Total Quality Management- 3.

Robust and Reliable Product Approach to Quality

Unit 3
Robust and Reliable Product Approach to Quality

Reliability
Reliability is the probability of a system or component to perform its intended function
under stated conditions for a specified period of time.
The definition based on four significant elements
(s) Probability
(ii) Intended Function
(iii Time
(iv) Operating Conditions.
• Probability
Consideration of variation makes reliability a probability. It is possible to identify the
frequency distribution of an item, which permits prediction of life of the item. e.g., the
probability of an item functioning is 0.85 for 60 hours indicates that only 85 times out of
100, item to be functioning for a period of 60 hours.
Intended Function
For an item to be reliable, it must perform a certain functions satisfactorily when called
to do, while considering the reliability of an item, the criteria of what is considered as
the required function have to be exactly spelled out in advance. Thus, criteria must be
established in all cases which clearly specify and define what is considered as intended
function.
• Time
Time is the most important factor in the assessment of reliability, since it represents a
measure of the period during which one can expect a certain degree of performance from
an item.
• Stated Conditions
The application arid operating circumstances under which an item is put to use is an
important component of reliability. As the operating conditions, change the reliability of an
item also changes. Operating conditions such as temperature, humidity, torque, and
corrosive atmosphere all have a definite effect on performance.
Thus, reliability can be stated as follows
“The reliability of a 60 watts incandescent bulb has been estimated to be 0.95 for
1200 hrs. providing 20 candles output under 180-230 volts and at normal
environmental condition.

FAILURE
Failure of an item represents unreliability. Thus, to compute the reliability of an item,
it is necessary to understand the concept of failure. A deviation in the properties of an item
from the prescribed conditions is considered as fault. A state of the fault is denoted as
“Failure”.
An item is considered to have failed under one of the following conditions
1. When it becomes completely inoperable.
2. Whet it is still operable, but no longer able to perform a required function,
3. When a serious deterioration makes the item unsafe for its continued use.

Causes of Failures
There are many specific causes of failures of components and systems. Due to the
complexity of the system, some are known and some are unknown. Some of the causes of
failures are
• Deficiencies in design.
• Improper selection of process and manufacturing technique.
• Lack of knowledge and experience.
• Errors of assembly.
• Improper service conditions.
• Inadequate maintenance.
• Variation in environmental and operating conditions.
• Human errors.

Nature of Failures

1 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
An item may fail in many ways. An understanding of these failures help in taking
appropriate corrective measures for achieving better reliability. The different modes of
failure are
1. Catastrophic Failures
In this case, a normally operating item suddenly becomes inoperative. Example blowing
of a fuse or electric bulb.
2. Degradation (Creeping) Failures
These failures occur gradually because of change in some parameter with time.
Example change in resistance will affect the performance of a resistor.
3. Independent Failures
These are the failures, which occur independently and does not depend on failure of
the other.
4. Secondary Failures
A secondary failure occurs as a result of some primary failure.
5. Failure due to Improper Handling and Misuse
e.g., overloading (stressing beyond the capacity).

Bath Tub Curve (Phases of Failures)

The bathtub shape is ‘characteristic of the failure rate curve of many well designed
products and components including the human body.
Analysis of failure data has shown that failures in general are of three types,
i) Early failures due to defects in poor design, manufacture, assembly and quality
control. These gradually decreases with life of product.
ii) Random failures due to random reason such as environmental changes, vibrations,
wrong use or maintenance etc. These are constant throughout life of product
iii) Wear out
failures due to Erosion
and corrosion of parts.
These gradually
increases with life of
product.
Total failures are
sum of failures due to
these three reasons.

When a large
number of units are put into operation, it is likely that there is large number of failures
initially. These failures are called initial Failures or Infant Mortality. After the initial
failures, for a long period of time of operation fewer failures are reported but it is difficult
to determine their cause. The failures during this period are often called random failures
or catastrophic failures. This is the period of normal operation. As the time passes, the
units get worn out due to wear and tear and begin to deteriorate. Here in this period, the
failures are due to wear and tear and due to ageing. This region is called be wear out region.
These three phases of failures are represented in following fig.

Infancy / Green / Debugging / Burn-in-period:

Many components fail very soon after they are put into service. These are called infant
mortality failures. In this period failure rate is high due to defects in poor design,

2 DAC Notes 2016

 Total Quality Management- 3. Robust and Reliable Product Approach to Quality
manufacture, assembly and quality control and negligible due to wear out. Good system
vendors will perform an operation called "burn in" where they put in operation and test a
system for several days to try to weed out these types of problems so the customer doesn't
see them. Warranty is provided to compensate these early failures.
Methods to reduce failures during infant mortality period are: Burn-in testing or
debugging testing, Environment Stress Screening, Quality control and Acceptance testing
and accelerated testing.
Chance failure / Normal Operating Life:
If a component passes its infancy and after restored enter into normal useful life. It will
generally tend to remain trouble-free over its operating lifetime. The failure rate during this
period is typically quite low and is constant because failures are mainly due to random
reasons only. These are caused due to sudden stresses beyond strength of material, the
inherent limitation of design, accidents caused by usage or poor maintenance and the
environment random load. This phase, in which the failure rate is constant, typically
represents the useful life of the product.
To reduce failures in useful life,
 good control on operation and maintenance procedures can be applied
 The basic reduction of failure rate requires a basic redesign.
 Using redundancy to reduce failures during ‘useful life’.
Wear out / Ageing:
The wear-out period is characterized by a rapid increasing failure rate with time. After
a component reaches a certain age, it enters the period where it begins to wear out, and
failures start to increase. The period where failures start to increase is called the wear out
phase of component life. In this region, the failure rate will increase firstly, and then
decrease to zero for the obsolescence of components.
To reduce the failures in this region it is necessary to bring preventive replacement of
the dying components into effects and all failed components will be replaced by identical
components without disturbing the system.

MEASURES OF RELIABILITY
1. Failure Rate
Failure rate is expressed in terms of failures per unit time i.e. as failures per hour, or
failures per 100 or 1000 hours. Failure rate is the ratio of number of failures (f) during a
specified S interval to the total test time of items undergoing test.
2. Mean Time Between Failures (MTBF)
MTBF is referred to as the average time of satisfactory operation of the system. Larger
MTBF, higher is the reliability of the system. It is applicable to repairable systems and
expressed in hours, e.g. If an item fails 8 times over a period of 40,000 hours of operation
MTBF would be 500 hrs.
3. Mean Time To Failure (MTTF)
This is applicable to non-repairable systems. The mean time to failure is expressed as
the average time an item is expected to function before complete failure.

