach. method was also extended by application to
The majority of papers relating to analytical interconnected systems. 5°'51
techniques considered systems without energy limita- HLI studies are an important area and extensions,
tions. This is usually acceptable for thermal systems modifications and new algorithms are and will
but less so for hydro systems and those using continue to be published regularly.
renewable energy sources. Energy limitations are
relatively easy to include in simulation tech- 2.5 Capacity planning in practice
niques. 12"3°'36 However, several analytical papers have
addressed this problem including Ref. 37 and more Risk increases as the load increases. A typical
recently for systems containing wind energy characteristic 11 for the IEEE RTS is shown in Fig. 2. If
s o u r c e s . 38,39 a maximum risk index is specified, then the maximum
Most analytical techniques are based on Calabrese peak load that can be supported by the generation
approach in which the generation model, represented system can be determined. This value is known as the
by a capacity outage probability table, is constructed peak load carrying capability (PLCC) shown in Fig. 2
using a recursive state enumeration technique. 1° for a specified risk level of RL. The variation of PLCC
However, alternative methods do exist. One is the with RL can be used as a measure to determine by
cumulant method first proposed for power system how much the load can be allowed to grow without
reliability application in Ref. 40. This technique has creating excessive risk. This concept can be extended
been widely used for adequacy assessment, production and applied to the study of alternative expansion
costing, and maintenance scheduling. Care is needed plans.
because considerable errors can arise including The use of these techniques varies considerably
apparent negative probabilities, under certain circum- around the world. In the UK, there is no central body
stances. Another approach is based on the use of fast since privatisation that is responsible for deciding
Fourier transforms 41 which, although slower computa- when additional capacity is required or how much.
tionally than the cumulant method, does not exhibit These decisions are left to individual private
the error problems. generators in response to market forces. The values of
The models for generating units are often simple LOLP and VOLL (value of lost load) embedded in
ones representing base loaded units. These are the pool pricing mechanism 52 are intended to be
inapplicable and pessimistic for peaking units and a indicators that encourage or discourage the installa-
6 Ron Allan

It should be noted that virtually all of the

techniques available at this time relate to adequacy
assessment: probabilistic security assessment is very
much in the research domain. Even with this
restriction, planning decisions can be improved
1'0 "~ considerably with the additional objective information
"o
derived from adequacy assessments.

3.2 ProbabiHstic criteria and indices

RL

Several HLII reliability indices can be calculated,

0.1
including system indices and load point indices. These
are complementary, not alternatives, and each serves
I
| an entirely different purpose. A typical set of load
i I i i !
2,4 / 2.6 2.8 3.0 3.2 point indices and system indices are listed in Figs 3
and 4.
p e a k load. G W
System indices are extremely valuable for decisions
PLCC regarding global observations and overall energy
management. They can be used to track the
Fig. 2. Effect of peak load on LOLE of IEEE RTS. 11
chronological changes in system behaviour, predict
and monitor the result of changing system operational
strategies, and compare the performance of different
tion of additional generation by these private systems and different areas within a system. They are
generators. It is too early to know how this will work. usually inappropriate for identifying the effect of
This process is very different to that operated in individual reinforcement schemes, particularly in the
Canada where each utility is essentially a provincial case of large practical systems when the change in
monopoly. Provincial regulatory bodies require their values resulting from a single reinforcement
decisions relating to capacity expansion plans to be scheme is usually very small compared with other
made on objective bases which have encouraged all contributions. However, this can be objectively
Canadian utilities to use some form of probabilistic gauged from the load point indices. The main problem
reliability assessments; 53 the criteria and indices vary in evaluating load point indices is where to allocate
considerably but the basic concept is the same. system deficiencies, e.g. at which BSP should energy
be curtailed if curtailment is needed. These may be
difficult questions but are still necessary if objective
3 HLII STUDIES decisions and cost-benefit analyses are to be made.

3.1 Requirements at HLIi 3.3 Evaluation techniques

Assessment at HLII is complex since it must consider The two main approaches, 9-1~ analytical and simula-
the integrated reliability or composite effects of tion, are both used in the adequacy assessment of
generation and transmission. These two entities composite systems. Both techniques assess the system
cannot be analysed separately at this level. To do so
could create misleading results and conclusions. This BASIC VALUES
does not mean they have to be owned by the same probability of failure
expected fitquency of failure, f/yr
company, but it is essential that one body has the role expected onmbcs of voltage violations
expented Io~l curUtiled, MW
of coordinating the planning and operation. expected energy not supplied, MWh
¢XlmCted duration of 1o~1 curtailment, h
The function of a composite system is to produce
electrical energy at the generation sources and then to MAXIMUM VALUES
maximum lcQd curtailed, MW
move this energy to the BSPs without violating system maximum energy curtailed, MWh
maximum duration of loed curtailment, h
operational constraints. In this case, the bulk
AVERAGE VALUES
transmission facilities must not only provide adequate average Iced curmiled/curmilmeat, MW/curmilment
transmission capacity to ensure the demand is satisfied average ¢am'gy not supplied/curtailment, MWh/curtailonmt
average duration of curtailment/curtailnumt, h/curtailment
and that voltage, frequency and thermal limits are
BUS ISOLATION VALUES
maintained (defined as adequacyS), but must also be exix~tednumberof cumilmmts
capable of maintaining stability following fault, exl~eted load curtailed, MW
exlat~2t~ emrgy not mpplied, MWh
switching and other transient disturbances (defined as h
expe~ed duration of load curtailment,
securityS). Fig. 3. Typical load point indices.
 Power system reliability assessment--a review 7

