Merged Presentation Choladeck

Quality Control Assurance &
Reliability
BITS Pilani Dr. Sudeep Kumar Pradhan, PhD.

Pilani Campus
Introduction to Quality Control
and Total Quality System
2
BITS Pilani, Pilani Campus
Evolution of Quality Control
• Egyptians demonstrated a commitment to quality in the

construction of pyramids
• Quality of Greek architecture of 5th century B.C was so
envied that it profoundly affected the subsequent
architectural constructions of Rome
• Middle ages and up to 1900: Operator Quality Control
– Production of goods and services were predominantly confined to
single person or a group
– Generally they are family owned business and hence responsibility
of maintaining quality was determined by individual who is
producing it
3
Evolution of Quality
Control – contd..
• Middle ages and up to 1900 – contd..
– Volume of production was limited
– Controlling the quality of the product was thus
embedded in the philosophy of the worker because
pride in workmanship was widespread
– Sculptures, Goldsmith etc.
• Early 1900 to about 1920: Foreman Quality
control
– Industrial revolution resulted in concept of mass
production
– Resulted in principle of specialization of labour and
hence individual not responsible for producing whole
product
4
Control – contd..
• Early 1900 to about 1920 – contd..
– Lacked in sense of accomplishment in their work, but
become skilled at a particular accomplishment
– But people are grouped together, where a supervisor
who directed that operation had the task of ensuring
that quality was achieved
• 1920 to 1940 : Inspection quality control phase
– Products and process were complicated and also the
volume increased
– Resulting in increase in number of workers under a
foreman and hence prevents him from keeping close
watch
5
Control – contd..
• 1920 to 1940 : Contd..
– Hence inspectors were designated to check quality of
the product after certain operations comparing it with
standards. If any problem, checked product is reworked
or discarded
– Foundation of statistical aspects of quality control were
developed
– W.A Shewhart developed control charts in 1924
– H.F. Dodge and H.G. Romig was working on acceptance
sampling plan and its application was started in late
1920s
6
Control – contd..
• 1940 to 1960 : Statistical quality control
– Production requirements increased during World war II
and hence 100% inspection was not feasible
– ASQC was formed in 1946 and developed sampling
inspection plans for military purposes MIL-STD-105A
– E. Deming visited Japan and lectured on importance of
statistical quality control in 1950
– J.M. Juran visited Japan in 1954 and impressed upon the
strategic role that management plays in achieving quality
program
– In 1959, Inspection and Quality control handbook was
updated with multi-level continuos sampling plan as well
as topics in life testing and reliability
7
Control – contd..
• 1960 to 1970 : Total quality control
– Gradual involvement of other departments and
management personnel in quality control process
– Attitude of quality is responsibility of inspection
department started changing
– Concept of ‘zero defect’ which emphasized productivity
through worker involvement emerged
– Use of Quality circles was started in Japan, which is based
on participative style of management
• 1970 to present : Total quality control
Organization wide
– Involves participation of everyone in the company from
the operator to supervisor to manager to CEO
8
Control – contd..
• 1970 to present : contd..
– A quality system was evolved which is
• Agreed on companywide operating work structure
• Documented integrating technical and managerial procedures
• Guiding the coordinated actions of men, machine and
information of the company
• To assure customer quality satisfaction and economical cost of
quality
– Expanded use of Ishikawa diagram
– G.Taguchi introduced the concept of quality
improvement through design of experiments
– With more advertisements on comparing competitors
brought quality into limelight
9
Control – contd..
• 1970 to present : contd..
– Top management saw the need for the marriage of quality
philosophy to production in all phases from determination
of customer needs, product design and customer service
– Training program in SQC methods for all workers
– With the wide use of computers, lot of quality control
software came and hence the emphasis on vendor quality
control, product quality audit were placed
– In this phase, customer will reign supreme as the
determiner of acceptable level of quality
– Industries need to adjust to this or lose market share
10
Quality - Definition
• Juran (1974): Fitness for use