RELIABILITY IMPROVEMENT
A high degree of reliability is an absolute necessity for complex and modern systems to
be used for industrial, military and other scientific purposes. There are many ways by
which reliability of a component or system can be enhanced. These methods are discussed
below.
1. Design and Safety Factor
In order to design reliability into products, reasons for product failures must be
analyzed thoroughly. Generally, a product fails prematurely because of inadequate design
features, manufacturing and part defects, abnormal stresses induced, environmental
condition and human error. Several methods are available for the design engineers to
accomplish the requirements. Higher reliability could be achieved through mature design.
Higher safety factors are being used in those cases where there is a doubt regarding
the ability of the certain structure/component to withstand a particular load. Nowadays,
exhaustive testing methods are available to achieve required degree of reliability.
2. Parts and Material Selection
Designer has to choose between selecting standard parts and manufactured specialized
parts with higher reliability and greater tolerances. The tradeoff is usually in cost but ease
of parts availability, ease of repair, energy requirements, weight and size may also be

3 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
considerations. The historical databases can assist in determining relative reliabilities
among competing parts.
Knowledge of material properties and the external stresses the system will experience
is important. The material properties of materials such as metals, polymers, ceramics and
composites include tensile strength, hardness, impact strength, fatigue life and creep.
3. Redundancy
When it is not possible either to manufacture a highly reliable component or the cost
associated with such manufacturing is too high, the system reliability can be improved by
the techniques of introducing redundancy. This involves the creation of additional parallel
paths in the system.
Generally, there are two types of redundancies - parallel and standby to improve system
reliability.
In a system of complex nature, redundancy can be applied at various levels.
The various approaches for introducing redundancy in the system are
(a) To provide a duplicate or an additional path for the entire system itself. This is
known as system or unit redundancy.
(b) To provide redundant path for each component individually which is called
component redundancy.

It is easier to introduce redundancy at unit level rather than at component level. However,
component level redundancy provides higher reliability than unit redundancy.
(c) Weak component should be identified and strengthened by reliability.
(d) Use a combination of the above methods depending upon the configuration called
mixed redundancy.
Parallel redundancy also referred to as “hot redundancy” is generally resorted to when
the individual unit doesn’t have the required reliability. If the unit is highly critical or
strategic, the more than one unit can be put in parallel redundancy to compensate for
unreliability. A compromise is to be struck between the improvement in reliability and the
cost of each additional unit.
Stand by Redundancy
In stand by redundancy the units are duplicated but one or more units more remain
idle (called secondary unit) until the primary unit fails. In standby redundancy the failed
equipment or unit is replaced manually or automatically by its equivalent and in such
cases the reliability of the operator or sensing and switching mechanism is to be taken into
consideration.

4. Marginal Testing
Marginal testing is prescribed by the designer as a method of predicting probability of
failure due to degradation. Marginal testing involves periodic testing on a programmed
basis. This helps to isolate degraded parts or components and are replaced before the
actual failure occurs.
5. Derating
Derating consists of using a component under stress significantly below its rated value.
This is proved to be more beneficial when applied to electronic components, in which case

4 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
the designed voltage or current strength of the part is well above the normal operating
level.
6. Quality Control and Z-D Program
7. Maintainability
Maintainability is used to provide high effective reliability. If parts are readily
interchangeable and replaceable, failures can be repaired at a faster rate by replacing
defective parts with operating spares. This increases the available of the system.

RELIABILITY TESTING
The basic objective of reliability testing is to obtain information concerning failures in
order to express reliability quantitatively and to ascertain whether reliability and safety
goals are met and to improve product reliability.
The factors that are to be considered before any reliability test is being carried out are
• Objective of the test.
• Types of the test to be performed.
• Operating and environmental conditions.
• Number of units to be tested (Sample size).
• Duration of the test.
The various types of tests ate described below
Life testing-
Life testing is a process done under controlled conditions to determine how and when
a product will fail in its intended environment. It improves a product’s reliability and finds
the weakest and strongest points of a product
The purposes of Life testing are
– A test may be performed to see whether a certain configuration or item is feasible.
– A test may be used to determine which of several configurations is the optimum with
respect to performance, reliability, cost, modes of behavior under varying conditions, etc.
– A test maybe used to make more sensitive comparisons to further improve economy,
maintenance, use of standard parts, and so on.
– A test may be used to demonstrate whether the item is adequate to meet the
requirements of performance and reliability.
– A test can be used as thorough investigation of the latent capabilities of the item
under severer or more diverse conditions than those immediately anticipated.
By analyzing the product's response to such tests, engineers can make predictions
about the service life and maintenance intervals of a product.
Test consists of the following:
i. Pick several combinations of the relevant stresses (the stresses that accelerate the
failure mechanism under investigation). Each combination is a "stress cell". Note that
you are planning for only one mechanism of failure at a time.
ii. Make sure stress levels used are not too high - to the point where new failure
mechanisms that would never occur at use stress are introduced. Picking a maximum
allowable stress level requires experience and/or good engineering judgment.
iii. Put random samples of components in each stress cell and run the components in
each cell for fixed (but possibly different) lengths of time.
iv. Gather the failure data from each cell and use the data to fit an acceleration model
and a life distribution model and use these models to project reliability at use stress
conditions.

1. Burn-in Testing
Burn-in is a test in which a system or component is made to run for an extended period
of time to detect problems. Burn-in may be conducted to ensure that a device or system
functions properly before it leaves the manufacturing plant or may be part of a repair or
maintenance routine.
An end user may use a burn-in process to ensure that new equipment is functioning
properly, for example, running a new computer for several days before committing it to
serious use. If any of the components are defective the problem will usually show up within
this initial testing time frame.
This test is carried out to eliminate or reduce infant mortality failures by accumulating
initial equipment hours and resulting failure before user acceptance.
Primary objective of burn in testing is to increase the mean residual life of components
as a result of being survival this test. Items that have failed during the bum in may be
discarded and replaced or be repaired. The burn-in testing requires testing of all the units
produce for the designated time, so it increases the production lead time and costs.

5 DAC Notes 2016

 Total Quality Management- 3. Robust and Reliable Product Approach to Quality
When a faulty device or system is brought into a repair facility, the malfunction might
not occur until the hardware has been warmed or cooled to a certain temperature.
Humidity may also be a factor. Under these circumstances, a burn-in must be performed
and environmental parameters varied until the problem is found. The equipment is
powered-up and then left to run continuously for hours (or sometimes for days) until the
malfunction is observed and can be diagnosed.
2. Accelerated Life Testing
Accelerated life testing is the process of testing a product by subjecting it to conditions
(stress, strain, temperatures, voltage, vibration rate, pressure etc.) in excess of its normal
service parameters in an effort to uncover faults and potential modes of failure in a short
amount of time. Failure may be due to mechanical fatigue, corrosion, chemical reaction,
diffusion, migration, etc. These are the same causes of failure under normal stress; the
time scale is simply different.
When there is true acceleration, changing stress is equivalent to transforming the time
scale used to record when failures occur. The transformations commonly used are linear,
which means that time-to-fail at high stress just has to be multiplied by a constant (the
acceleration factor) to obtain the equivalent time-to-fail at use stress.
Two types of this test are
Usage rate acceleration- this test is used for products that only operate for small
amounts of time. In this test the time is increased to make failures happen quicker than if
it were used normally. Care should be taken that the increased usage rate does not
increase the other stresses on the product too much, so that the failure mechanism
produced are the same as those seen under normal usage rates.
Overstress acceleration – it is sometimes used as Backup to usage rate test. This
consists of running a product at a higher than normal level of some accelerating stress or
stresses to shorten product life or to degrade product performance factor. Typical
accelerating stresses are temperature, voltage, mechanical load, thermal cycling, humidity
and vibration. It is also used if usage rate is not appropriate means. This test is used for
products that are constantly used.

Evaluating System reliability-

Series model-
A series system is a configuration such that, if any one of the system components
fails, the entire system fails. Conceptually, a series system is one that is as weak as its
weakest link. A graphical description of a series system is shown in Figure below.