BASIC VALUES amined in Refs 69-72. The quantitative results show

bulk power interntption index (BPII), MW/MW.yr
bulk power supply average MW eurtailment/disturbatace (BPSACI), MW/disturbance that station-originated events can create significant
bulk power energy curtailment index (BPECI), MWh/MW.yr
modified bulk power energy curtailment index (MBPECI)
increases in the load point and system adequacy
system minuteS, m indices. The impact of dependency effects due to
AVERAGE VALUES common-cause and weather-related outages are
average number of curtailments/load point
average load cut,tiled/load point, MW/Ioad point illustrated in Refs 73 and 74.
average e~rgy cortailed/lmd point, MWh/load point
average duration of load curtailed/load point, h/load point
The concept of adequacy and security was first
average number of voltage violations/load point presented in Ref. 75. An overview of these aspects
MAXIMUM VALUES together with other major concepts in HLII adequacy
maximum system load curtailed under any contingency condition, MW assessment are given in Refs 76-78. The first two of
maximum system energy not supplied under any contingency condition, MWh
Fig. 4. Typical system indices. these papers reviewed the goals, the time frame, and
the methods of evaluation as well as the conflicting
opinions that sometimes seem to occur between North
effect caused by contingencies involving single and American and European planners. The third paper
multiple outages. These outages may be due to described the philosophy rather than the modelling
failures of the generators and transmission lines details.
themselves or may be outaged due to the failure of Early work in measuring and reporting overall
other system components. The most important of reliability at HLII is described in Ref. 79. This was
these outages are listed below. extended in a three-stage activity sponsored by the
• Independent outages IEEE Subcommittee on the Application of Probability
• Dependent outages Methods. The first stage 8° identified techniques and
• Common mode outages approaches for monitoring and measuring reliability
• Station-originated outages indices in the operation of systems. The second
• Weather-related outages stage 81 identified predictive indices that could be
calculated, the degree to which they are used in
3.4 Historical developments at HLII practice, commonality and differences between uti-
lities, perceived problems in their practical use and
Considerable effort has been expended during the last management understanding. The third stage 82 iden-
two decades in developing techniques and criteria for tified the future needs and requirements of reliability
HLII assessment. Two main approaches exist, assessments and how these can be resolved by future
analytical (state enumeration) and (Monte Carlo) reliability assessment techniques and approaches.
simulation. Progress in the area of HLII evaluation has been
The initial concepts of the analytical technique is relatively slow as many conceptual, modelling and
illustrated in Refs 54, 55 which introduced the computational difficulties have had to be resolved.
reliability criteria based on the application of Many utilities still apply a deterministic approach such
conditional probability. The use of the indices in as the (n - 1 ) criterion. However, utilities around the
systems planning is illustrated in Ref. 56. Additional world are now in the transition phase from the
examples of alternative analytical approaches are traditional methods to the more modern probabilistic
presented in Refs 57-60. approaches. This will escalate as they recognise the
The (Monte Carlo) simulation used by ENEL is benefits to be gained and thus an increasing number of
described in Refs 61 and 62. Extensive work has also relevant models and techniques will continue to be
been done in France, 63 and more recently in Brazil 64 developed.
and the UK/Spain. 65
There has been considerable debate about the 3.5 HLII studies in practice
relative merits of the two approaches. Comparisons
between them using different computer programs are It is pertinent at this stage to illustrate the type of
described in Ref. 66 for the IEEE RTS and in Ref. 67 results that can be obtained using a simple illustrative
for the New Brunswick System. These provide a example taken from Ref. 10 which gives a full
deeper perspective of the individual merits and description of the data and the results. A summary of
differences and show that both have a part to play in the results is shown in Figs 5 and 6. These clearly
system analysis. Also, a recent application 68 of the illustrate the effect that various reinforcement
simulation approach to the RTS illustrates how schemes have on individual bus indices, for instance,
reliability evaluations can be used in the actual note the large change in frequency of interruptions at
planning at HLII to obtain more ecrnomical systems bus 5 when particular lines are added. The benefit of
structures. each reinforcement can therefore be objectively
The impact of terminal station failure events assessed. In this example, some, but not all, of the
(breakers, transformers, bus sections, etc.) is ex- system indices change quite significantly following
8 Ron Allan

