Basics of Reliability

Theory
Introduction, Decomposition of an Engineered Object,
Functions, Failures, and Faults, Characterization of
Degradation, Reliability Concept and Characterization, Linking
System and Component Failures
George Ebot Etta 1
 Introduction
• Every engineered object (product, plant, or infrastructure) is unreliable in the
sense that it degrades and eventually fails.
• The reliability of the object is determined by decisions made during the
design and building (manufacturing) of the object and is affected by factors
such as operating environment, usage mode, and intensity.
• Maintenance actions are needed to counteract the unreliability of the object.
Effective maintenance decision making needs to take into account the
reliability of the object
George Ebot Etta 2
 Decomposition of an Engineered Object
• Even the simplest engineered product (for example, a small consumer product is
comprised of several interacting elements and can be viewed as a multi‐level
system.

George Ebot Etta 3

 Decomposition of an Engineered Object

George Ebot Etta 4

 Decomposition of an Engineered Object

George Ebot Etta 5

 Functions, Failures, and Faults
• Functions
 Essential function - intended or primary function e.g. thermal power plant -
to provide electrical power.
 Auxiliary functions - required to support the primary function e.g. preserving
fluid integrity is an auxiliary function of a pump.
 Protective functions - two‐fold goal here is to protect people from injury and
protect against damage to the environment - relays offer protection against
current surges and scrubbers on smokestacks that remove particulate matter
to protect the environment.
George Ebot Etta 6
 Functions, Failures, and Faults
• Functions
 Information functions - comprise condition monitoring, gauges, alarms, and
so on. In a power plant, the main control panel displays various bits of
information about the different subsystems – for example, voltage and
current output of generators, pressure and temperature of steam in the
various parts of the plant, and so on.

George Ebot Etta 7

 Functions, Failures, and Faults
• Failures
 Failure is the termination of the ability of an item to perform a
required function (IEC 50(191), 1990)
 A system failure occurs due to the failure of one or more of its
components
 Primary failure - failure of a component due to natural causes (for
example, aging)

George Ebot Etta 8

 Functions, Failures, and Faults
 Secondary failure - failure of a component due to one or more of the
following causes: (i) the (primary) failure of some other component(s) in the
system, (ii) environmental factors, and/or (iii) actions of the user.
 Command failure - occurs when a component is in the non‐working (rather
than a failed) state because of improper control signals or noise e.g. a faulty
action of a logic controller switching off a pump).