Let’s use a block for each system element, each block having its own reliability for a
given mission time T;
If the component reliabilities differ
𝑅𝑠 = 𝑅1 𝑥𝑅2 𝑥 … … … … … 𝑥𝑅𝑛
If all components are identical,
𝑅𝑠 = 𝑅 𝑛
"System Success" = "Success of every individual component"
The following three assumptions are needed.
1. Each component operates or fails independently of every other one, at least until the first
component failure occurs.
2. The system fails when the first component failure occurs.
3. Each of the n (possibly different) components in the system has a known life distribution
model R(t).

Parallel or redundant model

A parallel system is a configuration such that, as long as not all of the system
components fail, the entire system works. Conceptually, in a parallel configuration the
total system reliability is higher than the reliability of any single system component. A
graphical description of a parallel system of "n" components is shown in following Figure.

6 DAC Notes 2016

 Total Quality Management- 3. Robust and Reliable Product Approach to Quality

In a parallel model for which all the components have to fail before the system fails.
If there are n components, any (n−1) of them may be considered redundant to the
remaining one (even if the components are all different). When the system is turned on, all
the components operate until they fail. The system reaches failure at the time of the last
component failure.
If the component reliabilities are different,
𝑅𝑝 = 1 − Π(1 − 𝑅𝑖 ) = 1 − ((1 − 𝑅1 )𝑥(1 − 𝑅2 )𝑥 … … … 𝑥(1 − 𝑅𝑛 )) or
If all "n" components are identical,
𝑅𝑝 = 1 − (1 − 𝑅)𝑛
The assumptions for a parallel model are:
1. All components operate independently of one another, as far as reliability is concerned.
2. The system operates as long as at least one component is still operating. System failure
occurs at the time of the last component failure.

R out of N model
An "r out of n" system contains both the series system model and the parallel system
model as special cases. The system has n components that operate or fail independently
of one another and as long as at least r of these components (any r) survive, the system
survives. System failure occurs when the (n−r+ 1)th component failure occurs.
When r=n, the r out of n model reduces to the series model. When r=1, the r out
of n model becomes the parallel model.
We treat here the simple case where all the components are identical.
Formulas and assumptions for the r out of n model (identical components):
1. All components have the identical reliability function R(t).
2. All components operate independently of one another (as far as failure is concerned).
3. The system can survive any (n−r) of the components failing. The system fails at the
instant of the (n−r+1)th component failure.
4. System reliability is given by adding the probability of exactly r components surviving
to time t to the probability of exactly (r + 1) components surviving, and so on up to the
probability of all components surviving to time t. These are binomial probabilities (with
p=R(t)), so the system reliability is given by:

Combined systems
Some systems are made up of combinations of several series and parallel
configurations. The way to obtain system reliability in such cases is to break the total
system configuration down into homogeneous subsystems. Then, consider each of these
subsystems separately as a unit, and calculate their reliabilities. Finally, put these simple
units back (via series or parallel recombination) into a single system and obtain its
reliability.
For example, assume that we have a system composed of the combination, in series,
of the examples developed in the previous two sections. The first subsystem A, therefore,
consists of two identical components in parallel. The second subsystem B consists of a
redundant (parallel) configuration, composed of five identical components.

7 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality

𝑅𝐴 = 1 − ((1 − 𝑅1 )𝑥(1 − 𝑅2 ) and

𝑅𝐵 = 1 − ((1 − 𝑅3 )𝑥(1 − 𝑅4 )𝑥(1 − 𝑅5 )𝑥(1 − 𝑅6 )𝑥(1 − 𝑅7 )
Total reliability of system,
𝑅 = 𝑅𝐴 𝑥𝑅𝐵

Failure Modes and Effects Analysis (FMEA)

The Failure Modes and Effects Analysis (FMEA), also known as Failure Modes,
Effects, and Criticality Analysis (FMECA), is a systematic method by which potential
failures of a product or process design are identified, analyzed and documented.
Once identified, the effects of these failures on performance and safety are
recognized, and appropriate actions are taken to eliminate or minimize the effects of these
failures. An FMEA is a crucial reliability tool that helps avoid costs incurred from product
failure and liability.
Typically, the main elements of the FMEA are:
• The failure mode that describes the way in which a design fails to perform as
intended or according to specification;
• The effect or the impact on the customer resulting from the failure mode; and
• The cause(s) or means by which an element of the design resulted in a failure mode.
It is important to note that the relationship between and within failure modes, effects
and causes can be complex. For example, a single cause may have multiple effects or a
combination of causes could result in a single effect. To add further complexity, causes
can result from other causes, and effects can propagate other effects.

Procedure of the Design FMEA

The following ten steps provide a basic approach that can be followed in order to
conduct a basic FMEA. An example of a table lamp is used to help illustrate the process.
Attachment A provides a sample format for completing an FMEA.
Attachment A
FMEA Form
Failure Modes & Effect Analysis
Product: Completed by: Date Completed: Revision No.:
Item/P Part Part Failure Failure Severity Causes Probability Current Con RPN Recomm
art No. Descript Functio Mode Effects of Controls trol ended
ion n Occurrenc effec Actions
e tive
ness

Step 1 Step 2 Step 3 Step 4 Step 5 Step 6 Step 7 S8 S9 Step 10

8 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality

Step 1: Identify components and associated functions

The first step of an FMEA is to identify all of the components to be evaluated. This
may include all of the parts that constitute the product or, if the focus is only part of a
product, the parts that make up the applicable sub-assemblies. The function(s) of each
part within in the product are briefly described.
Example:
Part description Part Function
Light bulb Provides x ± y lux of illumination
Plug 2 wire electrical plug
Cord Conducts power from outlet to lamp
Step 2: Identify failure modes
The potential failure mode(s) for each part are identified. Failure modes can include
but are not limited to:
• complete failures • intermittent failures
• partial failures • failures over time
• incorrect operation • premature operation
• failure to cease functioning at allotted time
• failure to function at allotted time
It is important to consider that a part may have more than one mode of failure.
Example:
Part Description Failure Mode
Cord Short circuit
Open circuit
Insulation failure
Step 3: Identify effects of the failure modes
For each failure mode identified, the consequences or effects on product, property
and people are listed. These effects are best described as seen through the eyes of the
customer.
Example:
Failure Mode Failure Effects
Short No light/ Electrical fire/ Blown fuse
Insulation fail Shock/injury hazard
Step 4: Determine severity of the failure mode
The severity or criticality rating indicates how significant of an impact the effect is
on the customer. Severity can range from insignificant to risk of fatality. Depending on
the FMEA method employed, severity is usually given either a numeric rating or a coded
rating. The advantage of a numeric rating is the ability to be able to calculate the Risk
Priority Number (RPN) (see Step 9). Severity ratings can be customized as long as they are
well defined, documented and applied consistently. Attachment B provides examples of
severity ratings.
Attachment B
Severity Ratings
Example 1
Critical Safety hazard. Causes or can cause injury or death.
Major Requires immediate attention. System is non-operational.

Minor Requires attention in the near future or as soon as possible. System

performance is degraded but operation can continue.
Insignificant No immediate effect on system performance.
Example 2
1 None Effect will be undetected by customer or regarded as
insignificant.