• Crosby (1979): Conformance to requirements or specifications
• Garvin (1984)
– Divides the definition of quality into 5 categories
• Transcendent (always improving or uplifting)
• Product-based
• User-based
• Manufacturing-based
• Value-based
– 8 attributes can be used to define quality
• Performance Features Reliability
• Conformance Durability Serviceability
• Aesthetics Perceived Quality
• Mitra (2000)
– Is the fitness of the product or service for meeting or exceeding its
intended use as required by customer
11
How to define quality?
• Defined using Quality Characteristics

• Quality characteristics
– Defined as the elements that define the
intended quality level of the product or service
12
How to define quality? – contd..
• Diameter
• Length
• Pressure
• Wall Thickness
13
How to define quality? – contd.
• Taste, Smell
• Price, Way of serving

etc.
14
Place B Place A
• Time
• Reliability
15
• Honesty
• Courtesy
16
Quality characteristics
• Can be grouped under 4 categories

– Structural characteristics: Length, Viscosity, Strength
– Sensory characteristics: Taste, smell, etc.
– Time related characteristics: Warranty, reliability
– Ethical characteristics: Honesty, courtesy etc.
• Are of two types
– Variables
• Measurable and expressed on numerical scale,
• Eg., Diameter, Resistance etc.
– Attributes
• If a characteristic is said to be conforming or non conforming or
that cant be measured
– Variable can be attribute, but attribute cannot be variable
• It results in economical measurement and time saving
17
Some Terminologies
• Nonconformity • Nonconforming unit
– A quality – A unit that has one or
characteristic that more nonconformities
does not meet its such that the unit is
stipulated unable to meet the
specifications intended standards and
– Eg. Thickness of is unable to function as
plate: 5 +/- 0.1, but if required
it is 5.2, then – Eg. A plate having both
nonconformity thickness and length
failing to meet
specifications
18
Some Terminologies
• Defect • Defective
– Associated with quality – Severity of one or more
characteristic that does not defects in a product may
meet certain standards cause it to be unacceptable
– According to ANSI (The
American National Standard
Institute), a defect is a
departure of quality
characteristic from its
intended level or state that
occurs with a severity
sufficient to cause an
associated product or
service not to satisfy
intended normal or
reasonable foreseeable
usage requirements
19
Some Terminologies – contd..
• Standard / specification • Standard
– Refers to a precise statement that – A prescribed set of conditions
formalizes the requirements of or requirements of general or
the customer, it may relate to a
product, process or a service broad application, established
• Specification by authority or agreement, to
– A set of conditions and be satisfied by a material,
requirements of specific and product, process, procedure,
limited application that provide a
detailed description of the test method etc and/or the
procedure, process, material, physical, functional,
product or service for use
primarily in procurement and performance or conformance
manufacturing characteristic thereof
– Standards may be referenced in – A physical embodiment of a
specification
unit of measurement (a
caesium block clock)
20
Aspect of Quality
21
Aspects of Quality
• Quality of design
– Implies that the product or service must minimally possess to
satisfy the requirements of the customer
– Design must be simplest and least expensive too
– Influenced by such factors as the type of product, cost, profit policy,
demand, availability of parts and materials, product safety etc.
– Eg: Customer requirement for yield strength of a cable is
100kg/cm2
– The parameters that influence the yield strength would be selected
to minimally satisfy this requirement
– In practice, the product is over designed with a safety factor k =
1.25, so the cable will be designed for 125 kg/cm2
– Increase in designed quality level will lead to increase in cost at an
exponential rate
22
Aspects of Quality – contd.
23
Aspects of Quality – contd..
Cost • Quality of design – Effect