10
frequency, f / y r individual customers to be evaluated and compared
against relevant design and operational criteria. This is
usually impractical because of the enormity of the
problem. Instead the assessment is usually done for
the distribution functional zone only. This is
0.1
acceptable because distribution networks generally
interface with the transmission system through one
0.01
supply point. Therefore the load point indices
evaluated in the HLII assessments can be used as
0.001
bus 2 bus 3 bus 4 bus 5 input values for the reliability evaluation of a
distribution system to give overall HLII indices. Also
i bale case i line 7 added
[-"7 linec 7 • 8 added ii common mode on I • 6
distribution systems are generally the major cause for
the outages seen by individual customers and
Fig. 5. Load point indices (frequency) for five-bus system. 1° therefore dominate the overall reliability indices. An
illustration of this effect is shown in Fig. 7; 83 the
relative effect being similar for most systems.
individual reinforcements. This is because the system The technical function of a distribution system is to
is very small and individual busbars make a significant take energy from BSPs and deliver it to individual
contribution to the global value. customers within certain quality constraints of voltage,
At present, HLII studies are not used extensively in frequency, harmonics, flicker, etc. It is also expected
practice. However, the interest has changed dramati- to achieve this with a reasonable level of reliability,
cally in recent times and such studies are likely to i.e. to keep the number and duration of outages
become of significant importance in the near future. It reasonably low. This can be quite difficult to achieve
should be remembered that they are very valuable economically particularly at the lower voltage levels
in comparing alternatives such as reinforcements, and in rural areas, because the system generally
maintenance schedules, operating strategies, etc. It is consists of single radial overhead lines which are
worth noting that individual utilities or regulatory exposed to adverse environmental conditions. They
bodies may need alternative indices in order to reflect are therefore prone to failure and frequently lengthy
particular system conditions and requirements, and outage times.
that they may require only a few, even one, for their Although these naturally occurring conditions
decision-making process. cannot themselves be avoided, the adequacy of the
supply to customers should be considered by

4 HLIII STUDIES contributor

gen and trans system

4.1 Requirements at HLIII

132kV system
Consideration of HLIII would enable the effect of
generation, transmission and distribution on
33/66kV systems
system index
1000 I

100~ 6.6/11kV systems

=
7
low voltage system

0.1

0.01
arranged shutdowns

i i i ~ i i
0.001
BPII BPECI BPSACl system minutes 0 10 20 30 40 50 60 70

B b i l l Clal I common mode on 1 • 6 ~ lined 7 & 8 added unavailability, mins/yr

Fig. 6. System indices for five-bus system. ]° Fig. 7. Typical customer unavailabilities.13
 Power system reliability assessment---a review 9

assessing objectively the effect of the available require realistic component data that includes relevant
alternatives including alternative reinforcement sch- failure rates and restoration times. This is not easily
emes, allocation of spares, improvements in main- obtained from some fault reporting schemes which
tenance policy, alternative operating policies. record information only when customers are inter-
rupted and not for equipment or component failures
4.2 Probabilistic criteria and indices when customer outages do not occur.
The usual method for evaluating the reliability
Most utilities collect measures of how distribution indices is an analytical approach 1° based on a failure
systems perform during the operational phase. modes assessment and the use of equations for series
Historically these are customer-related measures and parallel networks. 9 A simulation approach is
evaluated from system interruption data. The basic sometimes used for special purposes in order to
indices are failure rate, A, average outage duration, r, determine, for instance, the probability distributions
and annual unavailability, U, at individual load points. of the reliability indices, u As with HLII studies, the
A set of system indices can also be deduced. The effect of different outage types and weather effects are
terms vary but are conceptually the s a m e a s 1° studied.
• System average interruption frequency index
(SAIFI) 4.4 Historical developments in distribution systems
• System average interruption duration index
(SAIDI)
There have been considerable historical developments
• Customer average interruption frequency index
in the transmission and distribution functional zones.
(CAIFI)
However the transmission zone is now generally
• Customer average interruption duration index
absorbed into HLII because it is an integral part of
(CAIDI)
composite system assessment. The techniques de-
• Average service availability index (ASAI)
veloped for transmission and distribution are now
• Average (expected) energy not supplied
more applicable to the assessment of distribution
(AENS/EENS)
systems only.
These indices are excellent measures for assessing how Quantitative assessment techniques began sig-
well a system has performed its basic function of nificantly with Refs 84 and 85 which presented
satisfying the needs of its customers. The indices can important concepts. The aspect of failure bunching in
be calculated for the overall system or for subsets of parallel facilities due to storm associated failures is
the system depending on the requirements for the introduced in Ref. 84. A major contribution was the
performance measures. procedures for calculating the basic indices of failure
frequency (failure rate) and average duration in
4.3 Evaluation techniques addition to probability of failure. These indices are
now widely used for assessing distribution and
The same range of indices for future performance can customer reliability. A practical application of the
also be predicted. Theoretically this is relatively techniques available at that time is presented in Ref.
simple and straightforward 1° often not requiring 86.
complex computer programs. However, it does The application of Markov processes to transmis-
sion evaluation is illustrated in Ref. 87 which
considered the effect of adverse weather induced
- - 30
failures on simple parallel systems and compared the
25
calculated results with those of Ref. 84. The concept
of a consistent set of equations for series/parallel
20
system reduction including adverse weather and
permanent, temporary, maintenance and overload
O - - 03
outage modes is considered in Refs 88 and 89. The
,< latter also illustrates the concept of using minimal cut
if)
10 sets.
h
The incorporation of switching actions in the
assessment of transmission systems including protec-
tive elements is introduced in Ref. 90 and formalised
base 0 in Ref. 91 which presents a basic three-state model
+ fuses + isolators + backfeed
incorporating the switching-after-fault concept. This
study c a s e
basic framework is used in Ref. 92 to create a
Fig. 8. Results for radial system.9 procedure for evaluating substation and switching
10 Ron Allan

