Monte Carlo in Power System

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 14, Number 9 (2019) pp.
2252-2259
© Research India Publications. http://www.ripublication.com
Reliability Evaluation of Power System using Monte carlo simulation

in Pspice
Hemansu Patel
PG Student, Department of Electrical Engineering, Faculty of Technology and Engineering,
M. S. University of Baroda, Vadodara, 390001, India.
Anuradha Deshpande
Associate professor, Department of Electrical Engineering, Faculty of Technology and Engineering,
M. S. University of Baroda, Vadodara, 390001, India.
Abstract for complex system to be modeled. So, some significance can

be lost in resulting analysis. Due to this simulation technique is
Power systems reliability is a precondition step to design and important for reliability assessment. Simulation technique
planning of the modern power system. This paper focuses on estimates reliability by actual process and random behaviour of
evaluation of power system reliability with nonchronological
system. The techniques take into the planning, designing, and
load by Monte Carlo simulation (MCS) technique using Pspice.
operation of power system which considers random events like
In paper power system Reliability is found out in terms of
Probability of system success from the probability of system outages of elements, dependent events and component
failure. In this regard, failure probability of elements of the behaviour, queuing of failed component, load variation,
system is estimated using the probability distribution histogram variations in energy source and different operating conditions.
from Monte Carlo simulation. By applying outage conditions In simulation techniques Monte Carlo simulation (MCS)
on every element, considering most general possible failure method is used to evaluate system reliability. MCS method has
occurrence and Reliability in terms of probability is estimated. two approaches random and sequential. Random approach
To estimate failure probability random seed generation and simulates system lifetime intervals randomly in
sequences of trials generated from simulation histogram in nonchronological order and sequential approach simulates
Pspice is considered. The results of simulated system from interval in chronological order [2].
MCS approach is compared with the results of analytical
method. Non-sequential Monte Carlo simulation and contingency
enumerations are the common state selection techniques used
Keywords: Electric power system, Reliability Evaluation, in probabilistic reliability evaluation of bulk power systems.
Monte carlo simulation (MCS), Pspice. State enumeration is easier and the computational effort for
single order contingencies. Rei and Schilling [3] studied Monte
Carlo simulation requires larger computational effort but is
1 INTRODUCTION very versatile to model random behaviour of components. It
Reliability evaluation of power systems by simulation may be easiest way to evaluate adequately higher order
technique is helpful to find reliability of real system to supply contingencies of bulk system. To improve the computational
customers with reliable power service. Power system reliability speed. Kai Hou et al. [4] proposed approach of “impact-
evaluation is concerned with the determination of the adequacy increment-based decoupled reliability assessment (IID)”. In
of the combined generation and transmission system for which new formulation of reliability indices obtained by
providing a suitable supply at the load points. For reliability decoupling the reliability index of composite systems into
assessment, there are two main approaches analytical and generation adequacy and transmission reliability. Then,
simulation. Analytical methods are mostly used in past because impact-increment-based reliability assessment method is
simulation generally takes large amount of computing time and implemented to both and reduction technique used for the
results of analytical model sufficient for planners and designers higher order contingency states is developed for the
to make objective decisions for small system. But in recent transmission reliability evaluation. Then the number of
years, increasing interest towards modelling the actual system analysed contingencies can be significantly reduced, which
behaviour in simulating software to estimate system reliability improves the computational speed significantly.
used for all type of system [1]. In composite power system, Hua et al. [5] proposes a
In analytical techniques, system is represented by mathematical framework for “system reliability evaluation with subset
simulation. For that a small failure probability can be expressed
models to evaluate reliability using numerical solutions in
as a product of larger conditional probabilities and turning the
which assumptions are required for simplification of problems
problem of simulating a rare failure event into several
to produce analytical model. Assumptions required particularly
2252
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 14, Number 9 (2019) pp. 2252-2259
conditional simulations of more frequent intermediate failure In power system, number of components are more and
events. The inefficiency of Monte Carlo simulation in sometimes failure of component depends on other failure
simulating rare failure events is overcome by breaking the events. For that binominal distribution is used which associated
problem into estimating a sequence of conditional probabilities, with the combinational problems. Binominal distribution
