Chapter 27

Reliability Engineering
Chapter Outline
27.1 Functional Reliability 391 27.12 Probability Density Function 400
27.2 General Causes for Poor 27.13 Procedure of Establishing
Reliability 392 Reliability Based Product
27.3 Distinguishing Between Quality 401
Quality and Reliability 392 27.14 Reliability Prediction 402
27.4 What is RBM? 392 27.14.1 Ingredients for
27.5 Bath Tub Characteristics 393 Reliability
27.6 Basics of RBM 395 Prediction 402
27.7 Principles of Reliability 27.14.2 Purposes of
Engineering 395 Reliability
27.8 House of Reliability 396 Prediction 402
27.9 Types of Failures 397 27.15 Monte Carlo Simulation 403
27.10 Severity of Failures 397 27.16 Markov Analysis 404
27.11 Statistical Distribution 27.17 Conclusion 404
Curves of Failures 397 Further Reading 404

27.1 FUNCTIONAL RELIABILITY

Reliability engineering is an engineering discipline for applying scientific
know-how to a component, product, plant, or process in order to ensure that it
performs its intended function, without failure, for the required time duration in
a specified environment. It emphasizes dependability in the lifecycle manage-
ment of a product, which is the ability of a system or component to function
under stated conditions for a specified period of time. In other words, reliability
has two significant dimensions, the time and the stress. A product has to endure
for several years of its life and also perform its desired function, despite all the
threatening stresses applied to it, such as temperature, vibration, shock, voltage,
and other environmental factors.
In quality management, this principle is applied to a component, product,
plant, or process in order to assure that it performs its intended function, with-
out failure, for the required time duration in a specified environment. This is
called functional reliability and the application of these principles to achieve
high product life is called reliability engineering.

Total Quality Management: Key Concepts and Case Studies. http://dx.doi.org/10.1016/B978-0-12-811035-5.00027-1

Copyright 2017 BSP Books Pvt. Ltd. Published by Elsevier Inc. All rights reserved. 391
392 Total Quality Management: Key Concepts and Case Studies

Earlier, we said quality conformation and customer satisfaction are essential

for companies to survive in their business. Reliability assumes a major factor in
sustaining quality and we can say that, the only companies left in business will
be those that are able to control the reliability of their products. Increase in the
complexity of the product, as well as in the equipment, has led to an increasing
demand for higher reliability.
Reliability engineering is an engineering framework that enables the defini-
tion of a complete production regime and deals with the study of the ability of
the product to perform its required functions under stated conditions for a speci-
fied period of time. It characterizes measures and analyzes the failure and repair
of the systems to improve their use by increasing their design life, mitigating
defect risks, and reducing the likelihood of failures.

27.2 GENERAL CAUSES FOR POOR RELIABILITY

1. Increasing product complexity,
2. Overemphasis on the state of the art factor for the performance,
3. Too many features included in the design that would affect reliability,
4. More complex and severe environmental changes, field stresses, and
interactions,
5. Short-circuited development cycles in order to be the first in the market,
6. Rapid product obsolescence,
7. Rising customer expectations for guaranteed performance and endurance, and
8. Lack of financial incentives or penalties for reliability in performance.

27.3 DISTINGUISHING BETWEEN QUALITY

ANDRELIABILITY
Quality is:
Independent of time.

Patent failures are removed by quality control methods.

Lot dependent.

Reliability is:
Time-dependent.
Latent failures can be detected.

Numerical estimates like mean time between failures (MTBF), failure rates

are possible.
These numerical estimates help us to compare between two different designs

at the proposal stage itself.

27.4 WHAT IS RBM?

If the principles explained in Section27.1 are applied to the machinery and equip-
ment, we call it the probability of failure free operation of the system for a given
period of time, under specific conditions, and the ability of equipment to perform
 Reliability Engineering Chapter | 27393

a required function under stated conditions for a stated period of time, and in
other words, how often the breakdown occurs of the equipment or the failure of
the components as measured in time units, say, hours. In this context, reliability
is as significant to quality management as it is for maintenance management.
Since the quality of a product depends to a large extent on the reliability of the
equipment with respect to their breakdown-free performance, it is essential for us
to understand reliability with specific reference to the reliability based machine
maintenance (Fig.27.1). Reliability based maintenance; originally called by its
synonym reliability centered maintenance incorporates sound guidance for man-
agers who wish to attain high standards of maintenance at their operating plants.
The amount and type of maintenance which is applied depend strongly on:
the age of the machine or components
its replacement cost and
the cost and safety consequences of system failure

FIG.27.1 Factors for RBM.

