Power System Reliability Lecture No.

7. RELIABILITY PRINCIPLES
7.1 FAILURE RATE
Since every piece of equipment in a system will eventually fail if it is in service
for a long period, there is a failure rate associated with each one. For some
items, the failure rate is quite significant while for others it could be extremely
low.
Failure rate is defined as the number of expected failures per unit in a given
time interval. It is just an expected value. In calculating the failure rate of a
group of units, the total operating time of the units should be used instead of
the chronological time. The formula is

Example 7.1
Ten transformers were tested for 500 h each and four transformers failed after
the following test
time periods:
one failed after 50 h
one failed after 150 h
two failed after 400 h
What is the failure rate for these types of transformers?
Total operating time of units

Example 7.2
Thirty motors were tested for 200 h. Five motors failed during the test. The
failures occurred after the following test times:
Motor 1 60 h
Motor 2 71 h
Motor 3 157 h
Motor 4 160 h

1
Power System Reliability Lecture No.7

Motor 5 170 h
What is the estimated failure rate?
Solution:
Total number of unit operating hours

= 60+71þ+157+160+170+25 x 200
= 5618 unit h

= 5/5168 unit h
= 0:00092 failures/h

7.2 Concept of Bathtub Curve

The life of a man has the following three major distinguishable periods as
shown in Fig.7.1:
1. Infant mortality period
2. Useful life period
3. Wear-out period

Fig.7.1 Bathtub curve failure rate versus time for human life.

2
Power System Reliability Lecture No.7

Also the life of equipment has similar three major periods as shown in Fig.7.2:
1. Infant mortality period
2. Useful life period
3. Wear-out period

Fig.7.2. Bathtub curve failure rate versus time for components.

 In the infant mortality period, the failure rate is high due to the presence
of weak spots from the manufacturing process such as poor
workmanship, substandard components, and so on. As these
weaknesses are manifested one by one by the stress of operation, the
failure rate keeps decreasing until a low constant level is reached.
 The equipment then enters the useful life period where failures are due
to chance and occur at random times. Failures in this period are also
independent of the age of the equipment.
 When the components of the equipment start to wear out. From this
time on, the failure rate rises rather rapidly due to deterioration.

Most reliability work deals with the useful life period when the failure rate is
constant and the exponential distribution applies. one of the several possible
distributions.

3
7.3 Reliability Model
Consider the case in which a fixed number of identical components (No) are
tested.
Let ;
Ns (t) = number of component surviving at time t
Nf (t ) = number of components failed at time t.
No = Ns (t) + Nf (t )

The reliability is the probability of not failing in a specified time interval. If the
original population is No and Nf of them fail after time t, leaving Ns surviving,
the reliability at time t which is the probability of surviving is

and

The probability of failure is

Hence R (t) + Q (t) = 1 (They are complementary)

If we define as the instantaneous failure density function

(=probability density function) f(t) , from which :

By integrating both sides:

4
 1- R(t) = ........ (1) (since R(0) =1)

If the failure rate is constant, then we can define:

Since we have

Example 7.3
Two hundred capacitors were installed and at the end of each year, the
number of surviving units was tallied.

5
Power System Reliability Lecture No.7

Based on these figures, what is the reliability of the capacitors for 5 years?
The annual reliability of Year 4 ? Assuming the reliability function is
exponential, that is, R (t) = e-t, what is the failure rate for this formula?
Solution:
Reliability for 5 years = 169/200 = 0:845

Example 7.4
Ten thousand new oil circuit reclosers (OCRs) are put in service. They have a
constant failure rate of 0.1 per year. How many units of the original 10,000 will
still be in service after 10 years? How many of the original will fail in Year 10?
Solution:
Probability of survival is given by

In 10 years, probability of survival

6
Power System Reliability Lecture No.7

Out of 10,000 original units

10,000x0.3679=3679 should survive
Number of failures in Year 10

7
Power System Reliability Lecture No.8

8 . RELIABILITY OF COMPLEX SYSTEMS

8.1 Series Systems
Power system systems are made up of subsystems or components.
Functionally, these subsystems or components are arranged in series or
parallel connections or in combinations of the two.
For a series connection, every component is required to function for system
success. Let us consider n component system X1, X2, X3 ,.........., Xn as shown in
Fig 8.1.