station reliability. It also introduces the concept of we expect the two aspects, reliability and worth of
active and passive faults in systems containing supply, to become of significant importance in the
protective elements. These concepts are extended for very near future.
more general conditions in Ref. 93. The philosophy of
these techniques is illustrated in Ref. 94 for the
auxiliary systems of power stations. REFERENCES
The concepts of distribution system reliability
evaluation is extended in Ref. 95 by incorporating 1. Allan, R. N. & Billinton, R., Probabilistic methods
operational constraints such as load partial loss of applied to electric power systems--are they worth it?
lEE Power Engng J., May (1992) 121-9.
continuity and load transfers. The effect of depen- 2. Billinton, R., Bibliography on the application of
dency and common mode failures is addressed in Ref. probability methods in power system reliability
96. The concepts of minimal cut sets and the evaluation. IEEE Trans., PAS-91 (1972) 649-60.
associated equations to include the effects of common 3. IEEE Subcommittee Report, Bibliography on the
mode failures in parallel and meshed systems is application of probability methods in power system
reliability evaluation, 1971-1977. IEEE Trans., PAS-
included in Ref. 97. 97 (1978) 2235-42.
Data is of great concern in all predictive reliability 4. Allan, R. N., Billinton, R. & Lee, S. H., Bibliography
assessments. The approach used to obtain data by one on the application of probability methods in power
reporting group is described in Ref. 98 and this system reliability evaluation, 1977-1982. IEEE Trans.,
provides an illustration of the data that can be PAS-103 (1984) 275-82.
5. Allan, R. N., Billinton, R., Shahidehpour, S. M. &
collected. One difficulty in developing data collection Singh, C., Bibliography on the application of
schemes is the definition of terms and indices for probability methods in power system reliability
reporting and analysing outages. Recent work evaluation, 1982-1987. IEEE Trans. Power Systems, 3
regarding this is reported in Ref. 99 and finalised as (1988) 1555-64.
IEEE Standard 859-1987.1°° 6. Allan, R. N., Billinton, R., Briepohl, A. M. & Grigg,
C. H., Bibliography on the application of probability
The development of data collection systems and methods in power system reliability evaluation,
analytical techniques that can be used in planning, 1987-1991. IEEE Winter Power Meeting, Columbus,
design and operation of distribution systems is February 1993, paper 93 WM 166-9-PWRS.
expected to continue and to provide important system 7. Billinton, R., Allan, R. N. & Salvaderi, L. (eds),
and customer benefits. Applied Reliability Assessment in Electric Power
Systems. IEEE Press, New York, USA, 1991.
8. Billinton, R. & Allan, R. N., Power system reliability
in perspective. IEE J. Electronics Power, 30 (1984)
4.5 HLIII studies in practice 231-6.
9. Billinton, R. & Allan, R. N., Reliability Evaluation of
Engineering Systems: Concepts and Techniques (2nd
edn). Plenum Publishing, New York, USA, 1992.
Consider a simple radial system with four teed-off 10. Billinton, R. & Allan, R. N., Reliability Evaluation of
laterals 1° feeding individual load points. The basic Power Systems. Plenum Publishing, New York, USA,
function can be achieved by solid teed-points and no 1984.
isolators. This arrangement is perfectly adequate if no 11. Billinton, R. & Allan, R. N., Reliability Assessment of
failures occur. This however is not realistic and Large Electric Power Systems. Kluwer Academic
Publishers, Boston, MA, USA, 1988.
protection devices and isolators are usually installed in 12. Allan, R. N. & Roman Ubeda, J., Reliability
order to improve system reliability. Consider the assessment of hydro thermal generation systems
possibility of installing fuses at the tee points, isolators containing pumped storage plants. Proc. lEE, 138(part
at section points and a backfeed. This produces the C) (1991) 471-8.
results 1° shown in Fig. 8. Each reinforcement produces 13. Lyman, W. J., Fundamental consideration in preparing
master system plan. Electrical World, 101(24) (1933)
a further improvement in the reliability indices. The 788-92.
question is 'whether the improvement is worth it?', 14. Smith Jr, S. A., Spare capacity fixed by probabilities of
which again relates to reliability economics. 1 outage. Electrical World, 103 (1934) 222-5.
Many distribution systems are still designed 15. Smith Jr, S. A., Service reliability measured by
according to deterministic standards. These views are probabilities of outage. Electrical World, 103 (1934)
371-4.
changing quite significantly and there is not a positive
16. Smith Jr, S. A., Probability theory and spare
awareness of the need to assess system design equipment. Bull. Edison Electrical Institute March
alternatives in a probabilsitic sense. There is also a (1934).
rapidly growing appreciation, inside and outside the 17. Benner, P. E., The use of theory of probability to
industry, of the need to account for customers' determine spare capacity. General Electric Rev., 37(7)
(1934) 345-8.
expectations and their assessment of the worth of
18. Calabrese, G., Generating reserve capability deter-
supply. Since the latter cannot be objectively assessed mined by the probability method. AIEE Trans. Power
without adequate and objective reliability measures, Apparatus Systems, 66 (1947) 1439-50.
 Power system reliability assessment---a review 11