Cost of increase
or – Value of product increases
value at a decreasing rate, with
Valu the rate of increase
e approaching zero beyond a
certain quality level
– At the middle, it maximizes
the difference between
value and cost, given that
minimum requirement ‘A’
is met.
Quality of design or designed

quality level
24
Aspects of Quality – contd.
25
• Quality of conformance
– Implies that the manufactured product or service
rendered must meet the standards selected in the
design phase
– With respect to manufacturing, it is concerned with the
degree to which quality is controlled from the
procurement of raw material to the shipment of
finished goods
– It consists of 3 broad areas:
• Defect Prevention: deals with means to prevent the occurrence
of defects and is usually achieved using SPC techniques
• Defect Finding: Done through inspection, test and statistical
analysis of data from the process
• Defect analysis and rectification
26
• Quality of performance
– Is concerned with how well the product functions or
service performs, when put to use
– It measures the degree to which the product or service
satisfies the customer
– It is a function of both quality of design and quality of
conformance
– If a product does not function well enough to meet the
expectation of a customer or standards, then
adjustment need to be done in the design or
conformance phase
• Relation between QOD, QOC, QOP
27
Quality of
conformance
Quality of Quality of
design performance
28
• Relation between QOD, QOC, QOP – contd..

– QOD has an impact on the QOC
– Eg. If length of a shaft is 20 +/- 0.2mm, then the tools,
equipment and operations should be able to achieve
this design specification , ie. QOC should be capable of
meeting QOD, if not QOD is affected
– This result in constant exchange of information
between QOC and QOD or design and manufacturing
phases for a feasible design to be achieved
– Only single line interaction between QOC and QOP,
because, if QOD is met in QOC phase, then the QOP will
always be good and hence no double line
29
Quality Control
30
Quality control
• It is defined as a system that is used to maintain a

desired level of quality in a product or service
• Can be achieved by planning, design, use of proper
equipment and procedures, inspection and taking
corrective action in case of deviation
• 3 main sub areas
– Off-line quality control
– Statistical process control
– Acceptance sampling plans
31
Quality control – contd..
• Offline Quality Control

– Procedures dealing with measures to select and choose
controllable product and process parameters in such a
way that the deviation between the product or process
output and the standard will be minimized
– Design of products and processes within the constraints
of resources and environmental parameters, so that
output of production meet the standards
– Most of the product and process parameters are set
before production
– Examples: Experimental design, Taguchi methods
32
• Statistical Process Control

– Involves comparing the output of a process or service
with a standard and taking remedial action in case of a
discrepancy
– Also involves whether a process can produce a product
that meets desired specification
– Generally SPC are online control procedures
– The information is gathered about the product, process
or service while it is functional and when the output
differs from a norm, corrective action is taken during
operation
33
• Statistical Process Control – contd..

– It is most desirable to corrective action on a real time
basis, as it minimizes the
• No. of unacceptable items produced
• Time over which undesirable services rendered
– Prevailing theme of quality control is that “Quality has
to be designed into the product or service; it cannot be
inspected into it”
– Though offline quality control measures are taken, there
may be a need for online quality control, because
variation in the manufacturing stage of the product or
the delivery of a service is inevitable
– A combination of offline and online quality control
measures will lead to a desirable level of operation
34
• Acceptance sampling plans

– Deals with inspection of product or service
– Used when 100% inspection is not possible
– Requires a decision on how many items should a sample
contain?
– Sample is inspected
– Compared with acceptance number, i.e. no. of
nonconforming items allowed in a sample and if less,
the whole batch is accepted
– Acceptance criteria is based on meeting certain
stipulated conditions (such as the risk of rejecting a
good lot or accepting a bad lot)
35
Quality Assurance
• Quality is not the responsibility of one person in the

organization
• Role and purpose of the quality assurance function is to
have a “System that ensures all procedures that have been
designed and planned are followed”
• Objective of quality assurance is to have in place a formal
system that continually survey the effectiveness of the
quality philosophy of the company
• Definition: “It refers to all those planned or systematic
actions necessary to provide confidence that a product or
service will satisfy given needs”
• Quality assurance is conducted by carrying out a AUDIT to
check whether the given procedures are followed or not
36
Quality Improvement
Is a never ending process to reduce both the variability of process and the
production of nonconforming items
Process control deals with identification and elimination of special causes that
force a system to go out of control, while quality improvement relates to the
detection and elimination of common causes
• Common causes • Special causes