9 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality

2 Very minor A few customers may notice effect and may be annoyed.

3 Minor Average customer will notice effect.

4 Very low Effect recognized by most customers.

5 Low Product is operable, however performance of comfort or

convenience items is reduced.

6 Moderate Products operable, however comfort or convenience items are

inoperable.

7 High Product is operable at reduced level of performance. High degree

of customer dissatisfaction.

8 Very high Loss of primary function renders product inoperable. Intolerable

effects apparent to customer. May violate non-safety related
governmental regulations. Repairs lengthy and costly.

9 Hazardous – Unsafe operation with warning before failure or non-conformance

with warning with government regulations. Risk of injury or fatality.

10 Hazardous – Unsafe operation without warning before failure or

without nonconformance with government regulations. Risk of injury or
warning fatality.
Example:
Failure Effects Severity
No light 8-Very high
Shock/injury hazard 10-Hazardous-no warning
Step 5: Identify cause(s) of the failure mode
For each mode of failure, causes are identified. These causes can be design
deficiencies that result in performance failures, or induce manufacturing errors.
Example:
Failure Mode Cause
Insulation failure Cord pinched
Step 6: Determine probability of occurrence
This step involves determining or estimating the probability that a given cause or
failure mode will occur. The probability of occurrence can be determined from field data
or history of previous products. If this information is not available, a subjective rating is
made based on the experience and knowledge of the cross-functional experts.
Two of the methods used for rating the probability of occurrence are a numeric
ranking and a relative probability of failure. Attachment C provides an example of a
numeric ranking. As with a numeric severity rating, a numeric probability of occurrence
rating can be used in calculating the RPN. If a relative scale is used, each failure mode is
judged against the other failure modes. High, moderate, low and unlikely are ratings that
can be used. As with severity ratings, probability of occurrence ratings can be customized
if they are well defined, documented and used consistently.
Attachment C
Probability of Occurrence Ratings1

1 Unlikely ≤ 1in 1.5 million (≤ 0.0001 %)

Low (few failures)
2 1 in 150, 000 (≤ 0.001 %)

3 1 in 15, 000 (≤ 0.01 %)

Moderate (occasional failures)
4 1 in 2,000 (0.05%)

5 1 in 400 (0.25%)

10 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality

6 1 in 80 (1.25%)
High (repeated failure)
7 1 in 20 (5%)

8 1 in 8 (12.5%)
Very high (relatively consistent
9 1 in 3 (33%)
failure)
10 ≥ 1 in 2 (50%)
Note: if a failure rate falls between two values, use the lower rate of occurrence. For
example, if failure is 1 in 5, use a rating of 8.
Example:
Cause Prob. Of Occurrence
Cord pinched Too Low (few failures)
Step 7: Identify controls
Identify the controls that are currently in place that either prevent or detect the
cause of the failure mode. Preventative controls either eliminate the cause or reduce the
rate of occurrence. Controls that detect the cause allow for corrective action while controls
that detect failure allow for interception of the product before it reaches subsequent
operations or the customer.
Example:
Cause Current controls
Cord pinched Review CSA standards
Warranty data from preceding products

Step 8: Determine effectiveness of current controls

The control effectiveness rating estimates how well the cause or failure mode can be
prevented or detected. If more than one control is used for a given cause or failure mode,
an effectiveness rating is given to the group of controls. Control effectiveness ratings can
be customized provided the guidelines as previously outlined for severity and occurrence
are followed. Attachment D provides example ratings.
Attachment D- Control Effectiveness Ratings
1 Excellent; control mechanisms are foolproof.
2 Very high; some question about effectiveness of control.
3 High; unlikely cause or failure will go undetected.
4 Moderately high
5 Moderate; control effective under certain conditions.
6 Low
7 Very low

8 Poor; control is insufficient and causes or failures extremely unlikely

to be prevented or detected.
9 Very poor
Ineffective; causes or failures almost certainly not prevented or
10 detected.
Example:
Current controls Control effectiveness
Review CSA standards 5-Moderate
Warranty data from preceding
products
Step 9: Calculate Risk Priority Number (RPN)
The RPN is an optional step that can be used to help priorities failure modes for
action. It is calculated for each failure mode by multiplying the numerical ratings of the
severity, probability of occurrence and the probability of detection (effectiveness of
detection controls) (RPN=S x O x D). In general, the failure modes that have the greatest
RPN receive priority for corrective action. The RPN should not firmly dictate priority as

11 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
some failure modes may warrant immediate action although their RPN may not rank
among the highest.
Step 10: Determine actions to reduce risk of failure mode
Taking action to reduce risk of failure is the most crucial aspect of an FMEA. The
FMEA should be reviewed to determine where corrective action should be taken, as well as
what action should be taken and when. Some failure modes will be identified for immediate
action while others will be scheduled with targeted completion dates. Conversely, some
failure modes may not receive any attention or be scheduled to be reassessed at a later
date.
Actions to resolve failures may take the form of design improvements, changes in
component selection, the inclusion of redundancy in the design, or incorporation design
for safety aspects. Regardless of the recommended action, all should be documented,
assigned and followed to completion.

Fault Tree Analysis (FTA)

FTA is a technique by which many events that interact to produce other events can be
related using simple logical relationships.
It follow a top-down structure and represent a graphical model of the pathways within a
system between basic events that can lead to a foreseeable loss event (or a failure) referred
to as the top event
There is a need to analyze all the possible failure mechanisms in complex systems
(e.g. nuclear power plants) also perform probabilistic analyses for the expected rate of
failures and estimate probabilities of events that are modeled as logical combinations or
logical outcomes of other random events.
It is one of the principal methods of probabilistic safety (or risk) analysis (PRA)
developed by Bell Telephone Laboratories in 1962 for the U.S. Air Force Minuteman
system, later adopted and extensively used by Boeing Company
Fault tree diagrams are used most often as a system-level risk assessment
technique. It can model the possible combinations of equipment failures, human errors,
and external conditions that can lead to a specific type of accident
In this diagram the contributory events and conditions are interconnected using
standard logic symbols (AND, OR, etc.), also referred to as gates
Events that must coexist to cause the top event are described using the AND relationship
Alternate events that can individually cause the top event are described using the OR
relationship
The occurrence of a top event may or may not lead to a serious or adverse
consequence.
The relative likelihood of a number of potential consequences will depend on the
conditions or subsequent events that follow potential consequences can be systematically
identified using an event tree.
The basic symbols used in FTA are grouped as events, gates, and transfer symbols
are given as below
No Symbol Name Description
Primary Event symbols
Basic Event – a basic initiating fault requiring no
1. Circle
further development
Conditioning Event – specific conditions or
2. Oval restrictions that apply to any logic gate (used with
INHIBIT gate)
Undeveloped Event – an event that is not developed
3. Diamond further because it is of insufficient consequence or
because information is unavailable

External Event – an event which is normally

4. House
expected to occur (not a fault event)

Intermediate Event Symbols

A fault event that occurs as a result of the logical

5. Rectangle
combination of other events

12 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
Gate Symbols

The union operation of events, i.e. the output event

6. OR Gate
occurs if (at least) one or more of the inputs occur

The intersection operation of events, i.e. the output

7. AND Gate
event occurs if and only if all the inputs occur

The output event occurs if the (single) input event

INHIBIT occurs in the presence of an enabling condition (i.e.
8.
Gate Conditioning Event (oval) drawn to the right of the
gate)
Transfer Symbols
Indicates that the tree is developed further
9. Triangle-in
someplace else (e.g. another page)
Indicates that this portion of the tree is a sub-tree
Triangle-
10. connected to the corresponding Triangle-In(appears
out
at the top of the tree)