19. Lyman, W. J., Calculating probabiity of generating 35. Billinton, R. & Singh, C., Generation capacity
capacity outages, A I E E Trans. Power Apparatus reliability evaluation in interconnected systems using a
Systems, 66 (1947) 1471-7. frequency and duration approach: Part I--
20. Seelye, H. P., Outage expectancy as a basis for mathematical analysis. IEEE Trans. Power Apparatus
generator reserve. AIEE Trans. Power Apparatus Systems, PAS-90(4) (1971) 1646-54.
Systems, 66 (1947) 1483-8. 36. Insinga, F., Invernizzi, A., Manzoni, G., Panichelli, S.
21. Loane, E. S. & Watchorn, C. W., Probability methods & Salvaderi, L., Integration of direct probabilistic
applied to generating capacity problems of a combined methods and Monte Carlo approach in generation
hydro and steam system. A I E E Trans. Power planning. Proc. 6th Power System Computational
Apparatus Systems, 66 (1947) 1645-57. Conf., 1978. IPC Science and Technology Press,
22. Hall, J. D., Ringlee, R. J. & Wood, A. J., Frequency Guildford, UK, pp. 48-58.
and duration methods for power system reliability 37. Billinton, R. & Harrington, P. G., Reliability
calculations: Part I--generation system model. IEEE evaluation in energy limited generating capacity
Trans. Power Apparatus Systems, PAS-87(9) (1968) studies. IEEE Trans. Power Apparatus Systems,
1787-97. PAS-97(6) (1978) 2076-85.
23. Ringlee, R. J. & Wood, A. J., Frequency and duration 38. Allan, R. N. & Corredor-Avella, P., Reliability and
methods for power system reliability calculations: Part economic assessment of generating systems containing
II----demand model and capacity reserve model. IEEE wind energy sources. Proc. lEE, 132(1) (1985) 8-13.
Trans. Power Apparatus Systems, PAS-88(4) (1969) 39. Billinton, R. & Chowdury, A. A., Incorporation of
375-88. wind energy conversion systems in conventional
24. Galloway, C. D., Garver, L. L., Ringlee, R. J. & generating capacity adequacy assessment. Proc. IEE,
Wood, A. J., Frequency and duration methods for 139 (1992) 47-56.
power system reliability calculations: Part III-- 40. Rau, N. S. & Schenk, K. F., Application of Fourier
generation system planning. IEEE Trans. Power methods for the evaluation of capacity outage
Apparatus Systems, PAS-88(8) (1969) 1216-23. probabilities. IEEE Winter Power Meeting (1979)
25. Cook, V. M., Ringlee, R. J. & Wood, A. J., Frequency Paper No. A 79 103-3.
and duration methods for power system reliability 41. Allan, R. N., Leite de Silva, A. M., Abu-Nasser, A. A.
calculations: Part IV--models for multiple boiler- & Burchett, R. C., Discrete convolution in power
turbines and for partial outage states. IEEE Trans. system relliability. IEEE Trans. Reliab., R-30 (1981)
Power Apparatus Systems, PAS-88(8) (1969) 1224- 452-6.
32. 42. IEEE Task Group on Model for Peaking Units of the
26. Ringlee, R. J. & Wood, A. J., Frequency and duration Application of Probability Methods Subcommittee, A
methods for power system reliability calculations: Part four-state model for estimation of outage risk for units
V--models for delays in unit installations and two in peaking service. IEEE Trans. Power Apparatus
interconnected systems. IEEE Trans. Power Apparatus Systems, PAS-91(2) (1972) 618-27.
Systems, PAS-90(1) (1971) 79-88. 43. IEEE Subcommittee on the Applications of Probabi-
27. Baldwin, C. J., Gaver, D. P. & Hoffman, C. H., lity Methods, IEEE reliability test systems. IEEE
Mathematical models for use in the simulation of Trans. Power Apparatus Systems, PAS-98(6) (1979)
power generation outages: I--fundamental considera- 2047-54.
tions. A I E E Trans. Power Apparatus Systems, 78 44. Allan, R. N., Billinton, R. & Abdel-Gawad, N. M. K.,
(1959) 1251-8. The IEEE reliability test system---extensions to and
28. Patton, A. D., Blackstone, J. H. & Balu, N. J., A evaluation of the generating system. IEEE Trans.
Monte Carlo simulation approach to the reliability Power Systems, PWRS-I(4) (1986) 1-7.
modeling of generation systems recognizing operation 45. Anstine, L. T., Burke, R. E., Casey, J. E., Holgate, R.,
considerations. IEEE Trans. Power Systems, 3(3) John R. S. & Stewart, H. G., Application of
(1988) 1174-80. probability methods to the determination of spinning
29. Salvaderi, L. & Paris, L., Pumped storage plant basic reserve requirements for the Pennsylvania-New
characteristics: Their effect on generating system Jersey-Maryland interconnection. IEEE Trans. Power
reliability. Proc. Am. Power Conf., 35 (1974) 403-18. Apparatus Systems, PAS-68 (1963) 726-35.
30. Manzoni, G., Noferi, P. L. & Voltorta, M., Planning 46. Leffler, L. G., Cucchi, G. A., Ringlee, R. J., Reppen,
thermal and hydraulic power systems--relevant para- N. D. & Chambless, R. J., Operating reserve and
meters and their relative influences. CIGRE Paper generation risk analyses for the PJM interconnection.
32-16 (1972). IEEE Trans. Power Apparatus Systems, PAS-94(2)
31. Halperin, H. & Adler, H. A., Determination of (1975) 396-407.
reserve-generating capability. AIEE Trans. Power 47. Patton, A. D., A probability method for bulk power
Apparatus Systems, 77 (1958) 530-44. system security assessment--I: basic concepts. IEEE
32. Watchorn, C. W., The determination and allocation of Trans. Power Apparatus Systems, PAS-91(1) (1972)
the capacity benefits resulting from interconnecting 54-61.
two or more generating systems. A I E E Trans. Power 48. Billinton, R. & Jain, A. V., The effect of rapid start
Apparatus Systems, 69 (1950) 1180-6. and hot reserve units in spinning reserve studies. IEEE
33. Cook, V. M., Galloway, C. D., Steinberg, M. J. & Trans. Power Apparatus Systems, PAS-91(2) (1972)
Wood, A. J., Determination of reserve requirements 511-16.
of two interconnected systems. A I E E Trans. Power 49. Allan, R. N. & Nunes, R. A. F., Modelling of standby
Apparatus Systems, 82 (1963) 18-33. generating units in short-term reliability evaluation.
34. Pang, C. K., Wood, A. J., Watt, R. L. & Bruggeman, J. IEEE Paper, (1979) A79 006-8.
A., Multiarea generation reliability studies. IEEE 50. Billinton, R. & Jain, A. V., Interconnected system
Summer Power Meeting (1978) Paper No. A 78 spinning reserve requirements. IEEE Trans. Power
546-4. Apparatus Systems, PAS-91(2) (1972) 517-526.
12 Ron Allan

