Reliability Assessment of

Distribution Networks

Dr. Sanjib Ganguly

Associate Professor, Department of EEE,
IIT Guwahati (Email: sganguly@[Link] )
2 Basic definition: Outages

 Outage is the non-availability of operation of certain

equipment
 Outage may or may not cause interruption.
 Outage can be of:
 Forced outage: Equipment taken out-of-service on
emergency basis
 Scheduled outage: Equipment taken out-of-service for
preventive maintenance
 Partial outage: Equipment operational with reduced
capacity
3 Basic definition: Interruption

 Interruption is the loss of service to one or more consumers.

An interruption may occur due to one or more component
outages.
 Interruption can be of the following types:
 Scheduled interruption: Due to scheduled outages
 Momentary interruption: Duration few seconds
 Temporary interruption: Duration few minutes
 Sustained interruption: Duration few hours
4 Preparation of outage reports

According to an IEEE committee report, the following basic information

should be included in an equipment outage report:
 Type, design, manufacturer, and other descriptions for classification
purposes
 Date of installation, location on system, length in the case of a line
 Mode of failure (short-circuit, false operation, etc.)
 Cause of failure (lightning, tree, etc.)
 Times (both out of service and back in service, rather than outage
duration alone), date, meteorological conditions when the failure
occurred
 Type of outage, forced or scheduled, transient or permanent
5 Distribution system reliability

The major challenges of reliability assessment:

Insufficient data on reliability performance

Most of the interruptions of distribution networks are

caused by severe weather-related interruptions and
for inadequate maintenance.

Distribution system reliability can be improved by the

timely identification and response to failures.
6 Reliability Indices

 There are various types of reliability indices. They

are:
 Indices for sustained interruptions

 Indices for momentary interruptions

 Load and energy-based indices

7 Usage of the reliability indices
 Most utilities use one or more of the reliability indices to
understand how the distribution system is performing.

 For example, some of the reliability indices which are

formulated for sustained interruptions may not reveal
momentary interruptions.
8 Benefits of reliability-based study

 A reliability assessment study can help to quantify the impact

of design improvement options.
 Typical improvement options that can be studied based on a
predictive reliability model include the following:
 New feeders and feeder expansions
 Load transfers between feeders
 New substation and substation expansions
 New feeder tie points
 Locations for sectionalizing switches
 Feeder automation
 Replacement of aging equipment
9 Costs associated with system reliability

 Investment cost (utility side) to achieve a desired level

of reliability

 Customer interruption cost which includes the cost of

production loss for process industries

 Customer damage cost (utility side), if applicable

10 Reliability and Cost: Conflicts each
other
11 Reliability index: SAIFI

 System average interruption frequency index (SAIFI)

(sustained interruptions): This index is designed to provide the
information about the average frequency of sustained
interruptions per customer over a predefined service area.

𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛𝑠 𝑁𝑖

SAIFI= 𝑆𝐴𝐼𝐹𝐼 =
𝑁𝑇
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑

Where
Ni is the number of interrupted customers for each interruption event during reporting period
NT is the total number of customers served for the area being indexed
12 Reliability index: SAIDI

 System average interruption duration index (SAIDI): This index

is commonly referred to as customer minutes of interruption or
customer hours, and is designed to provide information about
the average time the customers are interrupted.

𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛 𝑑𝑢𝑟𝑎𝑡𝑖𝑜𝑛𝑠 𝑟𝑖 𝑁𝑖

SAIDI= 𝑆𝐴𝐼𝐷𝐼 =
𝑁𝑇
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑

Where ri is the restoration time for each interruption event.

13 Reliability index: CAIDI

 Customer average interruption duration index (CAIDI): It

represents the average time required to restore service to the
average customer per sustained interruption.

𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛 𝑑𝑢𝑟𝑎𝑡𝑖𝑜𝑛𝑠

CAIDI=
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛𝑠

𝑟𝑖 𝑁𝑖 𝑆𝐴𝐼𝐷𝐼
𝐶𝐴𝐼𝐷𝐼 = =
𝑁𝑖 𝑆𝐴𝐼𝐹𝐼
14 Reliability index: CTAIDI
 Customer total average interruption duration index (CTAIDI): For
customers who actually experienced an interruption, this index
represents the total average time in the reporting period they were
without service.
 This index is a hybrid of CAIDI and is calculated the same except that
customers with multiple interruptions are counted only once.

𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛 𝑑𝑢𝑟𝑎𝑡𝑖𝑜𝑛𝑠

CTAIDI= 𝐶𝑇𝐴𝐼𝐷𝐼 =
𝑅𝑖 𝑁𝑖
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑒𝑑 𝐶𝑁

Where, CN is total number of customers who have experienced a sustained interruption during the
reporting period.
15 Reliability indices: CAIFI
 Customer average interruption frequency index (CAIFI):
This index gives the average frequency of sustained
interruptions for those customers experiencing sustained
interruptions.
 The customer is counted only once regardless of the
number of times interrupted.

𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛𝑠 𝑁𝑖

CA𝐼𝐹𝐼 = 𝐶𝐴𝐼𝐹𝐼 =
𝐶𝑁
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑒𝑑
16 Reliability index: ASAI

 Average service availability index (ASAI): This index represents

the fraction of time (often in percentage) that a customer
had power without interruption during 1 year or the defined
reporting period.
𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 ℎ𝑜𝑢𝑟𝑠 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑖𝑙𝑖𝑡𝑦
A𝑆𝐴𝐼 =
𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 ℎ𝑜𝑢𝑟𝑠 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑑𝑒𝑚𝑎𝑛𝑑

ℎ𝑜𝑢𝑟𝑠
𝑁𝑇 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 − 𝑟𝑖𝑁𝑖
𝑦𝑒𝑎𝑟
𝐴𝑆𝐴𝐼 =
𝑁𝑇(𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 ℎ𝑜𝑢𝑟𝑠/𝑦𝑒𝑎𝑟)