George Ebot Etta 9

 Functions, Failures, and Faults
• Failures
Example: Consider the stereo system
shown in Fig. below, which is made up
of a CD player, tuner, amplifier, and two
speakers for stereophonic sound.
The system is working if one can listen
to either a CD or the radio (tuner) in
stereophonic sound.
George Ebot Etta 10
 Functions, Failures, and Faults
 Partial failures - Failures which result in a partial loss of function.
o One can listen to either a CD or the radio in monophonic sound (as one speaker
has failed).
o One can listen to only one source in stereophonic sound (either the CD or the radio
has failed).
o One can listen to only one source in monophonic sound (either the CD or the radio
and one of the speakers has failed).
 Complete Failure - Failures which result in a total loss of function.
o No sound (both the CD and the radio or the amplifier or both speakers have failed).
George Ebot Etta 11
 Functions, Failures, and Faults
• Faults
A fault is the state of an item characterized by its inability to perform
its required function (IEC 50(191), 1990).
NOTE: this excludes situations arising from preventive maintenance
or any other intentional shutdown.
 A fault is, hence, a state resulting from a failure.
George Ebot Etta 12
 Functions, Failures, and Faults
• Failure (Fault) Modes
A failure mode is a description of a fault (IEC 50(191), 1990) - identified by studying the
(performance) function of the item
 Intermittent failures - Failures that last for only a short time
 Extended failures - Failures that continue until some corrective action rectifies the
failure: Complete failures/Partial failures
 Sudden failures: Failures that occur without any warning.
 Gradual failures: Failures that occur with signals to warn of the occurrence of a failure
George Ebot Etta 13
 Functions, Failures, and Faults
 A complete and sudden failure is called a catastrophic failure and a gradual and partial
failure is designated a degraded failure.
Example
Hydraulic valves are used in refineries to control the flow of liquids. If a valve does not
shut properly, the flow is not reduced to zero and this can be viewed as a partial failure. If a
valve fails to operate (due, for example, to the spring not functioning properly), then the
failure is a complete failure. A valve usually wears out with usage and this corresponds to
a gradual failure.
George Ebot Etta 14
 Functions, Failures, and Faults
• Failure Causes and Severity
 Failure cause is the circumstances during design, manufacture, or use which have led to a
failure (IEC 50(191), 1990) - useful information in the prevention of failures or their
reoccurrence
 Design failure - due to inadequate design;
 Weakness failure: - due to weakness (inherent or induced) in the system so that the system
cannot stand the stress it encounters in its normal environment;
 Manufacturing failure - due to non‐conformity during manufacturing;
 Aging failure - due to the effects of age and/or usage;
 Misuse failure - due to misuse of the system (operating in environments for which it was not
designed);
George Ebot Etta 15
 Functions, Failures, and Faults
 Mishandling failure - due to incorrect handling and/or lack of care and
maintenance.
 The severity of a failure mode signifies the impact of the failure mode on the
system as a whole and on the outside environment
 Catastrophic - failures that result in death or total system loss;
 Critical - failures that result in severe injury or major system damage;
 Marginal - failures that result in minor injury or minor system damage;
 Negligible - failures that result in less than minor injury or system damage.
George Ebot Etta 16
 Characterization of Degradation
• The degradation in performance of an engineered object can be characterized by a
variable, X(t)
• X(t) indicates the state or condition of the item (system, component, or something in
between) as a function of age
 Two‐State Characterization (binary‐state characterization)
 X(t) = 1 corresponds to the object being in the working state (performance satisfactory
or acceptable)
 X(t) = 0 corresponds to the object being in the failed state (performance is unsatisfactory
or unacceptable).

George Ebot Etta 17

 Characterization of Degradation
• The item starts in the working state and changes to the failed state after a period T,
as shown
• T is the time to failure (or lifetime of the
component).
• T is a random variable, as the time instant
of change from working to failed is uncertain.
• Some items for which this characterization is
appropriate are (i) an electric bulb and (ii) the heating element in a kettle

George Ebot Etta 18

 Characterization of Degradation
• Multi‐State Characterization (Finite Number of States)
This is an extension of the above case, with 𝐾 = ∞ . X(t) is now a non‐decreasing,
continuous‐ time stochastic process. A higher value of X(t) implies greater degradation
and the item failure time is given by 𝑇 = inf{𝑡: 𝑋 𝑡 = 𝑥 ∗ } - Inf{} (infimum of a set) is
the greatest lower bound of the set
 X(t) can assume values from the set {1 2 , ,…,K} with:
 X(t) = 1 corresponding to item performance being fully acceptable (item is in a good
working state)
George Ebot Etta 19
 Characterization of Degradation
 X(t) = I, 𝑖 <1< 𝐾, corresponding to item performance being partially acceptable
(item is in a working state, with a higher value of i implying a higher level of
degradation)
 X (t) = K corresponding to item performance being unacceptable (item is in a
failed state).
As T increases
degradation
increases

George Ebot Etta 20

 Reliability Concept and Characterization
• The reliability of an item conveys the concept of dependability, successful operation
or performance, and the absence of failures. It is an external property of great
interest to both manufacturer and consumer.
• Unreliability (or lack of reliability) conveys the opposite.
 The reliability of an item is the probability that the item will perform its intended
function for a specified time period when operating under normal (or stated)
environmental conditions.
 design reliability - operational conditions are the same as the nominal conditions
 Field reliability - operational condition will differ from the nominal design
conditions
George Ebot Etta 21
 Reliability Concept and Characterization
• Time to First Failure
Time to first failure is the time elapsed between when a new item is put into operation
(in continuous mode) and when it fails for the first time
 Here, we assume continuous usage and an age‐based clock, where t = 0
corresponds to the object being put into operation
 The time to failure is a random variable, T, with cumulative distribution function,
F(t).
 Note that the item is in the working state at time t if 𝑇 > 𝑡 and in the failed state if
𝑇 ≤ 𝑡 and no action is being initiated to rectify the failure