51. Billinton, R. & Chowdhury, N. A., Operating reserve RTS. 1EEE Tram. Power Systems, 3(3) (1988)
assessment in interconnected generating systems. 1146-54.
IEEE Trans. Power Systems, 3(4) (1988) 1474-87. 69. Billinton, R. & Medicherla, T. K. P., Station originated
52. Dettmer, R., The UK electricity pool--A leap in the multiple outages in the reliability analysis of a
dark? IEE Rev., 37 (1991) 309-12. composite generation and transmission system. IEEE
53. Biilinton, R., Criteria used by Canadian utilities in the Trans. Power Apparatus Systems, PAS-100(8) (1981)
planning and operation of generating capacity. IEEE 3870-8.
Trans. Power Systems, 3 (1988) 1488-93. 70. Allan, R. N. & Adraktas, A. N., Terminal effects and
54. Billinton, R., Composite system reliability evaluation. protection system failures in composite system
IEEE Tram. Power Apparatus Systems, PAS-88(4) reliability evaIuation. IEEE Trans. Power Apparatus
(1969) 276-80. Systems, PAS-101(12) (1982) 4557-62.
55. Billinton, R. & Bhavaraju, M. P., Transmission 71. Billinton, R., Vohra, P. K. & Kumar, S., Effect of
planning using a reliability criterion: Part I--a station originated outages in a composite system
reliability criterion. IEEE Trans. Power Apparatus adequacy evaluation of the IEEE Reliability test
Systems, PAS-89(1) (1970) 28-34. system. IEEE Trans. Power Apparatus Systems,
56. Bhavaraju, M. P. & Billinton, R., Transmission PAS-104(10) (1985) 2649-56.
planning using a reliability criterion: Part II-- 72. Allan, R. N. & Ochoa, J. R., Modelling and assessment
transmission planning. IEEE Trans. Power Apparatus of station-originated outages for composite system
Systems, PAS-90(1) (1971) 70-8. reliability evaluation. IEEE Trans. Power Systems,
57. Dandeno, P. L., Jorgensen, G. E., Puntel, W. R. 3(1) (1988) 158-65.
& Ringlee, R. J., Program for composite bulk 73. Biilinton, R., Medicherla, T. K. P. & Sachdev, M. S.,
power electric system adequacy assessment. Proc. Application of common-cause outage models in
IEEE Conf. Reliab. Power Supply Systems 148 composite system reliability evaluation. IEEE Trans.
(1977). Power Apparatus Systems, PAS-100(7) (1981) 3648-
58. Marks, G. E., A method of combining high speed 57.
contingency load flow analysis with stochastic 74. BiUinton, R. & Cheng, L., Incorporation of weather
probability methods to calculate a quantitative effects in transmission system models for composite
measure of overall power system reliability. IEEE system adequacy evaluation. Proc. lEE, 133(6) (1986)
Paper, (1978) A78 053-1. 319-27.
59. Mikolinnas, T. A., Puntel, W. R. & Ringlee, R. J., 75. IEEE Subcommittee on the Application of Probability
Application of adequacy assessment techniques for Methods, Reliability indices for use in bulk power
bulk power systems. IEEE Trans. Power Apparatus supply adequacy evaluation. IEEE Tram. Power
Systems, PAS-101(5) (1982) 1219-28. Apparatus Systems, PAS-97(4) (1978) 1097-103.
60. Clements, K. A., Lain, B. P., Lawrence, D. J. & 76. Endrenyi, J., Albrecht, P. F., Billinton, R., Marks, G.
Reppen, N. D., Computation of upper and lower E., Reppen, N. D. & Salvaderi, L., Bulk power system