𝑟𝑖𝑁𝑖 represents total duration of interruption in a year

17 Reliability index: ASIFI

 Average system interruption frequency index (ASIFI): This index is

specifically designed to calculate reliability based on load rather than
number of customers.
 It is an important index for areas that serve industrial/commercial
customers.
 It is also used by utilities that do not have precise customer tracking
systems.
𝑐𝑜𝑛𝑛𝑒𝑐𝑡𝑒𝑑 𝑘𝑉𝐴 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑒𝑑
A𝑆𝐼𝐹𝐼 = 𝐴𝑆𝐼𝐹𝐼 =
𝐿𝑖
𝑡𝑜𝑡𝑎𝑙 𝑐𝑜𝑛𝑛𝑒𝑐𝑡𝑒𝑑 𝑘𝑉𝐴 𝑠𝑒𝑟𝑣𝑒𝑑 𝐿𝑇

Where
𝐿𝑖 is the total connected kVA load interrupted for each interruption event
𝐿𝑇 is the total connected kVA load served
18 Reliability index: ASIDI

 Average system interruption duration index (ASIDI): This index

is formulated with the same philosophy as ASIFI, but it provides
information on system average duration of interruptions. Thus,

𝑐𝑜𝑛𝑛𝑒𝑐𝑡𝑒𝑑 𝑘𝑉𝐴 𝑑𝑢𝑟𝑎𝑡𝑖𝑜𝑛 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑒𝑑

A𝑆𝐼𝐷𝐼 =
𝑡𝑜𝑡𝑎𝑙 𝑐𝑜𝑛𝑛𝑒𝑐𝑡𝑒𝑑 𝑘𝑉𝐴 𝑠𝑒𝑟𝑣𝑒𝑑

𝑟𝑖 𝐿𝑖
𝐴𝑆𝐼𝐷𝐼 =
𝐿𝑇
19 Reliability index: CEMIn

 Customers experiencing multiple interruptions (CEMIn): This

index is designed to track the number (n) of sustained
interruptions to a specific customer.
 Its purpose is to help identify customer trouble that cannot be
seen by using averages.
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑡ℎ𝑎𝑡 𝑒𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒𝑑 𝑚𝑜𝑟𝑒 𝑠𝑢𝑠𝑡𝑎𝑖𝑛𝑒𝑑 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛𝑠
C𝐸𝑀𝐼𝑛 =
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑
𝐶𝑁(𝑘>𝑛)
𝐶𝐸𝑀𝐼𝑛 =
𝑁𝑇
where CN(k>n) is the total number of customers who have experienced more than n sustained interruptions
during the reporting period.
20 Reliability index: MAIFI

 Momentary average interruption frequency index

(MAIFI): This index is very similar to SAIFI, but it tracks the
average frequency of momentary interruptions.

𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑚𝑜𝑚𝑒𝑛𝑡𝑎𝑟𝑦 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛𝑠

𝑀𝐴𝐼𝐹𝐼 =
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑

𝐼𝐷𝑖 𝑁𝑖
𝑀𝐴𝐼𝐹𝐼 =
𝑁𝑇

where IDi is the number of interrupting device operations.

MAIFI is the same as SAIFI, but it is for short-duration rather than long-duration interruptions.
21 Reliability index: MAIFIE

 Momentary average interruption event frequency index

(MAIFIE): This index is very similar to SAIFI, but it tracks the
average frequency of momentary interruption events.

𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑚𝑜𝑚𝑒𝑛𝑡𝑎𝑟𝑦 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛 𝑒𝑣𝑒𝑛𝑡𝑠

MAIFIE =
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑

𝐼𝐷𝐸 𝑁𝑖
𝑀𝐴𝐼𝐹𝐼𝐸 =
𝑁𝑇

where IDE is the interrupting device events during reporting period.

22 Reliability index: CEMSMIn

 Customers experiencing multiple sustained interruption and

momentary interruption events (CEMSMIn): This index is formulated
for tracking the number (n) of both sustained interruption and
momentary interruption events to a set of specific customers.

𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑡ℎ𝑎𝑡 𝑒𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒𝑑 𝑚𝑜𝑟𝑒 𝑡ℎ𝑎𝑛 𝑛 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛𝑠

C𝐸𝑀𝑆𝑀𝐼𝑛 =
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑

𝐶𝑁𝑇(𝑘>𝑛)
𝐶𝐸𝑀𝑆𝑀𝐼𝑛 =
𝑁𝑇
Where 𝐶𝑁𝑇(𝑘>𝑛) is the total number of customers who have experienced more than n
sustained interruption and momentary interruption events during the reporting period.
23 Load and energy based reliability indices

 There are also load- and energy-based indices. In determination

of such indices, one has to know the average load at each load
bus. This average load Lavg at a bus is found from:

𝐿𝑎𝑣𝑔 = 𝐿𝑝𝑒𝑎𝑘 𝐿𝐹

Where,
Lavg is the peak load (demand)
LF is the load factor
24 Load and energy based reliability index: ENS

 Energy not supplied index (ENS): This index represents

the total energy not supplied by the system and is
expressed as:

𝐸𝑁𝑆 = 𝐿𝑎𝑣𝑔,𝑖 𝑟𝑖

Where Lavg,i is the average load connected to load point i

25 Load and energy based reliability index:
AENS
 Average energy not supplied (AENS): This index represents
the average energy not supplied by the system.
𝑡𝑜𝑡𝑎𝑙 𝑒𝑛𝑒𝑟𝑔𝑦 𝑛𝑜𝑡 𝑠𝑢𝑝𝑝𝑙𝑖𝑒𝑑
A𝐸𝑁𝑆 =
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑

𝐿𝑎𝑣𝑔, 𝑖 𝑟𝑖
𝐴𝐸𝑁𝑆 =
𝑁𝑇

This index is the same as the average system curtailment index (ASCI).
26 Load and energy based reliability index:
ACCI
 Average customer curtailment index (ACCI): This index
represents the total energy not supplied per affected customer
by the system.
𝑡𝑜𝑡𝑎𝑙 𝑒𝑛𝑒𝑟𝑔𝑦 𝑛𝑜𝑡 𝑠𝑢𝑝𝑝𝑙𝑖𝑒𝑑
A𝐶𝐶𝐼 =
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑎𝑓𝑓𝑒𝑐𝑡𝑒𝑑

𝐿𝑎𝑣𝑔, 𝑖 𝑟𝑖
𝐴𝐶𝐶𝐼 =
𝐶𝑁

It is a useful index for monitoring the changes of average energy not supplied between one
calendar year and another.
 27 Example…