George Ebot Etta 22

 Reliability Concept and Characterization
• 𝐹 𝑡 = 𝑃𝑟 𝑇 ≤ 𝑡
• he probability density function for time to failure
𝑑𝐹(𝑡)
𝑓 𝑡 =
𝑑𝑡

George Ebot Etta 23

 Reliability Concept and Characterization
• Reliability Function - The reliability of an item is the probability that the item
does not fail before t
𝑅 𝑡 =1−𝐹 𝑡 =𝑃 𝑇 >𝑡 , t>0
 R(t) is a non‐increasing function of
t, 0 ≤ 𝑡 < ∞
 𝑅(0) = 1 and 𝑅 ∞ = 0.

George Ebot Etta 24

 Reliability Concept and Characterization
• Mean Time to First Failure
The mean time to first failure for an item is the expected value of the random
variable T

George Ebot Etta 25

 Reliability Concept and Characterization
• Failure Rate Function - Consider an item that has survived an interval of
time [0, t]. The conditional probability that the item will fail in the time
interval ( , t t t], given that it is functioning at time t is given by

George Ebot Etta 26

 Linking System and Component Failures
• A system is a collection of interconnected components.
• The failure of a system is due to the failure of one or more of the components of the
system. The linking of component failures to system failures can be done using two
different approaches.
 Forward (or bottom‐up) approach - starts with failure events at the component level and
then proceeds forward to the system level to evaluate the consequences of such failures on
system performance. Failure modes and effects analysis (FMEA) uses this approach.
 Backward (top‐down) approach - starts at the system level and then proceeds downward to
the part level to link system performance to failures at the part level. Fault tree analysis
(FTA) uses this approach.
George Ebot Etta 27
 Linking System and Component Failures
• Failure modes and effects analysis (FMEA) is a structured, logical, and systematic
approach involves in reviewing a system in terms of its subsystems, assemblies, and
so on, down to the component level, to identify failure modes and causes and the
effects of such failures on a system’s function
• Objectives of FMEA (IEEE Standard 352)
 To ensure that all conceivable failure modes and their effects on operational success
of the system have been considered;
 To list potential failures and identify the magnitude of their effects;
George Ebot Etta 28
 Linking System and Component Failures
• FMEA Procedure
The FMEA methodology is based on a hierarchical, inductive approach to analysis; to
determine how every possible failure mode of every system component affects the
system operation.
 Determining the item functions
 Identifying all item failure modes – usually these are the ways in which the item fails
to perform its functions
 Determining the effect of the failure for each failure mode, both on the component
and on the overall system being analyzed;

George Ebot Etta 29

 Linking System and Component Failures
 Classifying the failure by its effects on the system operation and mission;
 Determining the failure’s probability of occurrence;
 Identifying how the failure mode can be detected (this is especially
important for fault‐tolerant configurations);
 Identifying any design changes to eliminate the failure mode, or if that is not
possible, mitigate or compensate for its effects.

George Ebot Etta 30

 Linking System and Component Failures
• Illustrative FMEA Worksheet

George Ebot Etta 31

 Linking System and Component Failures
• Completion of an FMEA worksheet for a photocopier failure mode.

George Ebot Etta 32

 Linking System and Component Failures
• Fault Tree Analysis (FTA) is a deductive (top‐down) method of analysis
aimed at pinpointing the causes or combinations of causes that can lead to
the defined top event
• Objectives of FTA
 Identification of the causes or combinations of causes leading to the top
event
 Determination of whether a particular system reliability measure meets a
stated requirement
George Ebot Etta 33
 Linking System and Component Failures
 Determination of which potential failure mode(s) or factor(s) would be the
highest contributor to the system probability of failure (unreliability) or
unavailability, when a system is repairable, for identifying possible system
reliability improvements; and
 Analysis and comparison of various design alternatives to improve system
reliability

George Ebot Etta 34

 Linking System and Component Failures
• Construction of a Fault Tree
The construction of a fault tree begins with the top event and proceeds
downward to link to basic events through the use of different gates.