27.5 BATH TUB CHARACTERISTICS

The failure characteristics of a majority of the equipment follow the pattern shown
in Fig.27.2, sometimes called a bath tub pattern, which has three distinct phases:
Phase A or the burning in period: The major contributing factor for this
failure is the poor component quality. When the equipment is given initial trials,
there might be many initial failures due to poor design, workmanship, assembly
errors, etc. Damaged components and poor joints or connections also contribute
to this failure. These are tested and replaced generally at the manufacturer's
premises to improve the reliability.

FIG.27.2 Failure rate of equipment.

394 Total Quality Management: Key Concepts and Case Studies

Phase B or the useful life period: Here the failure rate is low, but may oc-
cur unexpectedly and at random intervals. They are known as random failures
or normal failures. It is during this period, that all our availability reliability
analysis is based on. The major contributing factor is the stress to which the
equipment or products are subjected to and could be due to operating stresses,
poor maintenance, operator abuse, and accidents.
Phase C or the wear out period: Beyond the useful period, the wear rate is
the major contributing factor because of aging or wear of the components of the
system, and could be due to weak design, poor lubrication, wear, fatigue failure,
corrosion, and insulation breakdown. In short, Table27.1 illustrates the contrib-
uting factors for each phase.

TABLE27.1 Contributing Factors for Failures

Phase A Phase B Phase C
Period Burning in Useful life Wear out

Failure occurrence Trial Random Excessive

Major contributing Low quality Stress Wear

factor

Other contributing Weak design, Operating stresses, Weak design,

factors assembly errors, poor maintenance, wear, fatigue,
damaged operator abuse, corrosion
components, poor accidents
joints/connections

A statistical representation of the probability that a product or system can

have maintenance-free performance for a given number of operating hours is
given in Fig.27.3.
Curve A shows a case where the system is not subjected to severe condi-
tions of services and tend to breakdown at nearly constant intervals following
the last repair. The statistical variation of these intervals is given by a nor-
mally distributed curve with its mean corresponding to the specified free run
time (Ta).
Curve B shows an ease of a system having more moving parts than in case
A, A's failure of any of the moving parts would result in the failures of the whole
machine, the variations expected in the average free time are more than the case
A, with the free time expected itself being lower. Thus, the distribution curve for
this will be a slanted type of normal curve.
Curve C shows wide variations in the free times, in the case of systems ne-
cessitating intricate and careful setting up for efficient performance.
 Reliability Engineering Chapter | 27395

FIG.27.3 Maintenance-free performance curves.

27.6 BASICS OF RBM

SAE JA1011, Evaluation Criteria for RBM Processes sets out the mini-
mum criteria that any process should meet, starting with the seven questions
below:
1. What is the item supposed to do and what are its associated performance
standards?
2. In what ways can it fail to provide the required functions?
3. What are the events that cause each failure?
4. What happens when each failure occurs?
5. In what way does each failure matter?
6. What systematic task can be performed proactively to prevent, or to dimin-
ish to a satisfactory degree, the consequences of the failure?
7. What must be done if a suitable preventive task cannot be found?

27.7 PRINCIPLES OF RELIABILITY ENGINEERING

The EBME website (http://www.ebme.co.uk) lists the following principles
upon which reliability based management (RBM) is based:
1. RBM is function-oriented and seeks to preserve system or equipment
function.
2. It is group focused and is concerned with maintaining the overall function-
ality of a group of devices, rather than an individual device.
3. It uses failure statistics in an actuarial manner to look at the relationship
between operating age and the failures. However, RBM is not overly con-
cerned with simple failure rate; it seeks to know the probability of failure
at specific ages.
396 Total Quality Management: Key Concepts and Case Studies