which reduces Computational burden by extracting the subset represented by expression (P + Q)n. If component has n trials
of system states with significant contribution to reliability with r success or (n - r) failure then probability can be evaluated
indices”. For very reliable system, González-Fernández et al. from,
[6] consider the method which combines the concepts of Cross-
𝑛!
Entropy and Importance Sampling to obtain an optimal 𝑃𝑟 = 𝑃𝑟 𝑄𝑛−𝑟
distortion for the probabilistic parameters of system 𝑟!(𝑛−𝑟)!
components. The state probabilities are, thus, “properly
distorted so that important failure events are sampled more to 𝑃𝑟 = 𝑛𝐶𝑟 𝑃𝑟 𝑄𝑛−𝑟 (3)
deal with large and complex power systems”.
In above equation, P (Availability of component) and Q
In bulk power system the number of possible outage states are (Unavailability of component) remain constant and n must be
extremely large. Billinton and Zhang [7] proposed approach of fixed number of trial.
the “state extension algorithm which extend the knowledge of
2.2 MCS method
the investigated system states to collectively include the effects
of a large number of the investigated system states. The This method is class of computational algorithm which relies
algorithm provides more accurate adequacy indices without on repeated random sampling to evaluate numerical results. In
investigating extensive system contingencies and therefore this method random numbers are generated by the random
without significantly increasing the required computational number generators in digital computer with uniform random
effort”. Sometimes bulk system has several outages for that, number found in interval of (0,1). These numbers are tested for
Nahman [8] proposed a method for “Modeling simultaneous component random behaviour and failure sequences are
multiple station originated and common cause transmission generated from failure event of component. These sequences
line outages for bulk-power system reliability analysis. The are helpful to estimate the failure probability near true value.
failure-event vs system-element outage correlation-matrix is 2.2.1 Generation of Random Number
defined and used for system outage Modeling”. P. Hu et al. [9]
and Parvini et al. [10] represent the impact of renewable Random numbers are essential in simulation technique. They
sources in assessment of operational reliability of power system are variables which values are uniformly distributed in range of
and analyse the effect of energy storage system for reliability (0,1). Expression for congruential generators is shown below in
improvement of the system. which the previous value Xi used to evaluated sequence of new
number Xi+1.
2 METHODOLOGY
Xi+1 = AXi + C (mod B) (4)
For reliability assessment mainly two approaches are used 1)
Analytical (probabilistic) 2) Failure sequences in MCS method. Where B (modulus), C (increment) and A (multiplier) are non
Probabilistic method represents reliability of system in terms of negative numbers such that A, C, X0 < B and A ≠ 0. Sequence
probability. Which is evaluated from data of components Xi+1 stars with value X0 known as seed. Meaning of modulo
availability and unavailability. MCS method estimate the notation given as,
failure probability of system component from failure sequences
Xi+1 = (AXi + C) – B, for 0 < Xi+1 < B (5)
using simulated samples. From failure probability, success
probability of system can be obtained which represents the Then sequence is produced automatically and repeats itself till
system reliability. the step value is equal to a value not greater than B. After
2.1 Analytical method obtaining sequence of random numbers Xi, a uniform random
number Ui in the interval (0,1) can be found as,
In this method, every component has at least two outcomes
from which one considered as successive outcomes and another Xi
considered as failure outcomes. From this probability of
𝑈𝑖 = (6)
B
component given as below.
2.2.2 Sequence Simulation of Two Component System
𝑁𝑜. 𝑜𝑓 𝑠𝑢𝑐𝑐𝑒𝑠𝑠
𝑝(𝑠𝑢𝑐𝑐𝑒𝑠𝑠) = 𝑃 = Consider a system with two identical components and one
𝑁𝑜. 𝑜𝑓 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑜𝑢𝑡𝑐𝑜𝑚𝑒𝑠
component is essential for success of system. If both
𝑁𝑜. 𝑜𝑓 𝑓𝑎𝑖𝑙𝑢𝑟𝑒 components are in failure, only that state cause the system
𝑝(𝑓𝑎𝑖𝑙𝑢𝑟𝑒) = 𝑄 = failure. For this example, let the availability is 0.8 and
𝑁𝑜. 𝑜𝑓 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑜𝑢𝑡𝑐𝑜𝑚𝑒𝑠
unavailability is 0.2 for both component. In practice this data
(1) can be obtained from experimental testing of individual
And, P + Q = 1 (2) component. Then analytically failure probability of system is
given as,
2253
Probability of system failure = 0.2 x 0.2 = 0.04 Where Si = sequence number at j trial, Iij = number of
cumulative overlapping failure, NMCSij = simulation trials. From
Now this probability can be obtained from failure sequences by Fig.2, implies that as the number of trials increases the
simulation. For that one simulated sequence is shown in Fig.1 sequences oscillates near the true value which is matching wit
in which ‘o’ represent component found in failure state at a analytical probability 0.04. The results continue to oscillate
trial. even after a large number of trials. sometimes oscillating above,
sometimes below and sometimes around the true value. So,
from this concept probability of system failure can be
estimated.
2.3 System Reliability
For power system reliability evaluation, obtain the failure
probability of system components using above concept with
possible outage combination. Then the failure probability of
power system is given as,
𝑄 = ∑𝑗 [𝑃(𝐵𝑗) 𝑃1𝑗]
Fig.1 Simulated failure sequence 𝑄 = ∑𝑗 𝑃𝑗 (8)