General Procedure for Fault Tree Analysis

Step 1. Define the system of interest.
Specify and clearly define the
 Intended Functions
 Physical Boundaries (to avoid overlooking key elements of a system at interfaces and
penalizing a system by associating other equipment with the subject of the study)
 Analytical Boundaries (to limit the level of analysis resolution, to explicitly exclude
certain types of events and conditions, such as sabotage, from the analysis)
Step 2. Define the top event for the analysis.
Specify the problem of interest that the analysis will address (e.g. a specific quality
problem, shutdown, safety issue, etc.).
Step 3. Define the treetop structure.
Determine the events and conditions (i.e., intermediate events) that most directly
lead to the top event.
Step 4. Explore each branch in successive levels of detail.
Determine the events and conditions that most directly lead to each intermediate
event. Repeat the process at each successive level of the tree until the fault tree model is
complete.
Step 5. Solve the fault tree for the combinations of events contributing to the top
event.
Examine the fault tree model to identify all the possible combinations of events and
conditions that can cause the top event of interest. A combination of events and conditions
sufficient and necessary to cause the top event is called a minimal cut set.
Step 6. Identify important dependent failure potentials and adjust the model
appropriately (qualitative common cause failure analysis).
Study the fault tree model and the list of minimal cut sets to identify potentially
important dependencies among events. Dependencies are single occurrences that may
cause multiple events or conditions to occur at the same time.
Step 7. Perform quantitative analysis (if necessary).
Use statistical characterizations regarding the failure and repair of specific events
and conditions in the fault tree model to predict future performance for the system.
Step 8. Use the results in decision making.
Use results of the analysis to identify the most significant vulnerabilities in the
system and to make effective recommendations for reducing the risks associated with those
vulnerabilities.

13 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality

FTA can be used to:

i. understand the logic leading to the top event / undesired state.
ii. show compliance with the (input) system safety / reliability requirements.
iii. prioritize the contributors leading to the top event - Creating the Critical
Equipment/Parts/Events lists for different importance measures.
iv. monitor and control the safety performance of the complex system (e.g., is a
particular aircraft safe to fly when fuel valve x malfunctions? For how long is it allowed
to fly with the valve malfunction?).
v. minimize and optimize resources.
vi. assist in designing a system. The FTA can be used as a design tool that helps to
create (output / lower level) requirements.
vii. function as a diagnostic tool to identify and correct causes of the top event. It can
help with the creation of diagnostic manuals / processes.

Example- construct a fault tree for the simple electric motor circuit,

Solution:
Step 1. Define the system of interest.
For this particular problem we have,
 Intended Function – the motor is used for some (unknown) purpose
 Physical Boundaries – power supply
 Analytical Boundaries – include all contributors in the above diagram
 Initial Conditions – switch closed, motor on
Step 2. Define the top event.
We are interested in the event that the motor fails to operate. Therefore, the top event is
defined as “Motor fails to operate”
Step 3. Construct the fault tree, starting from the top, i.e., define the treetop
structure. Identify the main contributing events, including all events and scenarios that
may cause the top event.

14 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
Step 4. Explore each branch in successive levels of detail, following the rules of fault
tree construction.
Gate 1. One primary failure event is the failure of the motor itself (for example, due to a
wiring failure within the motor or loss of lubrication to the bearings). This event is a basic
event because no details of the motor are given; therefore, the event cannot be developed
further. The other possibility is the event that no current is supplied to the motor.
Gate 2. The event “No current in motor” is the result of other events and is therefore
developed further. The lack of current to the motor can result from a broken connection
in any of the other four components in the circuit, including the failure of the wire or power
supply (basic events), the switch being open, or failure of the fuse.
Gate 3. The open switch may be due to a basic failure of the switch, or the event that the
switch was opened erroneously. The erroneous opening of the switch is due to human
error, which could be developed further into more basic events (i.e. operator is
inexperienced, under stress, etc.). However, due to insufficient information, the event is
not explored further. This purposely undeveloped event is therefore denoted with the
diamond symbol.
Gate 4. The fuse failure event may be caused by fuse failure under normal conditions
(primary failure) or due to overload from the circuit.
Gate 5. The secondary fuse failure can occur if the fuse does not open every time an
overload is present in the circuit (because all conditions of an overload do not necessarily
result in sufficient overcurrent to open the fuse). This is why a conditional gate, denoted
by the hexagon, is used. The condition, i.e. “Fuse fails open” is placed in the connecting
oval, and the conditional gate is treated similarly to an AND gate in subsequent tree
analysis.
Gate 6. The overload in the circuit may be caused either by a short or a power surge, both
of which are primary (i.e. basic) events.

Fault Tree Evaluation

Solution:

15 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
Let T denote the top event
Let P denote primary events (circles)
Let G denote intermediate events T

(rectangles)
+
Let S denote undeveloped events
(diamonds)
Let C denote conditioning events (ovals) P1 G1
Therefore
T = motor fails to operate +
P1 = defect in motor
P2 = wire failure (open)
P3 = power supply failure G2 P2 G3 P3
P4 = switch fails open
P5 = fuse failure under normal + +

conditions (open)
P6 = wire failure (shorted) G4
P4 S1 P5
P7 = power failure (surge)
G1 = no current to motor x C1
G2 = fuse fails open
G3 = switch open G5
G4 = fuse failure due to overload
G5 = overload in circuit +
S1 = switch opened erroneously
C1 = fuse fails to open
P6 P7

Writing equations for each gate of the tree

T = P1 + G1
G1 = P2 + P3 + G2 + G3
G2 = P4 + S1
G3 = G4 + P5
G4 = C1 · G5
G5 = P6 + P7
Using the top-down approach we get by substitution
T = P1 + G1
= P1 + P2 + P3 + G2 + G3
= P1 + P2 + P3 + P4 + S1 + G3
= P1 + P2 + P3 + P4 + S1 + G4 + P5
= P1 + P2 + P3 + P4 + S1 + (C1 · G5) + P5
= P1 + P2 + P3 + P4 + S1 + P5 + C1 · (P6 + P7)
= P1 + P2 + P3 + P4 + S1 + P5 + (C1 · P6) + (C1 · P7)
T = P1 + P2 + P3 + P4 + S1 + P5 + (C1 · P6) + (C1 · P7)
= 0.01 + 0.01 + 0.01 + 0.01 + 0.001 + 0.01 + (0.50)(0.01) + (0.50)(0.01)
= 0.061
0.061

0.01 0.051

0.011 0.01 0.02 0.01

+ +

0.01 0.001 0.01 0.01

x 0.50

0.02

0.01 0.01

16 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality

Another example

17 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality

Taguchi’s quality Engineering

Taguchi’s quality philosophy
The old methods for quality assurance relied heavily upon inspecting products as
they rolled off the production line and rejecting those products that did not fall within a
certain acceptance range. However, Taguchi was quick to point out that no amount of
inspection can improve a product; quality must be designed into a product from the start.
Taguchi quality philosophy has three main concepts-
(1) Quality should be designed and built into the product and not inspected
into it. Taguchi believed that quality improvement starts at the very beginning i.e., during
the design stages of a product or process, and continues through the production phase.
He observed that poor quality cannot be improved by the process of inspection, screening
and salvaging. He proposed an "off line" strategy for developing quality improvement.
(2)Quality is best achieved by minimizing the deviation from the target
His second concept deals with actual method of effecting quality. He stated that
quality is directly related to deviation of a design parameter from the target value, not to
conformance to some fixed specifications. By specifying a target value for the critical
property and developing manufacturing processes to meet the target value with little
deviation, the life expectancy of the product may be improved considerably.
(3) The cost of quality should be measured as a function /deviation from the
standard
His third concept calls for measuring deviations from a given design parameter in
terms of overall life cycle costs of the product. These costs would include the cost of scrap,
rework, inspection, returns, warranty, service calls and/or product replacement. These
costs provide guidance regarding the major parameters to he controlled.