George Ebot Etta 35

 Linking System and Component Failures
• Basic fault tree symbols

George Ebot Etta 36

 Linking System and Component Failures
• Example: Electric Motor

George Ebot Etta 37

 Linking System and Component Failures
• The top event is “motor fails to operate.”
 One primary failure is the failure of the motor itself
 This event is a basic event because no details of the motor are given and the
event cannot be developed further.
 The other possibility is the event that no current is supplied to the motor.
 The event “no current in motor” is the result of other events and is
therefore developed further.
George Ebot Etta 38
 Linking System and Component Failures

George Ebot Etta 39

 Linking System and Component Failures
• Reliability Block Diagram (RBD)
 A reliability block diagram (RBD) is an alternative way of describing a system that
is useful for reliability analysis.
 Each component is represented by a block with two end points.
 When the component is in its working state, there is a connection between the two
end points and this connection is broken when the component is in a failed state.
 A multi‐component system can be represented as a network of such blocks with
two end points.
George Ebot Etta 40
 Linking System and Component Failures
• The system is in a working state if there is a connected path between the two
end points. If no such path exists, then the system is in a failed state.
• Two well‐known network structures are the series and parallel configurations
• In a series configuration (with two or more components), the system is
functioning only when all of its components are functioning.
• In a parallel structure (with two or more components), the system is not
working (i.e., is in a failed state) only when all of its components are not
working.
George Ebot Etta 41
 Linking System and Component Failures
 Series and Parallel Configurations with Two Components

George Ebot Etta 42

 Linking System and Component Failures
• Link between RBD and FTA

George Ebot Etta 43

 Linking System and Component Failures
• Structure Function
 We confine our attention to the case where each component is characterized
as being in either a working or a failed state (the two‐state characterization)

 Let 𝑋 𝑡 = (𝑋1 (𝑡)𝑋2 (𝑡)𝑋3 (𝑡), … … , 𝑋𝑛 (𝑡)) with this n‐dimensional vector
denoting the state of the n components
George Ebot Etta 44
 Linking System and Component Failures
 The state of the system 𝑋𝑠 (𝑡) – a function of the states of the components
denoted by ∅(𝑋(𝑡))
1 𝑖𝑓𝑡ℎ𝑒 𝑠𝑦𝑠𝑡𝑒𝑚 𝑖𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑤𝑜𝑟𝑘𝑖𝑛𝑔 𝑠𝑡𝑎𝑡𝑒
𝑋𝑠 𝑡 = ∅ 𝑋 𝑡 =ቊ
0 𝑖𝑓𝑡ℎ𝑒 𝑠𝑦𝑠𝑡𝑒𝑚 𝑖𝑠 𝑖𝑛 𝑎 𝑓𝑎𝑖𝑙𝑒𝑑 𝑠𝑡𝑎𝑡𝑒
∅ 𝑋 𝑡 is called the structure function of the system
 One can derive the structure function using either a fault tree or the RBD

George Ebot Etta 45

 Linking System and Component Failures
• Series Configuration
The system is in a working state if and only if (iff ) all the components are in
the working state
𝑋𝑠 𝑡 = 1 iff 𝑋𝑖 ) 𝑡 = 1, 1 ≤ 𝑖 ≤ 𝑛
𝑛

∅ 𝑋 𝑡 = 𝑋1 𝑡 𝑋2 𝑡 … … 𝑋𝑛 𝑡 = ෑ 𝑋𝑖 (𝑡)
𝑖=1

George Ebot Etta 46

 Linking System and Component Failures
• Parallel Configuration
The system is in the failed state if and only if all the components are in the failed state
𝑋𝑠 𝑡 = 0 iff 𝑋𝑖 ) 𝑡 = 0, 1 ≤ 𝑖 ≤ 𝑛
𝑛

1−∅ 𝑋 𝑡 = (1 − 𝑋1 𝑡 )(1 − 𝑋2 𝑡 ) … … 𝑋𝑛 𝑡 = ෑ(1 − 𝑋𝑖 𝑡 )

