09/06/2015

S. N. Njoroge 2

EEE 2503 Course Outline

• Concepts of Reliability:
RELIABILITY • Definitions,
ENGINEERING • Failure Frequency,
• Distributions Of Failures,
Mr. S. N. Njoroge
• Failure Probability.
• Failure Rate

S. N. Njoroge 3 S. N. Njoroge 4

Course Outline Course Outline

• Statistical Reliability and Estimation for • System Design And Development:
Components System Selection: • Design Reliability Analysis,
• Exponential Formula for Chance and Wear-Out Failure • DC-rating Techniques.
• Confidence Limits, • Tolerance Design.
• Cumulative and Conditional Probability, • Design Simplification.
• Poisson Distribution, • Component Reliability Characteristics.
• Series and Parallel Redundancy
• Bayes Theorem

S. N. Njoroge 5 S. N. Njoroge 6

Course Outline Course Outline

• Failure Modes and Effect Analysis; • System Maintenance:
• Fault Free Analysis. • Planned and Preventive Maintenance,
• Prototype and Sequential Tests. • Man-hour Calculations,
• Evaluation of Systems Reliability Using Monte Carlo • System Utilization Factor,
Technique • Maintainability Design and Modelling,
• Equations for System Availability and Dependency.
• Spares and Manuals

1
 09/06/2015

S. N. Njoroge 7 S. N. Njoroge 8

References Evaluation
• Ramakumar R. (1993), Reliability Engineering: Fundamentals • CATs - 10%
and Applications, Prentice Hall.
• Assignments - 5%
• Aggarwal K. (1993), Reliability Engineering, Kluker Academic
• Ushakov ed. Igor (1994), Handbook of Reliability Engineering,
• Experiments – 15%
Wiley. • Exam – 70%
• Grant Ireson W. and Clyde F. Coombs, Jr., Handbook of
Reliability Engineering and Management, McGraw-Hill.

S. N. Njoroge 9 S. N. Njoroge 10

Reliability
“The ability of an item to perform a required function under stated
conditions for a stated period of time”
• “quality over time”

• Monitored according to BS4778 standard

• Has both quantitative and qualitative aspects;

CONCEPTS ON RELIABILITY
• Measurements of reliability are necessary for customer
S. N. Njoroge requirements compliance.
• “Measuring reliability does not make a product reliable, only by
designing in reliability can a product achieve its reliability
targets”

S. N. Njoroge 11 S. N. Njoroge 12

Reliability Failure
• Associated with unexpected failures of products or services “any event or collection of events that causes the system to lose
• Understanding why these failures occur is key to improving
its functionability”
reliability
Functionability
• The inherent characteristic of a product related to its ability to
perform a specified function according to the specified
requirements under the specified operating conditions
• Transition from reliability to failure can be instantaneous (tyre
burst, transformer explosion, transistor blowing)
• Can also be gradual (cracks in insulation, bearing wears, cable
corrodes)
• Health monitoring can prevent failure

2
 09/06/2015

S. N. Njoroge 13 S. N. Njoroge 14

Reliability Importance of Reliability

Reasons for Product Failures • Safety
• Product is not fit for purpose • Competitiveness
• Design is inherently faulty • Profit margins
• Item may be overstressed in some way • Cost of repair and maintenance
• Wear-out • Delays further up supply chain
• Variation from a specific operating environment • Reputation
• Wrong specifications • Good will
• Misuse of the item

S. N. Njoroge 15 S. N. Njoroge 16

Objectives of Reliability Engineering Measuring Reliability

• To apply engineering knowledge to prevent or reduce the Reliability metrics can be classified into:
likelihood or frequency of failures; • Basic Reliability Measures
• To identify and correct the causes of failure that do occur; • Mission Reliability Measures
• To determine ways of coping with failures that do occur; • Operational Reliability Measures
• To apply methods of estimating the likely reliability of new • Contractual Reliability Measures.
designs, and for analysing reliability data.

S. N. Njoroge 17 S. N. Njoroge 18

Basic Reliability Measures Mission Reliability Measures

• Predict the system's ability to operate without maintenance and • Predict the system's ability to complete mission.
logistic support.
These measures consider only those failures that cause mission
Measures include failure.
• Reliability Function Measures Include
• Failure Function • Mission Reliability,
• Maintenance Free Operating Period (MFOP)
• Failure Free Operating Period (FFOP)
• Hazard Function

3
 09/06/2015

S. N. Njoroge 19 S. N. Njoroge 20

Operational Reliability Measures Contractual Reliability Measures

• Predict the performance of the system when operated in a • Define, measure and evaluate the manufacturer's program.
planned environment • Contractual reliability is calculated by considering design and
• Includes the combined effect of design, quality, environment, manufacturing characteristics.
maintenance, support policy, etc • It is the inherent reliability characteristic of a product.