Taguchi’s concept of Quality Loss and Quality Loss Function

According to Taguchi, A products life cycle can be divided into two main parts. Life
before sale and life after sell.
All costs incurred prior to sale of the product are termed as manufacturing cost,
while all costs incurred after the sale are called as quality loss.
Taguchi defines Quality loss as “the loss imparted by the product to society from the
time the product is shipped.”
The loss is measured in monetary terms and includes all costs in excess of the cost
of a perfect product which may include various things such as a result of harmful effects
to society (e.g. pollution), failure to meet customer requirements (customer complaints],
failure to meet ideal performance, damage to company's reputation, added warranty costs,
cost of scrap, rework, inspection, returns, service calls and/or product replacement.
Quality loss results from customer dissatisfaction.
According to conventional concept quality loss occurs only when characteristics are
outside the upper and lower specification limits. Hence the conventional concept of quality
loss is a step function as shown in following figure.

0 𝑖𝑓 |𝑦−𝑚|≤𝛿
i.e. 𝐿(𝑦) = {𝐴𝑜 𝑖𝑓 |𝑦−𝑚|≤𝛿
Where L(y) is the loss associated with a particular parameter y,
Ao is repair cost.
m is target value (the nominal value of the parameter specification),
y is achieved value (the critical performance parameter value)
𝛿 is allowable variation
But according to Dr. Taguchi quality loss occurs when a product’s deviates from
target or nominal value. As deviation grows, then loss increases. Therefore Taguchi’s has
given a U-shaped loss Function Curve as shown below.

18 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality

Taguchi argues that product performance begins to gradually deteriorate as the

design parameter deviates from the target value. Before product goes to manufacturing
tolerances are set. Overall quality loss then increases by the square of deviation from the
target value, i.e. by the quadratic formula.
Taguchi’s loss function can be expressed in terms of the quadratic relationship:
L(y) = k (y - m)2 = 𝛿 2 (Eq. 2)
Where, k is a constant that depends on the cost at the specification limits (can be
determined conservatively by dividing the cost of scrap in Rs., by the square of the lower
or higher tolerance values).

Variations of the Quadratic Loss Function

1. Nominal the best type characteristic
Whenever the quality characteristic y has a finite target value, usually nonzero, and
the quality loss is symmetric on either side of the target. Their ideal value is equal to
nominal value, and as their value increases or decreases from nominal, the performance
becomes progressively worse. Such characteristics are
called smaller-the-better type quality
characteristics. The color density of a television set and
the output voltage of a power supply circuit are
examples of the nominal-the-best type quality
characteristic.
The quality loss in such situations can be approximated by the following function,
L(y) = k (y - m)2 = 𝛿 2
2. Smaller-the-better type characteristic
Some characteristics, such as radiation leakage from a microwave oven, can never
take negative values. Also, their ideal value is equal to zero, and as their value increases,
the performance becomes progressively worse. Such characteristics are called smaller-the-
better type quality characteristics.
The response time of a computer, leakage
current in electronic circuits, and pollution front an
automobile are additional examples of this type of quality
characteristic.
The quality loss in such situations can be
approximated by the following function, which is
obtained from Equation (2) by substituting m = 0:
L(y) = k (y)2 = 𝛿 2
Note this is a one-sided loss function because y cannot take negative values.
3. Larger-the-better type characteristic
Some characteristics such as the bond strength of adhesive do not take negative
value and zero is the worst value and as their value become larger performance becomes
progressively better that is, the quality loss becomes progressively smaller. Their ideal
value is infinity and at that point the loss is zero. Such characteristics are called larger the
better type quality characteristics. It is clear that the
reciprocal of such a characteristic has the same
qualitative behavior as a smaller-the-better type
characteristic. Thus, we approximate the loss
function for a larger-the-better type characteristic by
substituting 1/y for y in Equation of smaller the better.
1
L(y) = k ( 2)
𝑦
4. Asymmetric loss function

19 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
In certain situations deviation of the quality characteristic in one direction is much
more harmful than in the other direction. In such cases,
one can use a different coefficient k for the two
directions. Thus, the quality loss would be
approximated by the following asymmetric loss
function:
𝑘1 (𝑦−𝑚)2 𝑖𝑓 𝑦>𝑚
𝐿(𝑦) = { 𝑘2 (𝑦−𝑚)2 𝑖𝑓 𝑦<𝑚

Taguchi Philosophy of Robust design

Taguchi quality engineering is related to robust design.
Robust design, is defined as a design that has minimum sensitivity to variations in
uncontrollable factors.
The idea behind robust design is to improve the quality of a product by minimizing
the effects of variation without eliminating the causes (since they are too difficult or too
expensive to control).The sources of variation during these stages are noise i.e.
environmental variables, product deterioration, and production variations. The counter
measures against all these sources of variation are possible only at the product design
stage. To achieve desirable product quality by design Dr. Taguchi recommends a three
stage process:
1. System design.
2. Parameter design.
3. Tolerance design.

System design
System design is the development of the prototype.
The system design stage is where new ideas, concepts and knowledge in the areas of
science and technology are utilized by the design team to determine the right combination
of materials, parts, processes and design factors that will satisfy functional and economical
specifications.
System design is the conceptualization and synthesis of a product or process to be
used. It uses engineering and scientific knowledge to define initial setting of Product and
process parameters. Generally used techniques for this are QFD and Pugh’s concept
selection method.
To achieve an increase in quality at this level requires innovation. This is highly
creative step which needs experience and skill.
Usually only one technology is selected however for highly complex product two or
three are selected. Each is developed separately.
System design plays very important role in robust design and cost reduction.