– Inherent to the system and are – Not inherent to the system
always present – Impact on the output is not
– Impact on the output may be uniform
uniform – Are controllable by the
– Needs the attention of operator
management – Are those for which an
– Accounts for at least 90% of the identifiable reason can be
quality problems determined like tool wear, poor
raw materials etc.
37
Quality Improvement – contd..
• Supports the principle that no deviation from a

standard is acceptable, which is similar to the
principle of the loss function developed by Taguchi
• Methods for quality improvement are graphical
techniques like Pareto chart, fishbone diagrams,
Process capability, DOE etc.
• Three stages of quality improvement
– Commitment stage
– Consolidation stage
– Maturity stage
38
• Three stages of quality improvement – contd..

– Commitment stage
• Management agrees to take QIP
• Formal plan and policy are developed and organization structure is
put into place
• 2 stages: Introduction and Initial implementation
• Introduction (3 to 6 months)
– Policy statement
– Plans for creating infrastructure to implement quality improvement
methods
• Initial implementation (6 months to 1 year)
– Problem identified
– Elimination of special causes
– Process is brought into statistical control
• Time depends on business climate, type of company, product or
process, resource availability, management perception
39
• Three stages of quality improvement – contd..

– Consolidation stage
• Objective is to produce an item that conform to requirements
• Start a continual improvement in the efficiency and
productivity of the process
• Improving common causes is necessary, which improves
process capability
• Involves huge investment to reduce error
• Education and training are integral part
– Maturity stage
• Quality improvement and management are way of life
• Very few defects are produced and focus in on defect
prevention
40
Examples of quality problems
41
Examples of quality problems
42
References
• Text book
• www.iisd.org
• YouTube video
https://
www.youtube.com/watch?v=guJ4I3O8DeU
https://www.youtube.com/watch?v=Jxn2sdZ
TLH0
43
Thank You
44
Reliability

Pilani Campus
Quality Circle and Quality Improvement
Team
• Quality circle • Quality Improvement team
– Informal group – Formal group
– Participation is voluntary – Participation is mandatory
– People from different
– People from same area or department
dept. – Work towards improvement
– Consists of operators, of product or process from
supervisors and managers quality perspective
– Work towards improvement – Team gets dismantled after
of product, process or their completion of project
personal well being
– Team never gets dismantled
– Productivity improvement
tool
46
Design to Customer satisfaction
47
Benefits of Quality Control
• Cannot be realized immediately, but on long-term
perspective
• Benefits
– System is continually evaluated and modified to meet the
changing needs of customers and hence a mechanism
exists to rapidly modify product or process
– QC improves productivity as it reduces the scrap and
rework
– QC reduces cost in the long run and helps productivity and
cost reduction go hand in hand
– Improved delivery dates, due to reduction in lead time for
producing parts and SubAssembly are reduced
– Helps to stay competitive
48
Responsibility for Quality
• All department right from top management to

each worker should have responsibility
• For a high product quality, all these functions
should take responsibility
– Marketing and product planning
– Product design and development
– Manufacturing Engineering
– Purchasing
– Manufacturing
– Inspection and testing
– Packaging and storing
– Customer service
49
Total Quality System
• Quality system is defined as the collective plans, activities
and events that are provided to ensure that a product,
process or service will satisfy customer needs
• System approach integrates the various functions and
responsibilities of different units and provides a
mechanism to ensure that org. goals are met through
proper coordination of other departments
• Elements of total quality system (ANSI/ASQC STD Z1.15)
– Policy, planning organization and administration
• A quality policy is developed inline with org. goals
• Quality manuals are created, which gives the detailed procedure and
costs
• ANSI-American National Standards Institute
50
Total Quality System – contd..
• Elements of total quality system (ANSI/ASQC STD