 The information given in

the Tables is for a
Load Number of Average Load distribution network.
Point customers Connected (kW)(Lavg,i) Assume that the duration
(Ni)
of interruption is the same
1 250 2300 as the restoration time.
2 300 3700 Determine the following
3 200 2500 reliability indices:
4 250 1600 a. SAIFI, b. CAIFI, c. SAIDI, d.
CAIDI, e. ASAI, f. ASIDI, g.
ENS, h. AENS, i. ACCI
NT=1,000 LT=10100

Curtsey: T. Gonen. Electric Power Distribution System Engineering

 28 Example…
Annual Interruption Events
Load Point Number of Customers Load Interrupted
Affected Interrupted (Ni) (kW) (Li)
 The information given in
1 250 2300 the Tables is for a
2 200 2500
distribution network.
3 250 1600
4 250 1600
Assume that the duration
950 8000
of interruption is the
same as the restoration
Load point Duration of Customer Energy Not
Affected Interruptions Hours Supplied time. Determine the
(hr) (di=ri) Curtailed (kWh) (riXLi)
(riXNi) following reliability
1 2 500 4600 indices:
2 3 600 7500
a. SAIFI, b. CAIFI, c. SAIDI, d.
3 1 250 1600
CAIDI, e. ASAI, f. ASIDI, g.
4 1 250 1600
ENS, h. AENS, i. ACCI
1600 15300

CN, number of customers affected = 250+200+250 = 700

29 Solution…

𝑁𝑖 950
a. SAIFI = = = 0.95 interruptions/customer served
𝑁𝑇 1000

𝑁𝑖 950
b. CAIFI = = =1.357 interruptions/customer affected
𝐶𝑁 700

𝑟𝑖 𝑁𝑖 1600
c. SAIDI = = =1.6 hr/customer served = 96
𝑁𝑇 1000
min/customer served

𝑟𝑖 𝑁𝑖 1600
d. CAIDI = = =1.684 hr/customer interrupted = 101.05
𝑁𝑖 950
min/customer interrupted
30 Solution…

𝑁𝑇8760− 𝑟𝑖𝑁𝑖 1000×8760−1600

e. ASAI = = = 0.999817 unit hour
𝑁𝑇 8760 1000×8760
availability

𝑟𝑖𝐿𝑖 15300
f. ASIDI = = = 1.515
𝐿𝑇 10100

g. ENS= 𝐿𝑎𝑣𝑔,𝑖 𝑟𝑖 = 15300 kWh

𝐸𝑁𝑆 15300
h. AENS = = = 15.3 kWh/customer affected
𝑁𝑇 1000

𝐸𝑁𝑆 15300
i. ACCI = = = 21.857 kWh/customer affected
𝐶𝑁 700
31 Exercise:
Annual interruption event data
Load data Load point Number of Load Duration of
Load Number of Average load affected customers interrupted interruption
point customers connected (kW) interrupted (kW) (hr)
1 1800 8400 2 800 3600 3
2 1300 6000 3 600 2800 3
3 900 4600 3 300 1800 2
3 600 2800 1
Given that, number of customers 2 500 2400 1.5
interrupted/affected=2200 3 300 1800 1.5

Determine SAIFI, CAIFI, SAIDI, CAIDI, ASAI, ASIDI, ENS, AENS, ACCI
32
33 Basic mathematical concepts of
reliability: General Reliability function
 The probability of failure of a given component (or
system) is defined as a function of time as:

𝑃 𝑇≤𝑡 =𝐹 𝑡 𝑡≥0

where,
T is a random variable representing the failure time
F(t) is the probability that component will fail by time t
34 Basic mathematical concepts of
reliability: General Reliability function
 Here, F(t) is the failure distribution function, which can be
called as the unreliability function. Therefore, the probability
that the component will not fail at a given time t is defined as
the reliability of the component. Thus, the reliability function
can be expressed as:

𝑅 𝑡 =1−𝐹 𝑡
where,
R(t) is the reliability function
F(t) is the unreliability function
35 Basic mathematical concepts of
reliability: General Reliability function
 If the time-to-failure random variable T has a
density function f(t),

𝑅 𝑡 = 1 − 𝐹(𝑡)
𝑡 ∞

=1− 𝑓 𝑡 𝑑𝑡 = 𝑓 𝑡 𝑑𝑡
0 𝑡
36 Basic mathematical concepts of
reliability: General Reliability function
 The probability of failure of a given system in a particular
time interval (t1, t2) can be written in terms of the
unreliability function, as:

𝑡2 𝑡2 𝑡1
𝑓 𝑡 𝑑𝑡 = 𝑓 𝑡 𝑑𝑡 − 𝑓 𝑡 𝑑𝑡 = 𝐹 𝑡2 − 𝐹(𝑡1 )
𝑡1 0 0
37 Basic mathematical concepts of
reliability: General Reliability function
 The probability of failure of a given system in a particular
time interval (t1, t2) can be written in terms of the
reliability function, as:

𝑡2 ∞ ∞
𝑓 𝑡 𝑑𝑡 = 𝑓 𝑡 𝑑𝑡 − 𝑓 𝑡 𝑑𝑡 = 𝑅 𝑡1 − 𝑅(𝑡2 )
𝑡1 𝑡1 𝑡2
38 Basic mathematical concepts of
reliability: General Reliability function
 The rate of failures in a given time interval (t1, t2) is defined as the
hazard rate, or failure rate, during that interval. It is the probability
that a failure per unit time happens in the interval, provided that
a failure has not happened before the time t1,i.e., at the
beginning of the time interval. Therefore,

𝑅 𝑡1 − 𝑅(𝑡2 )
ℎ 𝑡 =
𝑡2 − 𝑡1 𝑅(𝑡1 )
39 Basic mathematical concepts of
reliability: General Reliability function
Let, the time interval be defined as: 𝑡1 = 𝑡, 𝑡2 = 𝑡 + ∆𝑡

𝑅 𝑡 − 𝑅(𝑡 + ∆𝑡)
ℎ 𝑡 = lim
∆𝑡→0 ∆𝑡. 𝑅(𝑡)

1 −𝑑𝑅(𝑡)
=
𝑅(𝑡) 𝑑𝑡

𝑓(𝑡)
=
𝑅(𝑡)