𝑖=1
𝑛

∅ 𝑋 𝑡 = 1 − (1 − 𝑋1 𝑡 )(1 − 𝑋2 𝑡 ) … … 𝑋𝑛 𝑡 = ෑ 1 − (1 − 𝑋𝑖 𝑡 )
𝑖=1

George Ebot Etta 47

 Linking System and Component Failures
• General Configuration
For many RBDs, the results for series and parallel configurations can be used
repeatedly to obtain the structure function
Example: Consider the stereo system given below

George Ebot Etta 48

 Linking System and Component Failures
Let 𝑋 𝑡 = (𝑋1 𝑡 , 𝑋2 𝑡 , 𝑋3 𝑡 , 𝑋4 (𝑡), 𝑋5 (𝑡))
The tuner and CD player can be viewed as a module with structure function
∅ 𝑋12 𝑡 = 1 − (1 − 𝑋1 𝑡 )(1 − 𝑋2 𝑡 )
Similarly, for the two speakers
∅ 𝑋45 𝑡 = 1 − (1 − 𝑋4 𝑡 )(1 − 𝑋5 𝑡 )
These two modules and the amplifier can be viewed as a series system
∅ 𝑋 𝑡 = 𝑋12 𝑡 𝑋3 𝑡 𝑋45 𝑡 =
1 − (1 − 𝑋1 𝑡 )(1 − 𝑋2 𝑡 ) 𝑋3 𝑡 1 − (1 − 𝑋4 𝑡 )(1 − 𝑋5 𝑡 )

George Ebot Etta 49

 Linking System and Component Failures
• System Reliability
𝐸 𝑋𝑠 (𝑡) = 1 𝑥 𝑃 𝑋𝑠 𝑡 = 1 + 0 𝑥 𝑃 𝑋𝑠 𝑡 = 0 = 𝑃 𝑋𝑠 𝑡 = 1
Note that 𝑃 𝑋𝑠 𝑡 = 1 is the reliability 𝑅𝑠 (𝑡)
𝑅𝑠 𝑡 = 𝐸 𝑋𝑠 (𝑡) = ∅( 𝐸(𝑋 𝑡 ) = ∅ 𝑅(𝑡)
Where R 𝑡 = (𝑅1 (𝑡)𝑅2 (𝑡)𝑅3 (𝑡), … … , 𝑅𝑛 (𝑡))

George Ebot Etta 50

 Linking System and Component Failures
 Series Configuration
The reliability of a system with a series configuration (and statistically independent component
failures) is given by:
𝑛

𝑅𝑠 (𝑡) = ∅ 𝑅 𝑡 = 𝑅1 𝑡 𝑅2 𝑡 … … 𝑅𝑛 𝑡 = ෑ 𝑅𝑖 (𝑡)
𝑖=1
Parallel Configuration
𝑛

𝑅𝑠 (𝑡) = ∅ 𝑅 𝑡 = 1 − (1 − 𝑅1 𝑡 )(1 − 𝑅2 𝑡 ) … … 𝑅𝑛 𝑡 = ෑ 1 − (1 − 𝑅𝑖 𝑡 )
𝑖=1

George Ebot Etta 51

 Linking System and Component Failures
• Example Stereo System

Consider the given stereo system, where failure is defined as “one cannot listen
to any music at all.” We consider the case where failed components are not
replaced and we are interested in the reliability of a new system over three
years. The component reliabilities for three years are as follows:

George Ebot Etta 52

 Linking System and Component Failures
𝑅1 (3) = 0.90, 𝑅2 (3) = 0.95, 𝑅3 (3) = 0.99, 𝑅4 (3) = 0.85, 𝑅5 (3) = 0.75
𝑅12 (3) = 1 − (1 − 𝑅1 3 )(1 − 𝑅2 3 )
𝑅45 (3) = 1 − (1 − 𝑅4 3 )(1 − 𝑅5 3 )
𝑅𝑆 3 = 1 − (1 − 𝑅1 3 )(1 − 𝑅2 3 ) 𝑅3 (3) 1 − (1 − 𝑅4 3 )(1 − 𝑅5 3 )

𝑅𝑆 3 = 0.995 𝑥 0.99 𝑥 0.9625 = 0.948

George Ebot Etta 53