Z1.15) – contd..
– Product design assurance, specification, development
and control
• Product design based on customer requirements
• Tolerances on product characteristics, based on manufacturing
capabilities is developed
• Procedures to check and control the product characteristics to
conform the design standards during production
– Control of purchased materials and component parts
• Procedures to evaluate the capability and performance of
vendors
• Setting of specifications for the incoming materials
• Inspection and checking procedure for incoming materials to
be followed
51
• Elements of total quality system (ANSI/ASQC STD

Z1.15) – contd..
– Production quality control
• Involves determining process specifications, selecting equipment,
training personnel, designing forms and charts, doing process
capability analysis etc.
• Having preventive maintenance and feed forward control, where
when changes are made in product or processes, the information is
conveyed to those who deal with subsequent operations, with
emphasis on concurrent control during the transformation process
• Developing product quality audits to detect departures from
specified standards
– User contact and field performance
• A system that exists to collect information from the consumer to
determine the level of performance of the product or service
• Developing policies for installation and servicing and determining
guidelines for compliance with these policies to provide customer
satisfaction
52
• Elements of total quality system
(ANSI/ASQC STD Z1.15) – contd..
– Corrective action
• Problems are to be detected, categorized and
systematically documented, similar to that of the
remedial actions
• Specific procedures for corrective action for entities like
materials, vendors, process, equipment should be
developed
– Employee selection, training and motivation
• Guidelines should be established to select people for
particular jobs and job manuals to train people
• Recognition of superior effort and use of motivational
program help reassure employees of the support of mgt.
53
• Elements of total quality system

(ANSI/ASQC STD Z1.15) – contd..
– Legal requirements – Product liability and user
safety
• Providing clear directions on usage of the product
• All advertising statements to be supported by
certified data
• Creation of warranty policy, and policy of
replacement, service etc. for the failure of products
– Sampling and other statistical techniques
54
Quality Cost
55
Quality Costs
56
Quality Costs
• Measure of performance of the total quality system

• 4 major categories for quality costs are:
– Prevention costs
– Appraisal costs
– Internal failure costs
– External failure costs
• Prevention costs
– Incurred in planning, implementing and maintaining a quality
system
– Includes salaries, development cost for product design, process,
equipment design, info. System design, education and training, cost
of a quality audit
– Prevention costs increase with the introduction of a quality system
and may be a significant proportion of the total quality system
– It increases with time
57
Quality Costs – contd..
• Appraisal costs
– Are those associated with measuring, evaluating or
auditing products, components, purchased materials to
determine the degree of conformance to the specified
standards
– It includes inspection and test of incoming material,
product inspection and testing, cost of calibrating and
maintaining measuring instruments etc.
– These costs occur during or after production, but before
the product is released to the customer
– It decreases with time
58
Quality Costs – contd..
• Internal failure costs

– Incurred when products or services fail to meet quality
requirements prior to the transfer of ownership to the
customer
– Includes scrap & rework costs for the materials, labor
and overhead associated with production.
– Also includes cost of lost production time, downgrading
cost, re-inspection cost etc.
• External failure costs
– Incurred when the product does not perform
satisfactorily after ownership is transferred to customer
– Includes cost of adjustment, investigation, costs due to
receipt handling, repair and replacement of parts,
warranty charges etc.
59
Quality Costs – Examples
• External quality cost
60
• Internal quality cost
61
• Prevention cost
62
• Appraisal cost
63
Measuring Quality Costs
• Magnitude of quality cost cannot be considered