𝑑𝑅(𝑡)
Where, f(t) is the probability density function −
𝑑𝑡
40 Basic mathematical concepts of
reliability: General Reliability function
 By substituting R(t) by F(t), we get,
𝑓(𝑡) 𝑑𝐹(𝑡)
ℎ 𝑡 = Hence, ℎ 𝑡 𝑑𝑡 =
1 − 𝐹(𝑡) 1 − 𝐹(𝑡)

Hence, 𝑡
𝑡
ℎ 𝑡 𝑑𝑡 = −ln[1 − 𝐹(𝑡)]
0 0

𝑡
1 − 𝐹 𝑡 = 𝑒𝑥𝑝 − ℎ 𝑡 𝑑𝑡
0
𝑡
Taking derivatives of this Equation, we get, 𝑓 𝑡 = ℎ 𝑡 𝑒𝑥𝑝 − ℎ 𝑡 𝑑𝑡
0
41 Basic mathematical concepts of
reliability: General Reliability function
 Thus, the generalized expression for the reliability
function can be written as,
𝑡
𝑅 𝑡 = 𝑒𝑥𝑝 − ℎ 𝑡 𝑑𝑡
0

ℎ 𝑡 = 𝜆 failures/unit time

 Assume that the hazard or failure function is

independent of time, i.e.,

We get the failure density function as, 𝑓 𝑡 = λ𝑒 −𝜆𝑡

42 Basic mathematical concepts of reliability:
basic single-component concepts

 Theoretically, the expected life, that is, the expected time

during which a component will survive and perform
successfully, can be expressed as:
∞
𝐸 𝑇 = 𝑡𝑓 𝑡 𝑑𝑡
0

∞
𝑑𝑅(𝑡)
𝐸 𝑇 =− 𝑡 𝑑𝑡
0 𝑑𝑡

∞ ∞ since, 𝑅(𝑡=0) =1 𝑅(𝑡=∞) =0

𝐸 𝑇 = −𝑡𝑅 𝑡 + 𝑅(𝑡)𝑑𝑡
0 0
43 Basic mathematical concepts of reliability:
basic single-component concepts

Therefore, the expected life can be expressed as:

∞ ∞ 𝑡
𝐸 𝑇 = 𝑅 𝑡 𝑑𝑡 Or, 𝐸 𝑇 = 𝑒𝑥𝑝 − 𝜆 𝑡 𝑑𝑡 𝑑𝑡
0 0 0

When there is a constant failure rate,

∞
−𝜆𝑡
1
𝐸 𝑇 = 𝑒 𝑑𝑡 =
0 𝜆
44 Basic mathematical concepts of
reliability: MTTF and MTBF
 If the system is simply replaced by a good system (i.e., no
maintenance required) the E(T) useful life is also defined as
the mean time to failure (MTTF) and denoted as:
1
𝑀𝑇𝑇𝐹 = 𝑚 =
𝜆
 If the system has undergone maintenance and repairs the
useful life, E(T) is also defined as the mean time between
failures (MTBF) and denoted as:
Where
𝑇 is the mean cycle time
𝑀𝑇𝐵𝐹 = 𝑇 = 𝑚 + 𝑟 𝑚 is the mean time to failure
𝑟 is the mean time to repair
45 Basic mathematical concepts of
reliability: MTTR
 the mean time to repair (MTTR) is defined as the
reciprocal of the average (or mean) repair rate and
denoted as:
1
𝑀𝑇𝑇𝑅 = 𝑟 =
𝜇

where, μ is the mean repair rate.

46 Computation of MTTR and MTTF in a
two-state model
 Consider a two-state model
shown in Figure. Assume that
the system is either in the up
(or in) state or in the down (or
out) state at a given time, as
shown in Figure. Therefore,
the mean time to failure
(MTTF) can be reasonably
estimated as:
𝑛
𝑖=1 𝑚𝑖
Where, MTTF= 𝑚 =
𝑚 is the mean time to failure 𝑛
𝑚𝑖 is the observed time to failure for ith cycle
n is the total number of cycles
47 Computation of MTTR and MTTF in a
two-state model
 Similarly, the mean time to
repair can be reasonably
estimated as:
𝑛
𝑖=1 𝑟𝑖
MTTR= 𝑟 =
𝑛
where,
𝑟 is the mean time to repair
𝑟𝑖 is the time to repair for ith cycle
n is the total number of cycles

Thus, we can write: MTBF=MTTF+MTTR

48 Mean cycle time

 The mean cycle time defines the average time that it

takes for the component to complete one cycle of
operation, that is, failure, repair, and restart.
𝑇 =𝑚+𝑟
1 1 𝜆+𝜇
𝑇= + or, 𝑇=
𝜆 𝜇 𝜆𝜇
 The reciprocal of the mean cycle time is defined as the
mean failure frequency and denoted as:
1 𝜆𝜇
𝑓= 𝑓=
𝑇 𝜆+𝜇
49 Availability and Unavailability

 In two-state model, it can be assumed that the component

is either up (i.e., healthy condition) or down (i.e., unavailable
for service). Therefore, it can be shown that:

𝐴+𝑈 =1

where,
A is the availability of component, i.e., the fraction of time component is up/healthy
𝐴 = 𝑈 is the unavailability of component, i.e., the fraction of time component is down
50 Availability and Unavailability

 Therefore, on the average, as time t goes to

infinity, it can be shown that the availability is:

𝑚 𝑀𝑇𝑇𝐹
𝐴≜ =
𝑇 𝑀𝑇𝐵𝐹
or 𝑚 𝑀𝑇𝑇𝐹
𝐴≜ =
𝑚 + 𝑟 𝑀𝑇𝑇𝐹 + 𝑀𝑇𝑇𝑅
 For constant failure rate and repair rate,
𝜇
𝐴=
𝜆+𝜇
51 Availability and Unavailability

 The unavailability can be expressed as:

𝑈 ≜1−𝐴
𝑟 𝑀𝑇𝑇𝑅
𝑈= =
𝑟 + 𝑚 𝑀𝑇𝑇𝐹 + 𝑀𝑇𝑇𝑅

 For constant failure rate and repair rate,

𝜆
𝑈=
𝜆+𝜇
52 Reliability of Unrepairable components
in series
 Two independent components are connected in series.
Therefore, the whole system will successfully operate (or
remains healthy) if both components remain healthy,
i.e., without a failure.
𝑅𝑠𝑦𝑠 =𝑃 [𝐸1 𝐸2]

𝑅𝑠𝑦𝑠 =𝑃 (𝐸1)𝑃(𝐸2)]

𝑅𝑠𝑦𝑠 = 2
𝑖=1 𝑅𝑖
Where
Ei is the event that component i (or subsystem i) operates successfully
Ri = P(Ei) is the reliability of component i (or subsystem i)
Rsys is the reliability of system (or system reliability index)
53 Reliability of Unrepairable components
in series
 For a series system with n independent components, the
generalized expression for the system reliability can be
written as:
𝑛
𝑅𝑠𝑦𝑠 = 𝑖=1 𝑅𝑖

 Note that, this is the product rule or the chain rule of

reliability. System reliability will always be less than or
equal to the least-reliable component, i.e.,

𝑅𝑠𝑦𝑠𝑖 ≤ min{𝑅𝑖 }
54 Reliability of repairable components in
series
 If two independent and repairable components are in series,
the availability or the steady-state probability of success (i.e.,
operation) of the system can be expressed as:

𝐴𝑠𝑦𝑠 = 𝐴1 𝐴2

where,
𝐴𝑠𝑦𝑠 is the availability of system
𝐴1 is the availability of component 1
𝐴2 is the availability of component 2
55 Reliability of repairable components in
series
𝑚1 𝑚2
 Since we know that, 𝐴1 = and 𝐴2 =
𝑟1 + 𝑚1 𝑟2 + 𝑚2

𝑚1 𝑚2
 We can write, 𝐴𝑠𝑦𝑠 = 𝐴1 𝐴2 =
𝑟1 + 𝑚1 𝑟2 + 𝑚2

Where
𝑚1 is the mean time to failure of component 1
𝑚2 is the mean time to failure of component 2
𝑚𝑠𝑦𝑠 is the mean time to failure of system
𝑟1 is the mean time to repair of component 1
𝑟2 is the mean time to repair of component 2
𝑟𝑠𝑦𝑠 is the mean time to repair of system
56 Reliability of repairable components in
series
 The average frequency of the system failure is the sum of
the average frequency of component-1 failing, given that
component 2 is healthy, plus the average frequency of
component 2 failing while component 1 is healthy. Thus,

𝑓𝑠𝑦𝑠 =𝐴2 𝑓1 + 𝐴1 𝑓2

where, 𝑓𝑠𝑦𝑠 is the average frequency of system failure

𝑓𝑖 is the average frequency of failure of component 𝑖
𝐴𝑖 is the availability of component i
57 Reliability of repairable components in
series
1 𝑚𝑖
 Since we know that, 𝑓𝑖 = and 𝐴𝑖 =
𝑚𝑖 + 𝑟𝑖 𝑟𝑖 + 𝑚𝑖

1 𝑚2 1 𝑚1
 We can write, 𝑓𝑠𝑦𝑠 = x + x
𝑚1+𝑟1 𝑚2+𝑟2 𝑚2+𝑟2 𝑚1+𝑟1

𝐴𝑠𝑦𝑠 = 𝑚𝑠𝑦𝑠 𝑓𝑠𝑦𝑠

Thus, the mean time to failure for a given series system with
two components can be expressed as
1
𝑚𝑠𝑦𝑠 =
1/𝑚1 + 1/𝑚2
58 Reliability of repairable components in
series
 Hence, the mean time to failure of a given series system
with n components can be expressed as:
1
𝑚𝑠𝑦𝑠 =
1/𝑚1 + 1/𝑚2 + ⋯ + 1/𝑚𝑛
 Hence, the mean time to failure of a given series system
with n components can be expressed as:
𝜆𝑠𝑦𝑠 =𝜆1 +𝜆2 +𝜆3 +…+𝜆𝑛
Similarly, the mean time to repair for an n-component
series system is: 𝜆 𝑟 + 𝜆 𝑟 + 𝜆 𝑟 + ⋯+ 𝜆 𝑟
1 1 2 2 3 3 𝑛 𝑛
𝑟𝑠𝑦𝑠 =
𝜆𝑠𝑦𝑠
59
60 Reliability of unrepairable components
in parallel
 Figure shows a block diagram for a system
that has two independent components
connected in parallel.
 Therefore, to have the system fail and not
be able to perform its desired function,
both components must fail simultaneously.
 Thus, the system unreliability is

𝑄𝑠𝑦𝑠 = 𝑃[𝐸1 ∩ 𝐸2 ] 𝑄𝑠𝑦𝑠 = 𝑃 𝐸1 𝑃(𝐸2 )]

2 Where
𝐸𝑖 is the event that component i fails
𝑄𝑠𝑦𝑠 = (1 − 𝑅𝑖 ) 𝑄𝑖 = 𝑃 𝐸𝑖 is the unreliability of component i
𝑖=1 𝑄𝑠𝑦𝑠 is the unreliability of system (or system unreliability index)
61 Reliability of unrepairable components
in parallel
 For m independent components, the system reliability
expression is:

𝑅𝑠𝑦𝑠 =1 − 𝑄𝑠𝑦𝑠

=1 − [𝑄1 𝑄2 𝑄3 …𝑄𝑚 ]= 1 − 𝑚
𝑖=1 𝑄𝑖

=1 −[(1 − 𝑅1 )(1 − 𝑅2 ) (1 − 𝑅3 )…(1 − 𝑅𝑚 )]

62 Reliability of repairable components in
parallel
 Consider a parallel system with two independent and
repairable components.
 Therefore, the unavailability or the steady-state
probability of failure of the system can be expressed as:
𝑈𝑠𝑦𝑠 = 𝑈1 𝑈2
𝜆1 𝑟1 𝜆2 𝑟2
where, 𝑈1 = 1 − 𝐴1 = 𝑈2 = 1 − 𝐴 2 =
1+𝜆1 𝑟1 1 + 𝜆2 𝑟2