Parameter design.
Parameter design is related to finding the appropriate design factor levels to make
the system less sensitive to variations in uncontrollable noise factors, i.e., to make the system
robust.
Parameter design is to determine best settings for control factors so system become
insensitive to uncontrollable factors (noise) and that don’t affect manufacturing cost and
which minimize quality loss. In this way the product performs better, reducing the loss to
the customer.
In parameter design the system variables are experimentally analyzed to determine
how the product or process reacts to uncontrollable “noise” in the system. Parameter
design is the main thrust of Taguchi’s approach.
In parameter design, there are two types of factors that affect a product’s functional
characteristic: control factors and noise factors.
Control factors are those factors which can easily be controlled such as material
choice, cycle time, or mold temperature in an injection molding process.
Noise factors are factors that are difficult or impossible or too expensive to control.
There are three types of noise factors: outer noise, inner noise, and between product
noise. Examples of each type of noise factor and controllable factors in product and
process design are listed in following Table.
Product design Process design

20 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
Outer  Consumer’s usage conditions  Ambient Temperature
Noise  Low temperature  Humidity
 High temperature  Seasons
 Temperature change  Incoming material variation
 Shock  Operators
 Vibration  Voltage change
 Humidity  Batch to batch variation
Inner  Deterioration of parts  Machinery aging
Noise  Deterioration of material Oxidation Tool wear
(rust)  Deterioration
Between Piece to piece variation where they Process to process variation
Product are supposed to be the same, e.g., where they are supposed to be
Noise  Young’s modulus shear modulus the same, e.g., variations in
allowable stress feed rate
Noise factors are primarily responsible for causing a product’s performance to
deviate from its target value.
Hence, parameter design seeks to identify settings of the control factors which make
the product insensitive to variations in the noise factors, i.e.,make the product more
robust, without actually eliminating the causes of variation.
Steps of parameter design are as follows,
1. Selection no. of the factors (both control and noise factors) to be evaluated.
2. Selection of the number of levels of each factor.
3. Selection of the appropriate orthogonal arrays.
4. Construct the “inner array” consists of the OA that contains the control factor
settings,
5. Construct the “outer array” consists of the OA that contains the noise factors and
their settings which are under investigation.
6. Then combine the “inner array” and “outer array” to form the “product array” or
“complete parameter design layout.”
7. Conduct the experiments as per product array to systematically test various
combinations of the control factor settings over all combinations of noise factors after
which the mean response and standard deviation is calculated for each run.
𝑛 𝑛
1 (𝑦𝑖 − 𝑦̅)2
𝑚𝑒𝑎𝑛 = 𝑦̅ = ∑ 𝑦𝑖 𝑎𝑛𝑑 𝑆𝑡𝑎𝑛𝑑𝑒𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 = 𝑆 = √∑
𝑛 𝑛−1
𝑖=1 𝑖=1

8. The preferred parameter settings are then determined through analysis of the
“signal-to-noise” (SN) ratio where factor levels that maximize the appropriate SN ratio are
optimal.

Orthogonal Array (OA)

Experimental design usually involves attempting to optimize a process which can
involve several factors (e.g., time, temperature, chemical composition etc.) at several levels
(e.g., two possible times, three possible temperatures, four possible compositions, etc.).
Orthogonal arrays (OA) represent a simplified method of putting together an
experiment. The original development of the concept was by Sir R.A. Fischer of England in
1930s.
Orthogonal Arrays (OAs), are employed in Taguchi’s approach to systematically vary
and test the different levels of each of the control factors.
An orthogonal arrays is a table consisting of a no. of rows and columns. An
orthogonal array is designated as L‘n’, e.g.L4, L8, L9 etc. In the designation, ‘n’ represents no
of the number of rows in OA.

21 DAC Notes 2016

 Total Quality Management- 3. Robust and Reliable Product Approach to Quality

Across the top of the orthogonal array consists of the maximum number of control
or noise factors that can be used. The columns in the OA indicate the factor and its
corresponding levels, and each row in the OA constitutes an experimental run which is
performed at the given factor levels or settings. Each row thus, represents a trial condition
with factor levels indicated by the numbers in the row.
For instance, in L4 OA of above Fig. experimental run no. 3 has Factor A at Level 2,
Factor B at Level 1, and Factor 3 at Level 2.
It is up to the experimental designer to establish the appropriate factor levels for
each control factor; typically either 2 or 3 levels are chosen for each factor.
Steps to determine suitable O.A.
1. Determine number of factors and levels
2. Determine degree of freedom
3. Select an orthogonal array

Degrees of freedom
Degrees of freedom decides the minimum number of treatment conditions.
Df = Number of factors x (number of levels - 1) + Number of interactions x (number of
levels-1) x (number of levels-1) + 1
Example – for two level factor, A, B, C, D and two suspected interactions, BC and CD,
Determine degree of freedom what is answer if factors are three level.
Df= 4(2-1) + 2(2-1)(2-1)+1=7
Df= 4(3-1) + 2(3-1)(3-1)+1=17
Maximum no. of degrees of freedom, Df = l f
Where l= number of levels and f= number of factors
The number of treatment conditions equal to number of rows in O.A. and must be
equal to or greater than Df.
e.g. for Df = 13, OA is L16.

Tolerance design.
Tolerance design occurs when the tolerances for the products or process are
established to minimize the sum of the manufacturing and lifetime costs of the product or
process.
In the tolerance design stage, tolerances of factors that have the largest influence on
variation are adjusted only if after the parameter design stage, the target values of quality
have not yet been achieved.
Tolerance design is a step used to fine tune the results of parameter design. When
parameter design is not sufficient for reducing the output variation, the last phase is
tolerance design. Narrower tolerance ranges must be specified for those design factors
whose variation imparts a large negative influence on the output variation. To meet these
tighter specifications, better and more expensive components and processes, identifying
the need for better materials, buying newer equipment, spending more money for
inspection etc. are usually needed. Because of this, tolerance design increases production
and operations costs.

Signal to Noise Ratio

Signal-to-noise ratio (often abbreviated SNR or S/N) is a measure used in science
and engineering that compares the level of a desired signal to the level of background noise.
It is defined as the ratio of signal power to the noise power, often expressed in
decibels.
There an many different types of S/N ratios. The important S/N ratios are
1. Nominal the best

22 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
2. Smaller the better
3. Larger the better
4. Dynamic
Computation of S/N Ratios
1. Nominal the best
𝑦̅ 2
S/N Ratio = -10 log (𝑆 2 )
2. Smaller the better
1
S/N Ratio = -10 log (𝑛 ∑𝑛𝑖=1 𝑦𝑖 2 )
3. Larger the better
1 1
S/N Ratio = -10 log (𝑛 ∑𝑛𝑖=1 𝑦 2 )
𝑖
Once all of the SN ratios have been computed for each run of an experiment, Taguchi
advocates a graphical approach to analyze the data. In the graphical approach, the SN
ratios and average responses are plotted for each factor against each of its levels. The
graphs are then examined to “pick the winner,” i.e., pick the factor level which (1) best
maximize SN and (2) bring the mean on target (or maximize or minimize the mean, as the
case may be). Using this information, the control factors can also be grouped as follows.
1. Factors that affect both the variation and the average performance of the
product.
2. Factors that affect the variation only.
3. Factors that affect the average only.
4. Factors that do not affect either the variance or the average.
Factors in the first and second groups can be utilized to reduce the variations in the
system, making it more robust. Factors in the third group are then used to adjust the
average to the target value. Lastly, factors in the fourth group are set to the most
economical level.
Finally, confirmation tests should be run at the “optimal” product settings to verify
that the predicted performance is actually realized.