Measures include Measures include

• Mean Time Between Maintenance (MTBM) • Mean Time To Failure (MTTF),
• Mean Time Between Overhaul (MTBO) • Mean Time Between Failure (MTBF)
• Maintenance Free Operating Period (MFOP) • Failure Rate
• Mean Time Between Critical Failure (MTBCF)
• Mean Time Between Unscheduled Removal (MTBUR)

S. N. Njoroge 21 S. N. Njoroge 22

The Bath – Tub Curve The Bath – Tub Curve

• Representation of the reliability performance of components or
non-repaired items.
• Captures the reliability performance of a large sample of
homogenous items entering the field over their lifetime without
replacement
• Has three distinct shapes or periods

S. N. Njoroge 23 S. N. Njoroge 24

Basic Rules of Probability Basic Rules of Probability

• The probability of any event must lie between zero and one
• In order to find the probability that A or B or both occur, the
inclusive:
0 ≤ 𝑃(𝐴) ≤ 1 probability of A, the probability of B, and also the probability
that both occur must be known, thus:
• For any event A, the probability of the complementary event,
𝑃 𝐴 ∪ 𝐵 = 𝑃 𝐴 + 𝑃 𝐵 − 𝑃(𝐴 ∩ 𝐵)
written A' , is given by
• If A and B are mutually exclusive events, so that 𝑃 𝐴 ∩ 𝐵 = 0,
𝑃 𝐴′ = 1 − 𝑃(𝐴)
then
• The probability of an empty or impossible event, ∅, is zero.
𝑃 𝐴∪𝐵 =𝑃 𝐴 +𝑃 𝐵
𝑃 ∅ =0
• If n events form a partition of S, then their probabilities must
• If occurrence of an event A implies that an event B occurs, so
add up to one:
that the event class A is a subset of event class B, then the 𝑛
probability of A is less than or equal to the probability of B: 𝑃 𝐴1 + 𝑃 𝐴2 + ⋯ + 𝑃 𝐴𝑛 = 𝑃 𝐴𝑖 = 1
𝑃(𝐴) ≤ 𝑃(𝐵) 𝑖=1

4
 09/06/2015

S. N. Njoroge 25 S. N. Njoroge 26

To Read Distributions Of Failures

• Measures Of Central Tendency • Probability Density Function
• Measures Of Dispersion
• Cumulative Distribution Function (CDF)
• Probability Density Function (PDF)

• The area under the curve of the distribution is equal to 1

∞
𝑓 𝑥 𝑑𝑥 = 1
−∞

S. N. Njoroge 27 S. N. Njoroge 28

Probability Density Function Distributions Of Failures

• The probability of a value falling between any two values 𝑥1 • The Cumulative Distribution Function or CDF, 𝐹 𝑡
and 𝑥2 is the area bounded by this interval, • since we are discussing time, change 𝑥 to 𝑡, 𝑇𝑇𝐹
𝑥2 𝑡
𝑝 𝑥1 < 𝑥 < 𝑥2 = 𝑓 𝑥 𝑑𝑥 𝐹 𝑡 = 𝑓 𝑡 𝑑𝑡
𝑥1 −∞
• gives the probability that a measured value will fall between
− ∞ and 𝑡
• as 𝑥 tends to ∞, 𝐹 𝑡 tends to 1.

S. N. Njoroge 29 S. N. Njoroge 30

Cumulative Distribution Function Failure Function 𝐹 𝑡

• “Probability that an item will fail before or at the moment of
operating time 𝑡”
• 𝑡 can have units such as miles, number of landings, flying
hours, number of cycles, number of trips, etc)
• probability that the time-to-failure (TTF) random variable will
be less than or equal a particular value 𝑡

5
 09/06/2015

S. N. Njoroge 31 S. N. Njoroge 32

Failure Function Failure Function - Properties

𝐹 𝑡 = 𝑃 𝑓𝑎𝑖𝑙𝑢𝑟𝑒 𝑤𝑖𝑙𝑙 𝑜𝑐𝑐𝑢𝑟 𝑏𝑒𝑓𝑜𝑟𝑒 𝑜𝑟 𝑎𝑡 𝑡𝑖𝑚𝑒 𝑡 • Failure function is an increasing function. i.e.
𝑡
𝑡1 < 𝑡2, => 𝐹(𝑡1) < 𝐹(𝑡2 )
𝐹 𝑡 = 𝑃 𝑇𝑇𝐹 ≤ 𝑡 = 𝑓 𝑡 𝑑𝑡 • It is assumed that the failure function value at time 𝑡 = 0,
0 𝐹(0) = 0. (not valid for “dead on arrival parts”)
• 𝑓(𝑡) is the probability density function of the time-to-failure • For 𝑡 = ∞, 𝐹 ∞ = 1
random variable TTF.
• 𝐹(𝑡) is the probability that an individual item will fail by time 𝑡.
Common PDF’s
• 𝐹(𝑡) is the fraction of items that fail by time 𝑡.
• Exponential
• Weibull • 1 − 𝐹(𝑡) is the probability that an individual item will survive
• Normal
up to time 𝑡.
• Lognormal
• Poisson