because conditions often change like no. of units
etc which affects the direct cost of labour and
materials and hence total costs
• 4 indices were developed as a measurement base
– Labour base index
• Is defined as the quality costs / direct labour hour
• Should be used for short periods, as during extended periods,
the impact of automation may be significant
• Mainly useful for line and middle management
64
Measuring Quality Costs – contd..
• 4 indices – contd..
– Cost base index
• Quality costs / dollar of manufacturing costs
• Includes direct labour, material and overhead costs
• More stable than labour base index because its not affected by
price fluctuations or changes in level of automation
• Important for middle management
– Sales base index
• Quality costs / dollar of sales
• Used by top management
• Not a good measure for short-term, but for long-term strategic
decisions, because sales lag behind production and are
subjected to seasonal variations
• Changes in selling price also affect this index
65
Measuring Quality Costs – contd..
• 4 indices – contd..
– Unit base index
• Quality costs / unit of production
• Used only when output from different production lines are
similar
• For dissimilar ones, the product lines have to be weighted and
standardized product measure computed
– For all these indices, a change in denominator causes
the value of the index to change, even if quality cost do
not change
– Cost of direct labour decreases due to productivity
improvement, labour base index decreases, and it
should not be treated as increased quality costs
66
Effect of Quality on..
• Productivity
– Quality decreases productivity is a misconception
– Quality relies on “make it right first time”, which
increases productivity
– Quality reduces waste (less scrap and rework) and
hence valuable resources can be utilized for production
of defect free goods.
• Effect on cost
– Prevention and appraisal cost
• Increases with initial improvements in productivity and due to
adequate process control procedures are installed
• With time, a reduction in appraisal costs is observed
• These costs are called costs of conformance to quality
requirements
67
Effect of Quality on..
68
Effect of Quality on.. – contd..
• Effect on cost – contd..

– Internal and External cost
• With improved quality, these costs decrease, which usually
offset the increase in prevention and appraisal cost
• Reduction in rework an scrap results in increased time for
productive output and hence the profitability
• Effect on Market
– An improvement in quality lead to increased market
shares, improved competitive position and increased
profits
– Due to decrease in external failure costs and increase in
performance of product
– High price can be charged and due to efficient resource
utilization, a firm can reduce cost of product
69
Management of Quality
• Achieved through functions of planning, organizing,
staffing, directing and controlling
• Eg. Product development
Management function Product phase Action to be taken
Plan Proposal - Develop quality policy

phase - Plan for quality
- Consider product liability / safety
Organize Design / - Develop an org. structure

planning - Design assurance
phase - Design change control
- Develop production quality planning
70
Management of Quality – contd..
Management Product phase Action to be taken
function
Staff Pre production - Select employee

phase - Train employee
- Motivate employees
Direct Production phase - Monitor purchased materials

- Monitor process control
- Direct final inspection
- Direct handling inspection
Control Production and Post - Obtain quality info.

production - Get data on field performance
- Take corrective action
- Conduct SQC
- Manage quality cost
71
Total Quality Environmental Management
• Social responsibility of the organization is to operate and

maintain safe and adequate environmental conditions
– Eg. Xerox uses only recyclable and recycled thermoplastics and
metals
• ISO 14000 standards promote this philosophy with the
objective of uniform environmental management
standards that do not create unnecessary trade barriers
• It involves
– management commitment to continuous improvement,
– compliance and pollution prevention,
– creating and implementing environmental policies,
– integrating environmental considerations in operating procedures,
– conducting audits of the environment management system etc.
72
The objectives of Total Quality Environmental
Management are to:
• reduce waste and improve continuously,
• reduce resource depletion,
• reduce or eliminate environmental pollution,
• design products for minimal environmental impact in production, use and
disposal,
• control environmental impact of raw material sourcing,
• control environmental impact of new developments,
• promote environmental awareness among employees and
• promote environmental awareness within the community.
73
Case study- BBC
• The BBC spends £700 million (nearly $1 billion) a year on