4. Acknowledges design limitations, recognizing that changes in reliability

are the province of design, rather than maintenance. Maintenance can only
achieve and sustain the level provided for by design.
5. RBM is driven by safety and economics. Safety must be ensured at any
cost; thereafter, cost-effectiveness becomes the criterion.
6. It defines failure as any unsatisfactory condition, as either a loss of function
(operation ceases), or a loss of acceptable quality (operation continues).
7. It uses a logic tree to screen maintenance tasks to provide a consistent ap-
proach to the maintenance of all kinds of equipment.
8. Its tasks must address the failure mode and consider the failure mode
characteristics.
9. RBM must reduce the probability of failure and be cost-effective.
10. RBM tasks are interval (time- or cycle-)-based and condition-based.
Here run-to-failure, is a conscious decision and is acceptable for some
equipment.
11. It is dynamic and gathers data from the results achieved, which is fed back
to improve future maintenance. This feedback is an important part of the
proactive maintenance element of the RBM program.

27.8 HOUSE OF RELIABILITY

The strength provided by Reliability Engineering to an organization can be il-
lustrated in Fig.27.4.

FIG.27.4 House of reliability.

 Reliability Engineering Chapter | 27397

27.9 TYPES OF FAILURES

Failures can be grouped into the following three categories. Understanding
these categories is critical when assigning maintenance tasks.
Induced
Intermittent
Wear out
Induced failures are a result of an outside force causing the failure
mode, like a soft foot condition on an equipment train causing coupling
misalignment, eventually leading to an inboard bearing failure. Soft foot
condition implies improper contact between a machine casing and the base-
plate used to support it. In case of rotating machines like the motors, such
soft foot condition causes heavy vibrations leading to major breakdowns
and accidents.
It is important to understand that induced failure must be recognized and
analysis performed to determine the root cause, as explained further by the fail-
ure mode and effects analysis (FMEA) concept.
Intermittent failures can happen at any time at random, and the MTBF can-
not be predetermined, and the repair cannot be effectively planned and sched-
uled. A plant can best detect these failure modes through process monitoring
and predictive maintenance to some extent.
Wear-out failures have a known MTBF and they occur when the useful life
of a component is expended. These types of failure modes are often detectable
through process monitoring and predictive maintenance. However, time-based
refurbishment or preventive maintenance sometimes could prove to be an effec-
tive maintenance strategy.

27.10 SEVERITY OF FAILURES

DOD-STD-2101 defines the characteristics of a component and system
defects as:
Critical, if the failures will have adverse impact on safety,
Major, if the defective characteristics will degrade with age, and
Minor, if the defects do not a have significant impact on the performance
This characteristic of severity of failure is dealt in more detail in Table26.1
of Chapter26.

27.11 STATISTICAL DISTRIBUTION CURVES OF

FAILURES
While the previous chapter gives physical illustrations and computed the fail-
ure rate, etc., by simple arithmetic computations, this chapter briefs the several
distribution curves that the failures conform to. While the detailed statistical
398 Total Quality Management: Key Concepts and Case Studies

explanation is beyond the scope of this book, the basic explanation of their
concepts to the extent an engineer should know is discussed in this chapter.
(a) The normal distribution is a probability distribution that associates the nor-
mal random variable around central value, called the mean. This is gener-
ally applied to analysis of variations, also known as Anova as a special
abbreviation around a nominal fixed value, like the variations in the ma-
chining of a bar to say, 50mm diameter. It is also called a bell curve, since
it looks like a bell with a central peak as in Fig.27.5.

FIG.27.5 Normal distribution.

(b) The Poisson distribution is a curve expressing the probability of a given dis-
creet number of events occurring in a fixed interval of time, or certain sample
sizes like the number of defects found in each of the several samples picked
up from a lot. This figure can be 0, 1, or 2, etc., unlike the normal distribution
which the variations cluster around a central figure. Here, like the of normal
distribution, the variation factor is represented by (lambda). As =1, it is a
simple c-shaped curve, reaching a maximum when the occurrence is zero and
smoothing out to a normal curve when the occurrences are high (Fig.27.6).

FIG.27.6 Poisson distribution.