Fig.1 indicate that component 1 suffers 5 failures and
component 2 suffers 4 failures in simulated period and two R=1–Q (9)
failures are overlapping for same trial. So, the system in failure
state at trials 5 and 9. If series system containing these two Where Bj = an outage condition in power system network,
components has 7 failures (overlapping component failures p(Bj) = failure probability of outage element and
responsible for system failure) and parallel system containing P1j = ‘probability of load at bus exceeding the maximum load
these two components has 2 failures (only overlapping failures which can be supplied at that bus without failure’. R is the
responsible for system failures). The simulated results for two reliability of power system.
parallel components shown in fig.2.
In simulated result three sequences are considered for 1000
trials. Each sequence starts with new seed generated randomly. 3 SYSTEM NETWORK and SOFTWARE MODELLING
From Fig.2 sequence 1 has two overlapping failures at first 100 3.1 Description of Power System
trial which gives probability of system failure as 2/100 = 0.02.
Single line diagram of Power system shown in Fig.3, consisting
In first 200 trials, it has 7 overlapping failure which is
two generating plant G1 and G2. Plant 1 has four generating
cumulative value considering the 2 overlapping failure from
units each of 20 MW and Plant 2 has two generating units each
100 trials. Then cumulative probability of failure at 200 trial is
of 30MW. Peak load of the system is 110 MW and assuming
7/200 = 0.035. Expression of sequence given as,
that it’s remaining constant.
Fig.2 Simulated sequence results

Fig.3 Three bus network
𝐼𝑖𝑗
𝑆𝑖𝑗 = (7)
𝑁𝑀𝐶𝑆𝑖𝑗
2254
Table1: Generation Data 3.2 Software Modelling
Plant No. of Capacity Unavailability Three phase Power system network of Fig.4 is modeled in
schematic of Pspice which is shown in Fig.5, in which plant 1
Units (MW)
has total generation capacity of 80 MW with four generating
1 4 20 0.01 units and plant 2 has 60 MW with two generating units. Total
generation of system is 140 MW and system peak load is 110
2 2 30 0.05 MW.
Table 2: Transmission Line Data 4 SIMULATION RESULTS and DISCUSSION
Line R X B/2 Unavailability The proposed system model analyses with MCS in Pspice and
failure sequences are obtained from MCS histogram for Outage
1 0.0912 0.4800 0.0282 0.00363667 elements. Samples generated from histogram for outage
condition of generation plant 1 is shown in Fig.6. for every state
2 0.0800 0.5000 0.0212 0.00454545 two sequences are considered and each sequence starts with
new seed generated randomly. Number of failures are obtained
3 0.0798 0.4200 0.0275 0.00341297 from probability distribution of samples from histogram.
In simulation, total 17 outage states are considered which are
For above system network results of analytical approach shown mentioned in analytical result. From which the result of outage
in table 1. There are 17 states of outage condition on network state at plant 1 (one unit in outage), plant 2 (two units in outage)
as Bj. Probability of outage element obtain from binominal and transmission line L3 shown in Fig.6,7, and 8 respectively.
distribution. Pj is failure probability of outage element and its Similarly, sequences of remaining states are obtained.
summation is the failure probability of system.
Table 3: Analytical Results of outage elements
State Bj Element on Outage Outage Probability of P(Bj) Failure probability Pj