bounds on reliability indices for bulk power systems. reliability assessment--why and how? Part I: why?
IEEE Tram. Power Apparatus Systems, PAS-103(8) IEEE Trans. Power Apparatus Systems, PAS-101(9)
(1984) 2318-25. (1982) 3439-45.
61. Noferi, P. L. & Paris, L., Quantitative evaluation of 77. Endrenyi, J., Albrecht, P. F., Billinton, R., Marks, G.
power system reliability in planning studies. IEEE E., Reppen, N. D. & Salvaderi, L., Bulk power system
Trans. Power Apparatus Systems, PAS-91(2) (1972) reliability assessment--why and how? Part II: how?
611-18. IEEE Trans. Power Apparatus Systems, PAS-101(9)
62. Noferi, P. L., Paris, L. & Salvaderi, L., Monte Carlo (1982) 3446-56.
method for power system reliability evaluation in 78. Bhavaraju, M. P., Albrecht, P. F., Billinton, R.,
transmission and generation planning. Proc. Ann. Reppen, N. D. & Ringlee, R. J., Requirements for
Reliab. Maintainab. Syrup., Jan (1975) 449-59. composite system reliability evaluation models. IEEE
63. Dodu, J. C. & Merlin, A., New probabilistic approach Tram. Power Systems, 3(2) (1988) 149-57.
taking into account reliability and operation security in 79. Winter, W. R., Measuring and reporting overall
EHV power system planning at EDF. IEEE Trans. reliability of bulk electricity systems. CIGRE Paper,
Power Systems, PWRS-I(3) (1986) 175-81. 32-15 (1980).
64. Cunha, S. H. F., Pereira, M. V. F., Pinto, L. M. V. G. 80. Fong, C. C., BiUinton, R., Gunderson, R. O., O'Neil,
& Oliveira, G. C., Composite generation and P. M., Raksani, J., Schneider Jr, A. W. & Silverstein,
transmission reliability evaluation in large hydroel- B., Bulk system reliability--measurement and indices.
ectric systems. IEEE Trans. Power Apparatus Systems, IEEE Winter Power Meeting, (1989) Paper no. 89
PAS-104(10) (1985) 2657-64. WM155-3 PWRS.
65. Roman Ubeda, J. & Allan, R. N., Sequential 81. IEEE APM Subcommittee Report (Chairman Allan,
simulation applied to composite system reliability R. N.) Bulk system reliability predictive indices. IEEE
evaluation. Proc IEE, 139 (1992) 81-6. Trans. Power System, PWRS-5 (1990) 1204-13.
66. Salvaderi, L. & Billinton, R., A comparison between 82. IEEE APM Subcommittee Report (Chairman Ringlee,
two fundamentally different approaches to composite R.) Bulk system reliability--Criteria and indices trends
system reliability evaluation. IEEE Trans. Power and future needs. IEEE Winter Power Meeting, 1993,
Apparatus Systems PAS-104(12) (1985) 3486-93. Paper 93 WM 180-9-PWRS.
67. Anon., Power System Reliability Analysis (Vol. 2: 83. Dixon, G. F. L. & Hammersley, H., Reliability and its
Composite power system reliability evaluation) cost on distribution systems, International Conference
(CIGRE TF 38.03.10), CIGRE Publicatins, Paris, on Reliability of Power Supply Systems. lEE Conf.,
France, 1992. 148 (1977).
68. Bertoldi, O., Salvaderi, L. & Scalcino, S., Monte Carlo 84. Gaver, D. P., Montmeat, F. E. & Patton, A. D.,
approach in planning studies: An application to IEEE Power system reliability: I--measures of reliability and
 Power system reliability assessment----a review 13