 Reliability Engineering Chapter | 27399

If X has a Poisson distribution with a mean of n (say 2) failures per year, then
the probability that no more than r (say 1) failures occur per year is given
by P(X=r)=X!
If n=2 and r=1, the n(X=1) would be 0.406 or 40.6% probability.
(c) The Weibull distribution (Fig.27.7) is similar to Poisson, but uses three
parameters.

FIG.27.7 Weibull distribution.

the shape parameter (), also known as the Weibull slope

the scale parameter ()
the location parameter ()
the most general expression of the Weibull pdf is given by the three-
parameter Weibull distribution expression, or:
b -1 b
b T -g T -g
f (T ) =
-
e
h h h

(T g )
or if t = , then
h
f (T ) = b ( t )
b 1
( )
exp t b

You notice that when =0, it is more or less similar to Poisson distribution,
but as increases, the curve assumes a normal shape and becomes steeper,
showing that the determination of failure would be more precise. The Weibull
distribution is very useful, not only for the failure analysis in determining the
equipment reliability, but also in survival analysis, in the insurance industry, in
industrial engineering to represent manufacturing and delivery times, and also
in weather forecasting (Fig.27.8).
400 Total Quality Management: Key Concepts and Case Studies

FIG.27.8 Probability of failure.

FIG.27.9 Probability of survival.

More statistical curves can be deduced and represented as follows, indicat-

ing the different conditions of failures.
(a) When a fraction of items are expected to fail by a time t that is the prob-
ability of failure f(t).
(b) When a fraction of items are expected to survive by a time t that is the prob-
ability of survival f(t) (Fig.27.9).

27.12 PROBABILITY DENSITY FUNCTION

The probability density function (PDF), or density of a continuous random vari-
able, is a function that describes the relative likelihood for this random vari-
able to take on a given value. This PDF is most commonly associated with
absolutely continuous univariate distributions and for the random variable to
fall within a particular region is given by the integral of this variables density
over the region. In the illustration given alongside, the probability density for
the median to fall within the limits of Q1 and Q2 is given by the darker shaded
area (Fig.27.10).
 Reliability Engineering Chapter | 27401

FIG.27.10 Probability density function. (Based on Wikipedia)

27.13 PROCEDURE OF ESTABLISHING RELIABILITY

BASED PRODUCT QUALITY
While the application of reliability in the design function is given in more detail
in Chapter32, a brief summary of the steps involved can be as follows.
1. Quantify reliability requirements as design goals or specifications.
2. Allocate and apportion the reliability requirements to specific system com-
ponents and parts.
3. Apply reliability design methods during the equipment design and develop-
ment. Perform reliability and maintainability analysis, such as block dia-
grams, stress-strength analysis, redundancy, etc.
4. Conduct FMEA and criticality analysis.
5. Participate in design reviews.
6. Establish test procedures and conduct reliability testing.
7. Perform reliability prediction and demonstration.
8. Develop a reliability plan.
402 Total Quality Management: Key Concepts and Case Studies

27.14 RELIABILITY PREDICTION

Reliability prediction, the process of forecasting the probability of success from
available data is one of the important techniques in knowing the reliability of an
equipment or system. It involves estimating the reliability (ie, performance of
the system over a period of time) based on the failure rate of the components.
Itthus helps in identifying weak areas in a design, and also in choosing the best
design from among alternate configurations.

27.14.1 Ingredients for Reliability Prediction

Reliability relationships
Reliability concepts
Constant failure rate
The Bathtub Failure Rate curve
System redundancy
Fault tolerance
Functional redundancy
Fault avoidance

27.14.2 Purposes of Reliability Prediction

1. Assuring the feasibility of reliability requirements (downtime, etc.) for the
design proposed
2. Comparing competing designs
3. Identifying potential reliability problems
4. Planning maintenance and logistic support strategies
5. Reliability predictions can be used to assess the effect of product
6. Reliability on the maintenance activity and on the quantity of spare
7. Units required for acceptable field performance of any particular system.
For example, predictions of the frequency of unit level maintenance can be
estimated
8. Estimating unit and system lifecycle costs
9. Provide necessary input to system level reliability models
10. Assist in deciding which product to purchase from a list of competing
products
11. Useful in setting standards for factory reliability tests and field performance
The failure rate of all the cards in the system are evaluated as per
QM115A Quality Manual on Guidelines to calculate theoretical reliability
failures for telecom equipment issued by Telecom QA circle, DOT, Issue
2, Jan. 1997.
In his address on Prevention of Problems on Reliability and Safety at
NIQR, Chennai, in January 2015, Professor Kazuyuki Suzuki of University
of Electro-Communications, Tokyo, emphasized that events that cannot be
 Reliability Engineering Chapter | 27403