A demonstration of Taguchi’s approach to parameter design serves as our case study

in the next section.
Case Study: Parameter Design of an Elastometric Connector
The following case study is taken from “The Taguchi Approach to Parameter Design,”
by D. M. Byrne and S. Taguchi, Quality Progress, Dec. 1987, pp. 19-26. The case uses
Taguchi’s parameter design approach to integrate product and process design decisions
for elastometric connector used in an automotive engine application.
The Problem
The experiment that is being conducted seeks to determine a method to assemble
an elastometric connector to a nylon tube while delivering the requisite pull-off
performance suitable for an automotive engineering application. The primary design
objective is to maximize the pull-off force while secondary considerations are made to
minimize assembly effort and reduce the cost of the connector and assembly.
Four control factors and three noise factors have been identified for the connector
and tube assembly. The control factors consist of the (A) interference, (B) connector wall
thickness, ( C) insertion depth, and (D) percent adhesive in connector pre-dip; a sketch of
the control factors for the connector and tube is given in Figure 3.

23 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality

Figure Control Factors for Connector and Nylon Tube Experiment

The noise factors in the experiment are (E) conditioning time, (F) conditioning
temperature, and (G) conditioning relative humidity. Each control factor is to be tested at
three levels while each noise factor is tested at two levels. The factors and levels of concern
for the experiment are listed in Table 3. In terms of product and process design, Factors
A and B represent product design parameters while Factors C-G represent process design
parameters. However, during routine operation the noise factors (E-G) are uncontrollable
are thus taken as “noise” which can adversely affect product performance. Fortunately,
these noise factors can able controlled for the purposes of this experiment. In this regard,
Taguchi’s parameter design approach can be used to help make product and process
design decisions to improve the robustness of a system.
Controllable Factors Levels
A. Interference Low Medium High
B. Wall thickness Thin Medium Thick
C. Insertion depth Shallow Medium Deep
D. Percent adhesive Low Medium High
Noise Factors Levels
E. Conditioning time 24 h 120 h
F. Conditioning 72ºF 150ºF
temperature
G. Conditioning relative 25% 75 %
humidity
Table 3 Factors and Levels for Connector and Tube
The Experiment
Following Taguchi’s method, two experimental designs are selected to vary (i) the
control factors and (ii) the noise factors. An L9 orthogonal array is selected for the
controllable factors while an L8 orthogonal array is chosen for the noise factors, see Table
4. The ones, twos, and threes in the L9 array in Table 4a correspond to the low, medium,
and high levels identified for each control factor and listed in Table 3. Similarly, the ones
and twos in the L8 in Table 4b corresponding the low and high levels for each of the noise
factors. Note that only the columns labeled E, F, and G in Table 4b are actually used in
the experiment. Since there are only three noise variables, the remaining columns in the
L8 array are used to estimate the interactions between certain noise factors (e.g., ExF
represents the interaction between conditioning time, E, and temperature, F). Finally, the
last column in the L8 array is used to estimate the variance in the experiment.

for the Control Factors for the Noise Factors

Table 4 Designs for the Control and Noise Factors
The total set of experiments that are performed is obtained by combining the L9 array
of control factors (the outer array) with the L8 array of noise factors (the inner array). The
total number of experiments is the product of the number of runs of each array, i.e., 9 x 8

24 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
or 72 experiments. For each experiment, the pull-off force is measured using the specified
settings for each control factor level and noise factor level. The average pull-off force for
each combination of the control factors A-D. Since the objective in the experiment is to
maximize the pull-off force, the signal-to-noise ratio for “Larger is Better” is also computed
for each set of runs. These results are summarized in Table 5 and discussed in the next
section.

Table 5 Pull-Off Force for Connector and Tube Parameter Design Experiment
Data Analysis
In this experiment, Taguchi’s graphical approach is used to plot the “marginal
means” of each level of each factor and “pick the winner” to determine the best setting for
each control factor. The average pull-off force and SN ratio for each level of each of the
control factors are plotted in Figure 4. These values are computed by averaging the mean
pull-off force or SNL for each factor for each level. For example, the average pull-off force
for the shallow setting (Level 1) of the insertion depth (Factor C) is obtained by averaging
Runs 1, 6, and 8 in Table 5, i.e., (17.525 + 19.225 + 18.838)/3 = 18.4 . The same procedure
is employed to compute the average SNL for each level of each factor and the remaining
pull-off force averages.

(a) Effect on SNL (b) Effect on Average Pull-off Force

Figure 4 Control Factor Effects
Figure 4 reveals that of the product design factors, interference (A) and wall
thickness (B), the interference has a larger impact on SNL and the average pull-off force.
The medium level for A (Amedium) is clearly the best choice for maximizing SNL and the
average pull-off force. As for the wall thickness, B, levels Bmedium and Bhigh are slightly better
than Blow; however, Bmedium is preferred to Bhigh in order to maximize the average pull-off
force, see Figure 4b.
For the process design parameters, insertion depth (C) and percent adhesive (D), the
insertion depth has a much larger impact on SNL and the average pull-off force than does
the percent adhesive. From Figure 4a, Cmedium and Cdeep yield nearly equal SNL, but Cdeep
should be selected to maximize the average pull-off force. What is the implication of two
levels of SNL being nearly equal? The answer is that the average pull-off force increases
when Cdeep is chosen, and SNL increases beyond its value when Cmedium is chosen. However,
the variance in the pulloff force also increases when Cdeep is chosen and SNL decreases.
The end result is that Cmediumand Cdeep yield nearly equal SNL. Notice that the SNL values
are nearly the same for Dmedium and Dhigh as well. It appears that level Dlow should be
selected for percent adhesive since it yields a slightly higher SNL and average pull-off force
compared to Dmedium and Dhigh.

25 DAC Notes 2016

Total Quality Management- 3. Robust and Reliable Product Approach to Quality
Hence, the best settings to maximize SNL are Amedium, Cdeep, Bmedium, and Dlow based
on the experimental results for maximizing pull-off force. In the actual study, further
analysis of the data revealed that (1) the variance in the experiment was not constant and
depended on the specific levels of each control factor, and (2) there were several
interactions between some of the control factors and noise factors (Byrne and Taguchi,
1987). A summary of the results is given in the next section for the full experiment.
Summary of Results
While it was easy to “pick the winner” from the effects plots in Figure 4, these factor
settings are only best for maximizing the pull-off force. In the complete set of experimental
results which also includes assembly effort and cost ratings, the choice of factor levels is
not as easy since many objectives are conflicting. A summary of the results is given in
Table 6 where insignificant effects are noted with a dash. Based on the full set of data, the
best levels for each factor are: Amedium, Blow, Cmedium, and Dlow. Only two of these choices
are the same as those obtained by examining Figure 4 since the most desirable settings for
reducing cost rating and assembly effort are not the same as those required to maximize
pull-off force. The final choice of factor level requires a trade-off between the multiple
competing objectives which is often the case in product and process design.
Factor Level Assembl Pull- Cost Performanc Rating Selecte
y off rating e Pull- d
effort force assembly off level
Interference 1 Low 8.1 18.7 Least Best Worst
(A) 2 8.3 20.7 – – Best X
Medium
3 High 8.7 19.8 Most Worst –
Wall 1 Thin – – Least – – X
Thickness (B) 2 – – – – –
Medium
3 Thick – – Most – –
Insertion 1 Shallow 7.7 18.4 Least Best Worst
Depth (C) 2 8.3 20.3 – – – X
Medium
3 Deep 9.1 20.6 Most Worst Best
Percentage 1 Low 8.3 20.5 Least Best Best X
adhesive(D) 2 8.4 19.5 -- Worst --
Medium
3High 8.4 19.2 Most Worst Worst
Table 6 Summary of Parameter Design Results for Elastometric Connector

26 DAC Notes 2016