X1 X2 X3 Xn

Fig. 8.1. A series connection.

The probability of success of the system is

Where Probability of success of = Reliability (t)

Probability of success of = Reliability (t)

Probability of success of = Reliability (t)

......................... ...

Probability of success of = Reliability (t)

If the components have exponential failure probabilities with corresponding

failure rates, λ1, λ2, λ3, and so on, then the system reliability

8
Power System Reliability Lecture No.8

The sum ( λ1 +λ2 +λ3+ -----+λn ) is a constant = λ. It is the composite failure rate
of the series system. From the system failure rate λ, the mean time between
failures of the system can be calculated:

8.2 Parallel Systems

A parallel system in a reliability sense means only one of the components in
the parallel connections has to work in order that the system will function. Let
us consider n component system X1, X2, X3 ,.........., Xn connected in parallel as
shown in Fig 8.2.

X1

X2
Fig. 8.2. A parallel connection

X3

Xn

The probability of success of the system is

where = Probability of failure of ith component =

9
Power System Reliability

Hence:

If all the elements have the same Q(t)’s

If all the elements of the parallel system have constant failure rates and hence
exponential reliability functions, as is usually assumed, the reliability of the
parallel system will be

If all λ are the same,

Example8.1:
What is the reliability of the following system as shown in Fig. 8.3?

Fig.8.3 . A parallel–series connected system.

10
Power System Reliability Lecture No.8

Solution:

Reliability of system

8.3 Availability Model

When a system fails, it will be out of service for some time until it is repaired or
replaced. Even for systems with spare units, the system can be “down” if a
failure occurs when no more spares are available. The percentage of time that
the system is functioning is called the availability of the system. It is usually
expressed as

Since on average, it takes a time interval equal to the MTTF for the system to
fail and a time interval equal to the MTTR for the system to be operational
again, the availability, defined as uptime / (uptime+ downtime), can be
expressed as

11
Power System Reliability Lecture No.8

and unavailability can be expressed as

For systems that can be treated as a single component with a constant failure
rate λ and repair rate μ, the availability

and unavailability

12
Power System Reliability Lecture No.9

9. Reliability of series parallel systems

 In reliability evaluations, we reduce the complicated configurations by
combining appropriate series and parallel branches of the model until a
single equivalent element remains. This approach is called successive
reduction method.
 For example, consider the structure shown below, the step of reduction
are shown in the diagrams (a) to f) :
0.95 0.92 0.88
A B A.B

0.85 0.85
C C
0.70
E G G
0.8 0.70 0.90 0.8 0.91 0.90
D F D E// F

(a) System (b) First reduction

0.88 0.88
A.B A.B

C 0.85 0.96 0.90

G
0.90
C// [ D . (E// F)} G
0.73
D . (E// F)

(c) Second reduction (d) Third reduction

0.88
A.B

A.B // (G.{C// [ D . (E// F)})

0.98
G.{C// [ D . (E// F)}

0.86

(e) Fourth reduction (f) Final reduction

13
Power System Reliability Lecture No.9

Networks reliability evaluation using set theory

1. Single element network:

Consider a single element A A

shown .For this element,

The probability of success P(S) = P(A) =Pa for which the element is available .

Also the probability of failure is P(f) =1- P(A) =Qa for which the element is
unavailable.

2. Two elements network:

(i) For two elements connected in series,

The availability of the network is A B

The unavailability of the network is

(ii) For two elements connected in parallel, A

The availability of the network is

C

The unavailability of the network is

2. Three elements network:

(i) Consider the case of three elements network as shown below:

The availability of the network is A

B

14
Power System Reliability Lecture No.9

The unavailability of the network is

(i) If the same three elements network are re-connected as shown below:

A B

The availability of the network is C

The unavailability of the network is

3. Four elements network:

(i) Case when the four elements connected as shown below

A B
The availability of the network is

C D

The unavailability of the network is

15
Power System Reliability Lecture No.9

(ii) Case when the four elements connected as shown below

A B

C D
The availability of the network is

The unavailability of the network is

16