S. N. Njoroge 33 S. N. Njoroge 34

Reliability Function 𝑅(𝑡) Reliability Function - Properties

• “the probability that the system will not fail during the stated
• Reliability is a decreasing function with time t. That is,
period of time, t, under stated operating conditions”
𝑅 𝑡 = 𝑃 𝑡𝑕𝑒 𝑠𝑦𝑠𝑡𝑒𝑚 𝑑𝑜𝑒𝑠𝑛′ 𝑡 𝑓𝑎𝑖𝑙 𝑑𝑢𝑟𝑖𝑛𝑔 0 , 𝑡 𝑓𝑜𝑟 𝑡1 < 𝑡2 ; 𝑅 𝑡1 > 𝑅(𝑡2 )
𝑅(𝑡) = 1 − 𝐹(𝑡) • It is usually assumed that 𝑅(0) = 1. As 𝑡 becomes larger and
• is valid only for new systems or those systems whose failures
larger 𝑅(𝑡) approaches zero, that is, 𝑅(∞) = 0.
are not age related in Mission Reliability • 𝑅(𝑡) is the probability that an individual item survives up to
time 𝑡.
• 𝑅(𝑡) is the fraction of items in a population that survive up to
time 𝑡.

S. N. Njoroge 35 S. N. Njoroge 36

Hazard Function 𝑕(𝑡) Hazard Function

• Conditional probability of failure in the interval
𝑡 − (𝑡 + 𝑑𝑡)
given that no failure has occurred by time 𝑡
𝑓 𝑡 𝑓 𝑡
𝑕 𝑡 = =1−
𝑅 𝑡 𝐹 𝑡
• Can be decreasing, constant or increasing

6
 09/06/2015

S. N. Njoroge 37 S. N. Njoroge 38

Distributions Of Failures Exponential Distribution

• Exponential Distribution • When an item is subject to failures that occur in random
• Weibull Distribution intervals and the expected number of failures is the same for
• Normal Distribution
long periods of time then the distribution of failures is said to fit
an exponential distribution.
• Lognormal Distribution
• The PDF, Failure and Reliability functions are;
• Poisson Distribution
𝑓 𝑡 = 𝜆𝑒 −𝜆𝑡
𝐹 𝑡 = 1 − 𝑒 −𝜆𝑡
𝑅 𝑡 = 𝑒 −𝜆𝑡
• The Hazard Function is
𝑓 𝑡 𝜆𝑒 −𝜆𝑡
𝑕 𝑡 = = −𝜆𝑡 = 𝜆
𝑅 𝑡 𝑒

S. N. Njoroge 39 S. N. Njoroge 40

Exponential Distribution PDF Exponential Distribution CDF

S. N. Njoroge 41 S. N. Njoroge 42

Exponential Distribution Weibull Distribution

• Takes account of a non-constant hazard function
• The hazard function is not a function of time but a constant • Has a Reliability Function
equal to 𝜆 in the exponential distribution 𝑡 𝛽
−
𝑅 𝑡 =𝑒 𝜂
• For repaired items,𝜆, is the failure rate
• 1/𝜆 is called the mean time between failures (MTBF), • 𝛽 is the shape parameter
sometimes denoted as 𝜃. • η is the scale parameter or characteristic life. (life at which 63.2%
of the population will have failed)
• 63.2% of items will have failed by time 𝑡 = 𝜃.
• When 𝛽 = 1, the hazard function is constant like an exponential
• The failure rate can be calculated as the total number of failures 1
distribution with 𝜂 =
divided by the total operating time. 𝜆
• When 𝛽 < 1, we get a decreasing hazard function and
• When 𝛽 > 1, we get a increasing hazard function (bath-tub curve)
• When 𝛽 > 3.5, the Weibull distribution is an approximation of the
Normal distribution

7
 09/06/2015

S. N. Njoroge 43 S. N. Njoroge 44

Weibull Distribution PDF Weibull Distribution CDF

S. N. Njoroge 45 S. N. Njoroge 46

Normal Distribution Normal Distribution PDF

• Has a Reliability Function
𝜇−𝑡
𝑅 𝑡 =Φ
𝜎
• 𝜇 is the mean
• 𝜎 is the standard deviation
• Probability values obtained from tables

S. N. Njoroge 47 S. N. Njoroge 48

Normal Distribution CDF Normal Distribution - Example

• The time to failure distribution of a computer memory chip
follows normal distribution with mean 9000 hours and standard
deviation 2000 hours. Find the reliability of this chip for a
mission of 8000 hours.
Solution
𝜇−𝑡
𝑅 𝑡 =Φ
𝜎
9000 − 8000
=Φ
2000
= Φ 0.5 = 0.6915

8
 09/06/2015

S. N. Njoroge 49 S. N. Njoroge 50

Weibull Distribution - Example TO READ

• The time to failure distribution of a steam turbo generator can • Lognormal Distribution
be represented using Weibull distribution with 𝜂 = 500 hours • Poisson Distribution
and 𝛽 = 2.1. Find the reliability of the generator for 600 hours
of operation
Solution
𝑡 𝛽
−
𝑅 𝑡 =𝑒 𝜂
600 2.1
−
𝑅 𝑡 = 𝑒 500
𝑅 𝑡 = 0.2307