𝜆1 𝑟1 𝜆2 𝑟2
𝑈𝑠𝑦𝑠 =
1+𝜆1 𝑟1 1+𝜆2 𝑟2
63 Reliability of repairable components in
parallel
 However, the average frequency of the system failure is:

𝑓𝑠𝑦𝑠 = 𝑈2 𝑓1 +𝑈1 𝑓2

where
𝑓𝑠𝑦𝑠 is the average frequency of system failure
𝑓𝑖 is the average frequency of failure of component I
𝑈𝑖 is the unavailability of component i
64 Reliability of repairable components in
parallel
𝜆1 𝜆2
 Since, 𝑓1 = 1 + 𝜆 𝑟 𝑓2 =
1 + 𝜆2 𝑟2
1 1

 We get the following expression:

𝜆1 𝜆2 (𝑟1 + 𝑟2 )
𝑓𝑠𝑦𝑠 =
(1 + 𝜆1 𝑟1 )(1 + 𝜆2 𝑟2 )

 The system unavailability can be expressed as:

𝑟𝑠𝑦𝑠 𝑈𝑠𝑦𝑠 = 𝑟𝑠𝑦𝑠 𝑓𝑠𝑦𝑠

𝑈𝑠𝑦𝑠 ≜
𝑇𝑠𝑦𝑠
65 Reliability of repairable components in
parallel
 The average repair time (or downtime) of the two-
component parallel system can be expressed as:

𝑟1 𝑟2
𝑟𝑠𝑦𝑠 =
𝑟1 + 𝑟2
 The system unavailability can be expressed as

𝑟𝑠𝑦𝑠
𝑈𝑠𝑦𝑠 ≜
𝑟𝑠𝑦𝑠 + 𝑚𝑠𝑦𝑠
66
67 Reliability of repairable components in
parallel
 The average time to failure (or operation time, or
uptime) of the parallel system can be expressed as:
1 + 𝜆1 𝑟1 + 𝜆2 𝑟2
𝑚𝑠𝑦𝑠 =
𝜆1 𝜆2 (𝑟1 + 𝑟2 )
1
 The failure rate of the parallel system is 𝜆𝑠𝑦𝑠 ≜
𝑚𝑠𝑦𝑠

𝜆1 𝜆2 (𝑟1 + 𝑟2 )
𝜆𝑠𝑦𝑠 =
1 + 𝜆1 𝑟1 + 𝜆2 𝑟2
68 Example…

 The Figure shows a 4 mile long distribution feeder.

Approximately 1 mile of the feeder has been built underground,
while the rest of the feeder is overhead. The underground
feeder has two termination points. Two faults per circuit-mile for
the overhead section and one fault per circuit-mile for the
underground section of the feeder have been recorded in the
last 10 years. The annual cable termination fault rate is given as
0.5% per cable termination. Furthermore, based on past
experience, it is known that the average repair times for the
overhead section, underground section, and each cable
termination are 3, 30, and 3 h, respectively.
Ref.: T. Gonen. Electric Power Distribution System Engineering
69 Example…

Using the given information,

determine the following:
 a. Total annual fault rate of
the feeder
 b. Average annual fault
restoration time of the
feeder in hours
 c. Unavailability of the
feeder
 d. Availability of the feeder
Ref.: T. Gonen. Electric Power Distribution System Engineering
70 Solution…

 a. Total annual fault rate of the feeder is

𝜆𝐹𝐷𝑅 = 3
𝑖=1 𝜆𝑖 = 𝜆𝑂𝐻 + 𝜆𝑈𝐺 + 2𝜆𝐶𝑇

Where
𝜆𝑂𝐻 is the total annual fault rate of overhead section of feeder
𝜆𝑈𝐺 is the total annual fault rate of underground section of feeder
𝜆𝐶𝑇 is the total annual fault rate of cable terminations

2 1
𝜆𝐹𝐷𝑅 =3 +1 + 2 0.005
10 10

=0.71 faults/year
71 Solution…

 b. Average fault restoration time of the feeder per fault is:

𝑟𝐹𝐷𝑅 = 3
𝑖=1 𝑟𝑖 = 𝑟𝑂𝐻 + 𝑟𝑈𝐺 + 2𝑟𝐶𝑇

Where
𝑟𝑂𝐻 is the average repair time for overhead section of feeder, h
𝑟𝑈𝐺 is the average repair time for underground section of feeder, h
𝑟𝐶𝑇 is the average repair time per cable termination, h

𝑟𝐹𝐷𝑅 =3+30+2(3)
=39 hr
72 Solution…

 However, the average annual fault restoration time of

the feeder is:
3
𝑖=1 𝜆𝑖 𝑟𝑖
𝑟𝐹𝐷𝑅 = 3
𝑖=1 𝜆𝑖

𝑙𝑂𝐻 𝜆𝑂𝐻 𝑟𝑂𝐻 + 𝑙𝑈𝐺 𝜆𝑈𝐺 𝑟𝑈𝐺 + (2𝜆𝐶𝑇 )(𝑟𝐶𝑇 )

𝑟𝐹𝐷𝑅 =
𝜆𝐹𝐷𝑅
3×0.2 3 + 1×0.1 30 +(2×0.005)(3) 4.83
= =
0.71 0.71

=6.8 hr
73 Solution…

 Unavailability of the feeder is:

6.8
𝑟𝐹𝐷𝑅 𝑈𝐹𝐷𝑅 =
𝑈𝐹𝐷𝑅 = 6.8 + 12338
𝑟𝐹𝐷𝑅 +𝑚𝐹𝐷𝑅

= 0.055%
1 1
𝑚𝐹𝐷𝑅 = = 8760 ∗ = 12338
𝜆𝐹𝐷𝑅 0.71
74 Solution…

 d. Availability of the feeder in percentage is:

𝐴𝐹𝐷𝑅 = 100 − 𝑈𝐹𝐷𝑅

=100−0.055

= 99.945%
75 Reliability for the Parallel-Series Combination
Simple combinations of series and parallel
subsystems (or components) can be analyzed by
successively reducing subsystems into equivalent
parallel or series components.
76 Reliability for the Parallel-Series Combination
 The equivalent reliability of the system with m
parallel paths of n components each can be
expressed as:
𝑅𝑠𝑦𝑠 =1 −(1 − 𝑅𝑛 )𝑚