P1j
1 G1 0.03462309 0 -
2 G1, G1 0.00052449 1 0.00052449
3 G1, G2 0.00364454 1 0.00364454
4 G1, L1 0.00012648 0 -
5 G1, L2 0.00015810 0 -
6 G1, L3 0.00011857 0 -
7 G2 0.09020227 1 0.09020227
8 G2, G2 0.00237374 1 0.00237374
9 G2, L1 0.00032951 1 0.00032951
10 G2, L2 0.00041188 1 0.00041188
11 G2, L3 0.00030891 1 0.00030891
12 L1 0.00313030 0 -
13 L1, L2 0.00001430 1 0.00001430
14 L1, L3 0.00001072 1 0.00001072
15 L2 0.00391288 0 -
16 L2, L3 0.00001340 1 0.00001340
17 L3 0.00293466 0 -
2255
Fig.5 Power system model in Pspice
Fig.6 MCS histogram
2256
Simulated sequences oscillated around the true value and shown in table 4. G1 is generating unit of plant1, G2 is
failure probability of element estimated. Form Fig.6,7 and 8, generating unit of plant2 and L1, L2 and L3 are transmission
the results are in error with less number of trials. This error line. Here, general outages of power system are considered.
decreases as the number of trial increases. These results of Each state elements are on outage except that remaining
probability compared with the analytical results which are elements are considered as fully reliable.
G1
0.04
Cumulative probability of system failure
0.035
0.03
0.025
0.02
0.015
0.01
0.005
0
100 200 300 400 500 600 700 800 900 1000
NO. of Trials
Sequence 1 Sequence 2
Fig.6 Failure sequence of outage G1
G2, G2
0.007
0.006
0.005
0.004
0.003
0.002
0.001
0
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
NO. of Trials
Sequence 1 Sequence 2
Fig.7 Failure sequence of outage G2, G2
2257
L3

0.007
0.006
0.005
0.004
0.003
0.002
0.001
0
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
NO. of Trials
Sequence Sequence2
Fig.8 Failure sequence of line L3
Simulated sequences oscillated around the true value and Each state elements are on outage except that remaining
failure probability of element estimated. Form Fig.6,7 and 8, elements are considered as fully reliable.
the results are in error with less number of trials. This error
decreases as the number of trial increases. These results of From table 4, we can see that generating units has high failure
probability level compared to transmission line failure
probability compared with the analytical results which are
probability. Results estimated from MCS technique are nearly
shown in table 4. G1 is generating unit of plant1, G2 is
similar with analytical results. From that the failure probability
generating unit of plant2 and L1, L2 and L3 are transmission
of system is obtained using eq. (8) and from that the reliability
line. Here, general outages of power system are considered. of system is obtained using eq. (9) which is quite similar with
analytical result.
Table 4: Comparison of Analytical and MCS result
State Bj Element on Outage Outage Probability of P(Bj) P1j Failure probability Pj

Analytical MCS Analytical MCS
1 G1 0.03462309 0.03100 0 - -
2 G1, G1 0.00052449 0.00062 1 0.00052449 0.00062
3 G1, G2 0.00364454 0.00350 1 0.00364454 0.00350
4 G1, L1 0.00012648 0.00025 0 - -
5 G1, L2 0.00015810 0.00025 0 - -
6 G1, L3 0.00011857 0.00028 0 - -
7 G2 0.09020227 0.08400 1 0.09020227 0.08400
8 G2, G2 0.00237374 0.00260 1 0.00237374 0.00260
9 G2, L1 0.00032951 0.00043 1 0.00032951 0.00043
10 G2, L2 0.00041188 0.00042 1 0.00041188 0.00042
11 G2, L3 0.00030891 0.00042 1 0.00030891 0.00042
12 L1 0.00313030 0.00370 0 - -
13 L1, L2 0.00001430 0.00012 1 0.00001430 0.00012
14 L1, L3 0.00001072 0.00011 1 0.00001072 0.00011
15 L2 0.00391288 0.00400 0 - -
16 L2, L3 0.00001340 0 1 0.00001340 0
17 L3 0.00293466 0.00260 0 - -
Q= 0.09783386 0.09222
Reliability = 1 – Q = 0.90216614 0.9078
2258
5 CONCLUSION [8] Nahman, J.M.: ‘Modeling simultaneous outages for

