Introduction to Power System Reliability Evaluation
Availability (AV) and Forced Outage Rate (FOR)
Experience has shown that no machine is so reliable and dependable that it is
available and in operating conditions all the time. That means that the machine
needs to be out of service (off service) for maintenance or it may be off due to
some problems.
Scheduled outages (planned outages) : purposely unit is taken out of service for
maintenance or replacement.
Forced outages : when unit (s) is out of service due to failure called also
unscheduled (unplanned) outage.
The last one is actually the most severe and important factor in power system
planning and operation, and can be defined as
Time unit is on forced outages
Forced outage Rate (FOR) =
Time unit on service + Time on outages
Therefore, Availability can be defined as
Availability (AV) = Time unit on service
Total time
From the above definitions, it is obvious that :
FOR + AV = 1
The two terms availability and forced outage rate represent probability of
events occurrence. From the probability theory, it is known that the product
AV1 * AV2 represents the probability that both unit 1 and unit 2 are
simultaneously in operation during a specified period of time, also
AV1 * AV2 * AV3 means 1 & 2 & 3 are in operation at the same time
FOR1 * FOR2 * FOR3 means that units 1 & 2 & 3 are out of service in the same
time.
AV1 * FOR2 means the probability that unit 1 is available (in service) and unit 2
is unavailable (out of service) in the same time.
Example
For the following units, find out the availability
a) One unit of 300 MW
b) Two units of 150 MW each
c) Three units of 100 MW each
The FOR for each unit is 0.02
Solution
a) For 1 unit
Capacity out Capacity in Availability
0 300 0.98
300 0 0.02
1 .00
b) For 2 units
0 300 0.98*0.98 =.9604
150 150 .98*.02 =.0196
150 150 .02*.98 =.0196
300 0 .02*.02 =.0004
1.0000
c) For 3 units
0 300 .98*.98*.98 =.941192
100 200 .02*.98*.98+
.98*.02*.98+ =.057624
.98*.98*.02
200 100 (.02*.02*.98)*3 =.001176
300 0 .02*.02*.02 =.000008
1.000000
Considering system C, as an example, the probability of 2 units being out of
service i.e a capacity of 200 MW out of service, is
P= C (AV)n-r (FOR)r
N r
n!
Where C =
N r (n r )!r!
P = reliability of system availability
n = no of units
r = no of units unavailable (being on forced outage, from 0 to n)
AV = availability (probability of being in-service)
= 1-FOR
FOR = unavailability (forced outage rate)
Capacity Outage Probability Table (COPT)
The COPT is a table contains all the capacity states in an ascending order of
outages magnitude. Each outage (capacity state) is multiplied by its probability. If
the system contains identical units, Binomial distribution can be used. If the units
are not identical, the procedure in the following example can be used.
Example (1):
A generating system has the following units:
(a) 10 MW (FOR = 0.02).
(b) 15 MW (FOR = 0.03).
(c) 20 MW (FOR = 0.05).
It is required to build the COPT for the system.
solution:
COPT For units (a) will be as follows:
cap out probability
0 0.98 x 0.97 x 0.95 = 0.90307
10 0.02 x 0.97 x 0.95 = 0.01843
15 0.98 x 0.03 x 0.95 = 0.02793
20 0.98 x 0.97 x 0.05 = 0.04753
25 0.02 x 0.03 x 0.95 = 0.00057
30 0.02 x 0.97 x 0.05 = 0.00097
35 0.98 x 0.03 x 0.05 = 0.00147
45 0.02 x 0.03 x 0.05 = 0.00003
---------
1.00000
Example (2):
A generating system has the following data:
(a) 2X20 MW and (b) 1X30 MW. The FOR for each unit is 0.1. It is required to
establish the COPT for the system.