Where
𝑅𝑠𝑦𝑠 is the equivalent reliability of system
𝑅𝑛 is the equivalent reliability of a path
𝑅 is the reliability of a component
𝑛 is the total number of components in a
path
𝑚 is the total number of paths
77 Reliability for the Series-Parallel Combination

The equivalent reliability of the system of n series units

(or banks) with m parallel components in each unit
(or bank) can be expressed as:

𝑅𝑠𝑦𝑠 =[1 −(1 − 𝑅)𝑚 ]𝑛

Where
𝑅𝑠𝑦𝑠 is the equivalent reliability of system
1-(1 − 𝑅)𝑚 is the equivalent reliability of a parallel
unit (or bank)
𝑅 is the reliability of a component
𝑚 is the total number of components in a parallel
unit (or bank)
𝑛 is the total number of units (or banks)
78 Example…
 Consider the various combinations of the reliability
block diagrams shown in Figure. Assume that they are
based on the logic diagrams of each subsystem and
that the reliability of each component is 0.85. Determine
the equivalent system reliability of each configuration.

Ref.: T. Gonen. Electric Power Distribution System Engineering

79 Example…
80 Solution…

 (a) The equivalent system reliability for the series system is:

𝑅𝑒𝑞 = 𝑅𝑠𝑦𝑠 = 4
𝑖=1 𝑅𝑖 = 0.85 4 = 0.5220

 (b) The equivalent system reliability for the parallel–series

system is: 𝑅𝑒𝑞 = 1 − (1 − 𝑅4)2= 1 − 1 − 0.85 4 2
= 0.7715
81 Solution…

(c) The equivalent 𝑅𝑒𝑞 = [1 − (1 − 𝑅2 )2 [1 − (1 − 𝑅2 )2]

system reliability = [1 − (1 − 0.852 )2 [1 − (1 − 0.852 )2]
for the mixed-
parallel system is: = 0.8519

𝑅𝑒𝑞 = [1 − (1 − 𝑅)2 [1 − (1 − 𝑅3 )2]

(d) The equivalent
system reliability = [1 − (1 − 0.85)2 [1 − (1 − 0.853 )2]
for the mixed-
= 0.8320
parallel system is:
82 Solution…

(e) The equivalent system reliability for the series–

parallel system from Equation is:

𝑅𝑒𝑞 = [1−(1−R)2]4

=[1−(1−0.85)2]4

= 0.9130
83 Example…
 Assume that a system has five components, namely, A, B, C, D,
and E, as shown in Figure, and that each component has
different reliability as indicated in the figure. Determine the
following:
 a. The equivalent system reliability.
 b. If the equivalent system reliability is desired to be at least 0.8, or
80%, design a system configuration to meet this system
requirement by using each of the five components at least once.

Ref.: T. Gonen. Electric Power Distribution System Engineering

84 Solution…

 (a) The equivalent system reliability is:

5

𝑅𝑒𝑞 = 𝑅𝑖
𝑖=1

= (0.80)(0.95)(0.99)(0.90)(0.65)

=0.4402 or 44.02%
85 Solution…
 (b) In general, the best way of improving the overall
system reliability is to protect the less-reliable
components with parallel components. Therefore, since
the relatively less-reliable components are A and E,
these can be backed by parallel redundancy as shown
in Figure. Therefore, the new equivalent system reliability
becomes

𝑅𝑠𝑦𝑠 = 5
𝑖=1 𝑅𝑖

= 1 − 1 − 0.80 2 0.95 (0.99) 0.90 [1 − 1 − 0.65 4]

= 0.8004 𝑜𝑟 80.04%
86 Example…
 Assume that a three-phase transformer bank consists of
three single-phase transformers identified as A, B, and C
for the sake of convenience. Assume that (1) transformer
A is an old unit and therefore has a reliability of 0.90, (2)
transformer B has been in operation for the last 20 years
and therefore has been estimated to have a reliability of
0.95, and (3) transformer C is a brand new one with a
reliability of 0.99. Based on the given information and
assumption of independence, determine the following:

Ref.: T. Gonen. Electric Power Distribution System Engineering

87 Example…

 a. The probability of having no failing transformer at any

given time.
 b. If one out of the three transformers fails at any given time,
what are the probabilities for that unit being the transformer A,
or B, or C?
 c. If two out of the three transformers fail at any given time,
what are the probabilities for those units being the
transformers A and B, or B and C, or C and A?
 d. What is the probability of having all three transformers out
of service at any given time?
88 Solution…

 (a) The probability of having all transformers operational

without fail at any given time is:

𝑃 𝐴∩𝐵∩𝐶 =𝑃 𝐴 𝑃 𝐵 𝑃 𝐶

= (0.90)(0.95)(0.99)

= 0.84645
89 Solution…
 (b) If one out of the three transformers fails at any given time, the
probabilities for that unit being the transformer A, or B, or C are:

𝑃 𝐴 ∩ 𝐵 ∩ 𝐶 = 𝑃 𝐴 𝑃 𝐵 𝑃(𝐶)
= 0.10 0.95 0.99
= 0.09405

𝑃 𝐴 ∩ 𝐵 ∩ 𝐶 = 𝑃 𝐴 𝑃 𝐵 𝑃(𝐶)
= 0.90 0.05 0.99
= 0.04455

𝑃 𝐴 ∩ 𝐵 ∩ 𝐶 = 𝑃 𝐴 𝑃 𝐵 𝑃(𝐶)
= 0.90 0.95 0.01
= 0.00855
90 Solution…
 If two out of the three transformers fail at any given time, the
probabilities for those units being the transformers A and B, or
B and C, or C and A are

𝑃 𝐴 ∩ 𝐵 ∩ 𝐶 = 𝑃 𝐴 𝑃 𝐵 𝑃(𝐶)
= 0.10 0.05 0.99
= 0.00495
𝑃 𝐴 ∩ 𝐵 ∩ 𝐶 = 𝑃 𝐴 𝑃 𝐵 𝑃(𝐶)
= 0.90 0.05 0.01
= 0.00045

𝑃 𝐴 ∩ 𝐵 ∩ 𝐶 = 𝑃 𝐴 𝑃 𝐵 𝑃(𝐶)
= 0.10 0.95 0.01
= 0.00095
91 Solution…

 The probability of having all three transformers out of

service at any given time is:

𝑃 𝐴∩𝐵∩𝐶 =𝑃 𝐴 𝑃 𝐵 𝑃 𝐶

= (0.10)(0.05)(0.01)

= 0.00005
92 Summary of the solution…
Summary of the Computations
Number of System Probability
failed Modes
Transformers

0 𝐴∩𝐵∩𝐶 0.84645
1 𝐴∩𝐵∩𝐶 0.09405
𝐴∩𝐵∩𝐶 0.04455
𝐴∩𝐵∩𝐶 0.00855
2 𝐴∩𝐵∩𝐶 0.00495
𝐴∩𝐵∩𝐶 0.00045
𝐴∩𝐵∩𝐶 0.00095
3 𝐴∩𝐵∩𝐶 0.00005

= 1.00000
93 Reliability modelling: Markov process

 In reliability studies, the variable t represents time, and X(t)

describes the state of the system at time t. The states at a
given time tn actually represent the (exhaustive and
mutually exclusive) outcomes of the system at that time.
 A Markov process is a stochastic system for which the
occurrence of a future state depends on the immediately
preceding state and only on it.
 The Markovian process is characterized by a lack of
memory.
94 Reliability modelling: Markov process

 A discrete parameter stochastic process, {X(t); t = 0, 1,

2,…}, or a continuous parameter stochastic process,
{X(t); t ≥ 0}, is a Markov process if it has the following
markovian property:

𝑃 𝑋 𝑡𝑛 = 𝑥𝑛 𝑋 𝑡1 = 𝑥1 , 𝑋 𝑡2 = 𝑥2 , … , 𝑋 𝑡𝑛−1 = 𝑥𝑛−1
= 𝑃 𝑋 𝑡𝑛 = 𝑥𝑛 𝑋 𝑡𝑛−1 = 𝑥𝑛−1
95 Reliability modelling: Markov process
 For any set of n time points, t1 < t2 < … < tn in the index
set of the process, and any real numbers x1, x2,..., xn. The
probability is called the transition probability and
represents the conditional probability of the system
being in xn at tn, given that it was xn−1 at tn−1. It is also
called the one-step transition probability due to the fact
that it represents the system between tn−1, and tn.

𝑃𝑥𝑛−1,𝑥𝑛 = 𝑃 𝑥 𝑡𝑛 = 𝑥𝑛 𝑋 𝑡𝑛−1 = 𝑥𝑛−1

96 Reliability modelling: Markov process

 For any set of n time points, t1 < t2 < … < tn in the index
set of the process, and any real numbers x1, x2,..., xn. The
probability of k-step transition is,

𝑃𝑥𝑛 ,𝑥𝑛+𝑘 = 𝑃 𝑋 𝑡𝑛+𝑘 = 𝑥𝑛+𝑘 𝑋 𝑡𝑛 = 𝑥𝑛

97 Reliability modelling: Markov process

 A Markov chain is defined by a sequence of discrete-

valued random variables, {X(tn)}, where tn is discrete-
valued or continuous. Therefore, one can also define the
Markov chain as the Markov process with a discrete
state space.
𝑝𝑖𝑗 = 𝑃 𝑋 𝑡𝑛 = 𝑗 𝑋 𝑡𝑛−1 = 𝑖

 As the one-step transition probability of going from state

i at tn−1 to state j at tn assumes that these probabilities
do not change over time. The term used to describe this
assumption is stationarity.
98 Reliability modelling: Markov process

 A Markov chain is completely defined by its transition

probabilities, of going from state i to state j, given in a
matrix form: 𝑝00 𝑝01 𝑝02 𝑝03 … 𝑝0𝑛
𝑝10 𝑝11 𝑝12 𝑝13 … 𝑝1𝑛
𝑝20 𝑝21 𝑝22 𝑝23 … 𝑝2𝑛
𝑃= 𝑝 𝑝31 𝑝32 𝑝33 … 𝑝3𝑛
30 𝑛
… … … … … …
𝑝𝑖𝑗 = 1 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑖
𝑝𝑛0 𝑝𝑛1 𝑝𝑛2 𝑝𝑛3 … 𝑝𝑛𝑛
𝑗

𝑝𝑖𝑗 ≥ 0 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑖𝑗

The matrix P is called a one-step transition matrix (or

stochastic matrix), since all the transition probabilities pij’s are
fixed and independent of time.
99 Reliability modelling: Markov process

 Transition diagram:
100 Reliability modelling: Markov process
 Example: The records of distribution utility indicate that
only 2% of the transformers that are presently down and
therefore being repaired now will be down and
therefore will need repair next time. The records also
show that 5% of those transformers that are currently up
and therefore in service now will be down and therefore
will need repair next time. Assuming that the process is
discrete, markovian, and has stationary transition
probabilities, determine the transition diagram.

Ref.: T. Gonen. Electric Power Distribution System Engineering

101 Reliability modelling: Markov process

 Let t and t + 1 represent the present time (i.e., now) and

the next time, respectively. Therefore, the associated
conditional probabilities are:

𝑃 𝑋𝑡+1 = 𝑑𝑜𝑤𝑛 𝑋𝑡 = 𝑑𝑜𝑤𝑛 = 0.02

𝑃 𝑋𝑡+1 = 𝑢𝑝 𝑋𝑡 = 𝑑𝑜𝑤𝑛 = 0.98
𝑃 𝑋𝑡+1 = 𝑑𝑜𝑤𝑛 𝑋𝑡 = 𝑢𝑝 = 0.05
𝑃 𝑋𝑡+1 = 𝑢𝑝 𝑋𝑡 = 𝑢𝑝 = 0.95
102 Reliability modelling: Markov process

 Let numbers 1 and 2 represent the states of down and

up, respectively.

p11 = 0.02 p12 = 0.98

p21 = 0.05 p22 = 0.95

𝑝11 𝑝12 0.02 0.98

𝑃= 𝑝
21 𝑝22 = 0.05 0.95
103 Reliability modelling: Markov process

 State-transition diagram:
104 Reference

 T. Gonen. Electric Power Distribution

System Engineering; CRC Press, 3rd Edition,
2014.
105

Thank You…