bulk-power system reliability-analysis’, IEEE Trans.
This paper has presented a simulation approach to estimate the
Reliab., 1985, 4, (3), pp. 554–559.
power system reliability using Monte Carlo simulation in
Pspice. By applying Outages on different element of power [9] P. Hu, R. Karki and R. Billinton, "Reliability evaluation
system modelled in Pspice and simulating the failure states to of generating systems containing wind power and
obtain failure probability for that element which is obtained energy storage," in IET Generation, Transmission &
from probability sequences generated using simulation samples
Distribution, vol. 3, no. 8, pp. 783-791, Aug. 2009.
in MCS process. For generating unit and transmission line
outages considered whose occurrences are highly possible in [10] Z. Parvini, A. Abbaspour and M. Fotuhi-Firuzabad and
the system. M. Moeini-Aghtaie “Operational Reliability Studies of
The simulation results showed that the proposed simulation Power Systems in the Presence of Energy Storage
method significantly estimate the reliability of given system. Systems” in IEEE Transactions on Power Systems, vol.
Also, Monte Carlo simulation in Pspice confirms that and can 33, no. 4, pp. 3691-3700, July2018.
estimate reliability of modelled system. In this paper [11] PSpice® User Guide, April 2016.
considered system has less number of elements compare to the
practical power system, to verify the simulation results with the
analytical results. So, this proposed approach can be used to
estimate power system with higher number of components.
REFERENCES
[1] Billinton R. and Allan R.N., “Reliability Evaluation of

Power Systems” 2nd Edition, Plenum Press, New York,
1984.
[2] Billinton R. and Allan R.N., “Reliability Evaluation of

Engineering Systems” 2nd Edition, Plenum Press, New
York 1992.
[3] A.M. Rei and M. T. Schilling, “Reliability assessment of

the Brazilian power system using enumeration and
Monte Carlo,” IEEE Trans. Power Syst., vol. 23, no. 3,
pp. 1480–1487, Aug. 2008.
[4] Kai Hou, Hongjie Jia, Xue Li, Xiandong Xu, Yunfei
Mu, Tao Jiang and Xiaodan Yu, “Impact-increment
based decoupled reliability assessment approach for
composite generation and transmission systems” in IET
Generation, Transmission & Distribution, vol. 12, no. 3,
pp. 586-595, 2018.
[5] Hua, B., Bie, Z., Au, S.K., et al.: ‘Extracting rare failure
events in composite system reliability evaluation via
subset simulation’, IEEE Trans. Power Syst., 2015, 30,
(2), pp. 753–762.
[6] González-Fernández, R.A., da Silva, A.M.L., Resende,

L.C., et al.: ‘Composite systems reliability evaluation
based on Monte Carlo simulation and cross-entropy
methods’, IEEE Trans. Power Syst., 2013, 28, (4), pp.
4598–4606
[7] Billinton, R., Zhang, W.: ‘State extension for adequacy

evaluation of composite power systems-applications’,
IEEE Trans. Power Syst., 2000, 15, (1), pp. 427–432.
2259

Monte Carlo in Power System

Uploaded by

Monte Carlo in Power System

Uploaded by

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 14, Number 9 (2019) pp.

Reliability Evaluation of Power System using Monte carlo simulation

Abstract for complex system to be modeled. So, some significance can

Fig.1 Simulated failure sequence 𝑄 = ∑𝑗 𝑃𝑗 (8)

Fig.2 Simulated sequence results

Table1: Generation Data 3.2 Software Modelling

Table 2: Transmission Line Data 4 SIMULATION RESULTS and DISCUSSION

Table 3: Analytical Results of outage elements

State Bj Element on Outage Outage Probability of P(Bj) Failure probability Pj

Fig.5 Power system model in Pspice

Fig.6 MCS histogram

Fig.6 Failure sequence of outage G1

Fig.7 Failure sequence of outage G2, G2

Cumulative probability of system failure

Fig.8 Failure sequence of line L3

State Bj Element on Outage Outage Probability of P(Bj) P1j Failure probability Pj

5 CONCLUSION [8] Nahman, J.M.: ‘Modeling simultaneous outages for

[1] Billinton R. and Allan R.N., “Reliability Evaluation of

[2] Billinton R. and Allan R.N., “Reliability Evaluation of

[3] A.M. Rei and M. T. Schilling, “Reliability assessment of

[6] González-Fernández, R.A., da Silva, A.M.L., Resende,

[7] Billinton, R., Zhang, W.: ‘State extension for adequacy

You might also like