redicted, cannot be prevented. But careful consideration of the following

p
would provide an inductive approach to understand the situation for more ac-
curate prediction.
Sharing of problem information beyond organization
Abstraction and generalization of individual problems
Implementation of PDCA cycle
Practical use of incident information

27.15 MONTE CARLO SIMULATION

The Monte Carlo method is a broad class of computational algorithms that re-
lies on repeated random sampling to obtain numerical results, letting us account
for risk in the quantitative analysis and decision-making. Just as we keep play-
ing and recording our results in a real casino situation, the simulation is run
many times over in order to calculate the probabilities. Hence, the name Monte
Simulation. It is specifically useful in systems with several degrees of free-
dom with uncertainty in the inputs, such as the calculation of risk in purchasing
shares.
In fact, this author nostalgically remembers his visit to Monte Carlo Casino
in Monaco in 1983, when he practically saw, what he was taught in 1968, how
the players in the roulette table kept on recording their and others bets and
results, and used them to decode their next bets.
When Monte Carlo simulations have been applied in space exploration
and oil exploration, their predictions of failures, cost overruns, and sched-
ule overruns are routinely better than human intuition or alternative soft
methods.
Because testing of statistical significance of any activity or parameter is
time-consuming, simulation techniques can effectively be adapted for known
probabilistic models to predict the outcome. Monte Carlo simulation is one such
technique. The steps involved are:
1. Define a domain of possible inputs.
2. Generate random numbers between 0 and 1 to represent the reliability of a
component.
3. Replace the reliability as a function of time into the model and calculate the
corresponding time to failure by deterministic computation.
4. Generate another random number and repeat the process until a sufficient
number of trials are made.
5. Summarize the results.
Monte Carlo simulation can be successfully applied, not only in engineer-
ing situations such as microelectronics engineering, geo metallurgy, aerospace
engineering, etc., but also in medical situations like biological systems, such as
proteins, membranes, images of cancer, etc.
404 Total Quality Management: Key Concepts and Case Studies

27.16 MARKOV ANALYSIS

Markov Analysis is an analytical method to determine the reliability and avail-
ability of a system whose components exhibit strong dependencies or con-
straints, similar to the tree analysis discussed elsewhere in this book.
This method proposed by Russian mathematician Andrei Andreyevich
Markov is used to forecast the value of a variable whose future value is indepen-
dent of its past history, and is a sure method for forecasting random variables.
Some typical constraints that can be considered for using Markov models are:
Components in cold or warm standby
Common maintenance personnel
Common spares win limited in-site stock

27.17CONCLUSION
In general, the concept of quality is related to control of lower-level product
specifications and manufacture, whereas the concept of reliability is related
to the systems engineering, manifested by the day-by-day operation for many
years. Quality is therefore related to Manufacturing, and Reliability is more re-
lated to the validation of systems or sub-systems inherent to design and lifecycle
solutions.

On the Lighter Side

Adjustmentality is a word coined by this author to
illustrate the philosophy to accept whatever is the
outcome of your actions, without getting unduly
worked up. This mentality of adjusting oneself to the
situation would reduce the blood pressure of a person,
but of course is in contradiction to most of the quality
management principles professed in this book!

FURTHER READING
[1] ReshetovD, IvanovA, FadeevV. Reliability of machines. Moscow: Mir Publishers; 1990.
[2] EbelingCE. Reliability and maintainability engineering. Noida: Tata McGraw Hill; 2000.
[3] MurthyMN. Excellence through quality & reliability. Chennai: Applied Statistical Centre;
1989.
[4] http://en.wikipedia.org/wiki/Reliability_engineering.
[5] www.iitb.ac.in.
[6] www.weibull.com/basics/reliability.htm.