goods and services, with no fewer than 20,000 suppliers.
• It employs 24,000 people, who collectively travel 17
million kilometres a year by road and 38 million by air.
• The BBC occupies more than 500 buildings in the UK.
• In London alone, it produced 4,600 tonnes of waste in
2001, 85% of which was landfilled.
74
Environment impact group
• Waste
• Utilities (energy & water)
• Transport
• Supply chain and procurement
• Property
75
Challenges
• BBC is the ongoing transition from analogue to

digital broadcasting.
• The latter has a higher electricity demand, which
makes it difficult to achieve a reduction in overall
greenhouse gas emissions.
• To make matter worse, both systems are being run
simultaneously until the changeover to digital is
complete.
76
Action
Waste
• Complete a comprehensive audit of waste throughout the BBC,
concentrating on paper, tapes and toner cartridges;
• Recycle at least 25% of office waste by 2003, rising eventually to 55%.
Utilities
• Each building manager to select an energy efficiency initiative, to be
run as a pilot programme;
• Reduce carbon dioxide emissions by 1% in the next year
Transport and travel
• Reduce car travel at the London White City site by 20% over four years,
by encouraging public transport use, cycling, teleworking and walking;
• Target the business units which require the most intensive transport
use.
77
Supply chain and procurement
• Conduct an environmental risk analysis on suppliers and products;
• Identify what the BBC spends most money on, and prioritise those areas for assessment;
• Set up partnerships with suppliers to improve environmental performance;
• Develop an assessment scheme that enables the BBC to identify environmentally active
suppliers;
• Update the BBC's purchasing policy to incorporate environmental objectives;
• Introduce awareness training for all staff who procure goods and services.
Property
• Run a site compliance programme to set a baseline for future improvements;
• Develop an environmental brief for building 'churn', the process of teams coming together in a
space over a fixed time period;
• Improve specification and tendering documentation;
• Include environmental considerations at financial approval stage;
• Implement the 'Considerate Constructors Scheme' on BBC contracts;
• Develop guidance notes for projects, construction and facilities management staff;
• Build all new buildings to BREEAM (Building Research Establishment Environmental
Assessment Method) 'Excellent' standards
78
Trends in Quality
• Quality certification
• Quality as project
• Quality in role of strategy
• Quality and sustainability(people and QMS)
79
80
81
References
• Text book
• www.iisd.org
• GEMI practice
• Curkovic, Sime, and Robert Sroufe. "Total
quality environmental management and total
cost assessment: An exploratory
study." International Journal of production
economics 105.2 (2007): 560-579.
• YouTube video
82
Homework
• TQEM practice for your organization

• Identifying cost of poor quality
• Quality consciousness
83
Thank You
84
Reliability