Solution:
COPT For units (a) will be as follows:
cap out probability
0 0.92 = 0.81
20 2(0.9 x 0.1) = 0.18
40 0.12 = 0.01
--------
1.00
COPT For units (b) will be as follows:
cap out probability
0 0.9
30 0.1
---
1.0
The two COPTs can be combined in one single COPT, as follows:
Cap out probability
0 0.81 x 0.9 = 0.729
20 0.18 x 0.9 = 0.162
30 0.81 x 0.1 = 0.081
40 0.01 x 0.9 = 0.009
50 0.18 x 0.1 = 0.018
60 ---
70 0.01 x 0.1 = 0.001
-----
1.000
Loss of Load Expectation (LOLE)
The LOLE risk index is the most widely accepted an used probabilistic method in
system reliability evaluation for generating systems. Two models are required and
employed. One is the previously studied Load Duration Curve (LDC), and the
other is the COPT. These two models are convolved (combined) in the process.
The units of the LOLE is in days per year (d/y). The LOLE evaluation method is
expressed in the following mathematical formula:
n
LOLE = t i pi (d/y) ( L max > C)
i=1
Installed capacity
Reserve
Peak load
ENS (Ei) Outage (Oi)
ti
Minimum load
MW
Total energy demanded
0 Time 100%
It is clear from the above load characteristics that capacity outages less than the
reserve will not cause a loss of load. Consider now:
Oi = the ith outage state in the COPT.
pi = the probability of this ith outage.
ti = the number of time units for which this outage cause loss of load.
EXAMPLE (1):
Consider a system having 5 x 60 MW units, each with an FOR of 0.03. The load
model can be assumed as linear with a maximum load (Lmax) of 240 MW, and a
minimum load (Lmin) of 100 MW. Evaluate the system LOLE.
Solution:
The COPT and the loss of load expected can be convolved as follows:
(1) (2) (3) (2x3)
cap out probability time ti loss of load expected
(MW) (%)
------------------------------------------------------------------------------------
0 0.858734 0 --
60 0.132794 0 --
120 0.008214 42.86 0.352052
180 0.000254 85.72 0.021770
240 0.000004 100.00 0.000400
------------ ------------
1.000000 0.374222%
------------------------------------------------------------------------------------
If 100% of time represents 365 days, The LOLE will be:
LOLE = 0.374222 x 365/100
= 1.365910 d/y
In system planning, the above figure is considered to be excessive (above normal),
Therefore, a value often considered to be normal LOLE risk index is considered as
0.1 d/y.
EXAMPLE (2):
Consider now the same system, but with each unit is having an FOR of 0.01.
Calculate the LOLE of the system.
Solution:
(1) (2) (3) (2x3)
cap out probability time ti loss of load expected
(MW) (%)
------------------------------------------------------------------------------------
0 0.950990 0 --
60 0.048030 0 --
120 0.000970 42.86 0.041574
180 0.000010 85.72 0.000857
240 0.000000 100.00 0.000000
----------- ------------
1.000000 0.042431%
------------------------------------------------------------------------------------
If 100% of time represents 365 days, The LOLE will be:
LOLE = 0.042431 x 365/100
= 0.154874 d/y
The next flowchart shows the systematic steps implemented (followed) in the
evaluation of the reliability level (LOLE) in power system planning process in
order to determine the appropriate capacity reserve margin in each year of the
planning horizon. [the LOLE level is based on the level prescribed by the utility
management decision (i.e. the LOLEp)].
Expected Demand Not Served (DNS)
In power system planning, we need sometimes another reliability index beside the
LOLE, to know the magnitude of load that has been lost due to severe outages (i.e.
when Lmax > C). So, the DNS can be found as follows:
n
DNS = Oi pi MW/y ( L > C)
max
i=1
Expected Energy Not Served (ENS)
Since the power systems are in fact energy system, where energy sale is the real
revenue for the electric company, so, another essential and most needed reliability
index known called the ENS can be deduced as follows:
( L > C&)
n
ENS = ENS i pi MWh/y max
i=1
Energy Index of Reliability (EIR)
The ratio of expected energy not served (ENS) to the system Total Energy
Demanded (TED) can be found as:
ENS
ENS pu =
TED
This ratio, in fact is so small because of the small nature of the ENS and large
nature of the TED, so, we can deduce another important reliability index called the
EIR, and can be expressed as follows:
EIR = 1 - ENS pu
100
LOLE 10
(days/y
1 A
B
C
0.1
Reliability risk level (0.1 dly)
0.01
0,001
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
100
10
LOLE
1 A
days/y
B
C
0.1
Reliability risk level (0,1 d/y)
0.01
0.001
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