methods of calculation. IEEE Trans. Power Apparatus Reliability evaluation of electrical systems with
Systems, 83(7) (1964) 727-37. switching actions. Proc. IEE, 123 (1976) 325-30.
85. Todd, Z. G., A probability method for transmission 94. Allan, R. N., De Oliveira, M. F. & Billinton, R.,
and distribution outage calculations. IEEE Trans. Reliability evaluation of the auxiliary electrical systems
Power Apparatus Systems, 33(7) (1964) 696-701. of power stations. IEEE Trans. Power Apparatus
86. S. A. Mallard & V. C. Thomas, A method for Systems, PAS-96(5) (1977) 1441-9.
calculating transmission system reliability. IEEE 95. Allan, R. N., Dialynas, E. N. & Homer, I. R.,
Trans. Power Apparatus Systems, PAS-87(3) (1968) Modelling and evaluating the reliability of distribution
824-34. systems. IEEE Trans. Power Apparatus Systems,
87. Billinton, R. & Bollinger, K. E., Transmission system PAS-09(6) (1979) 2181-9.
reliability evaluation using Markov processes. IEEE 96. IEEE Committee Report, Common mode forced
trans. Power Apparatus Systems PAS-87(2) (1968) outages of overhead transmission lines. IEEE Trans.
538-47. Power Apparatus Systems, PAS-95(3) (1976) 859-64.
88. Billinton, R. & Grover, M. S., Reliability assessment of 97. Allan, R. N., Dialynas, E. N. & Homer, I. R.,
transmission and distribution systems. IEEE Trans. Modelling common mode failures in the reliability
Power Apparatus Systems, PAS-94(3) (1975) 724-32. evaluation of power system networks. IEEE Paper,
89. Billinton, R. & Grover, M. S., Quantitative evaluation (1979) A79 040-7.
of permanent outages in distribution systems. IEEE 98. Lauby, M. G., Khu, K. T., Polesky, R. W., Vandello,
Trans. Power Apparatus Systems, PAS-94(3) (1975) R. E., Doudna, J. H., Lehman, P. J. & Klempel, D.
733-41. D., MAPP bulk transmission outage data collection
90. Ringlee, R. J. & Goode, S. D., On procedures for and analysis. IEEE Trans. Power Apparatus Systems,
reliability evaluation of transmission systems. IEEE PAS-103(1) (1984) 213-21.
Trans. Power Apparatus Systems, PAS-89(4) (1970) 99. Forrest, D. W., Albrecht, P. F., Allan, R. N.,
527-37. Bhavaraju, M. P., Billinton, R., Landgren, G. L.,
91. Endrenyi, J., Three state models in power system McCoy, M. F. & Reppen, N. D., Proposed terms
reliability evaluation. IEEE Trans. Power Apparatus for reporting and analyzing outages of electrical trans-
Systems, PAS-90(4) (1971) 1909-16. mission and distribution facilities. IEEE
92. Grover, M. S. & Billinton, R., A computerized Trans. Power Apparatus Systems, PAS-104 (1985)
approach to substation and switching station reliability 337-48.
evaluation. IEEE Trans. Power Apparatus Systems, 100. IEEE Standard 859-1987, Terms for Reporting and
PAS-93(5) (1974) 1488-97. Analyzing Outage Occurrences and Outage States of
93. Allan, R. N., BiUinton, R. & De Oliveira, M. F., Electrical Transmission Facilities.