Pilani Campus
Chapter - 2
• Some philosophies and their

impact on Quality
86
Some Quality Gurus
87
Deming’s contribution
• He emphasized on the role of management
• His approach is not a quick fix, rather a plan of
action to achieve long term goals
• Heart of his philosophy is that management and
workers should speak a common language – the
language of statistics
• His 14 points will form a framework for action and
it needs to be installed for the program to be
successful
• He identified the key components for continuous
improvement / “System of profound knowledge”
– Knowledge of the system and the theory of
optimization
– Knowledge of the theory of variation
– Exposure to the theory of knowledge88
– Knowledge of psychology BITS Pilani, Pilani Campus
<Course Code> 89
<Course Code> 90
Deming’s contribution –
contd..
• “System of profound knowledge” – contd..
– Knowledge of the system and the theory of
optimization
• Management needs to understand that optimization
of total system is the objective and not optimizing
the sub system to have sub-optimal total system
– Knowledge of theory of variation
• Management should understand special causes and
common causes in the variation of process
• Special causes can be controlled by operator or
engineer
• Common causes should be reduced by management
only, which makes the system stable and
91
in control
Deming’s contribution –
contd..
• “System of profound knowledge” – contd..
– Exposure to the theory of knowledge
• Experience and intuition are not of value to
management unless they can be interpreted and
explained in the context of a theory
• A data-analysis oriented approach to problem solving
is needed to suggest remedial action based on results
– Knowledge of psychology
• It is required to understand the behaviour and
interactions of people, and also the interactions of
people with their work environment
• People are motivated by intrinsic and extrinsic factors
– Job satisfaction and motivation: Intrinsic
– Reward and Recognition: Extrinsic
• Management should create the right mix of these
factors 92
Deming’s 14 points for
management
• Focus is on Management
• He expects a fundamental change in the style of
management and corporate culture
• Corporate culture should be in such a way that the
workers feel comfortable enough to recommend
changes
• Point 1: “Create and publish to all employees a
statement of the aims and purposes of the company”
– Management must demonstrate constantly their commitment
to this statement
– Emphasis is on long-term strategic plans and also on mission
statements that need to be understood by employers,
consumers, vendors and investors too.
– Can be achieved by
• Product Improvement cycle
• Constancy and consistency of purpose
93
management
94
management
95
Deming’s 14 points – contd..
• Point 1 – contd..
– Product Improvement cycle (Figure)
• It integrates the phases of QOD, QOC, QOP
• All customer needs must be quantified
• Company must focus on the customer and not on the
competitor
– Constancy and consistency of purpose
(Figure)
• Implies setting a course (say all dept. follow the
quality improvement) and keeping to it
• Allocates resources for long-term planning, employee
training and education, which ensures that it can
maintain a competitive position for years96 to come
– Constancy and consistency of purpose – contd..
• Company should not deviate from long-term
objectives
• i.e. it should accept that variability exists in any
operation and hence it should try to reduce
• Doing your best is not good enough, you have to know
what to do, then do your best
• Point 2
– Learn the new philosophy, top management and
everybody
• Quality consciousness must be everything to everyone
• Acceptable level of defect should be abandoned and
one should work towards reducing defects
continuously and also address the need of customer
97
• Point 3
– Understand the purpose of inspection, for
improvement of processes and reduction of cost
• Inspection does not increase quality, it just separates out
acceptable and non acceptable items
• It does not address the root cause of problem
• Hence emphasis should be placed on defect prevention
rather than defect detection
– Drawbacks of mass inspection
• Even 100% inspection will not eliminate all the defectives if
more than one person is responsible for inspection
• Reason being that, it is only human to assume that others
will find what you have missed
• Also the inspector fatigue should be taken into account
98
• Point 4
– End the practice of awarding business on the basis of
price tag alone
• Companies should review the bidders approaches to quality
control, and if possible should know what is the fraction of non
conforming items and the stability of process etc.
• Generally contract should be based on effective price per unit
• Eg. 16Rs. With 8% defective and 17Rs with 2% defective
– Principle of vendor selection
• While selecting a vendor, the total cost should be taken into
account
• Vendors and buyers should work as a team to choose methods
and materials to improve customer satisfaction
• It is desirable to have reduced no. of suppliers
• Having multiple suppliers results in mistrust between vendor
and buyer and hence no long-term relationships
• Price not quality is the driving factor
• Also results in variability in the incoming quality and hence
increased cost due to changes in setup, no99 volume discounts
etc. BITS Pilani, Pilani Campus
• Point 5
– Improve constantly and
forever the system of
production and service
• Focus is on defect
prevention and
process improvement
which are carried out
by the use of statistical
methods
• Process improvement
should be based on
Deming’s cycle which
consists of 4 stages:
Plan Do Check and Act
(PDCA)
100
• Point 5
– Plan
• Opportunities for
improvement are
recognised and
operationally defined
• Since customer
satisfaction is
important, degree of
difference between
customer needs
satisfaction and
process performance
are analysed
• Goal is to reduce the
difference and the
possible relationship
between the variables
in the process and
their
101 effect are
hypothesized
– Do
• Cause of action
developed in
planning is put
into action
• If necessary trial
runs or prototype
are conducted
• Feedback is
obtained from
process and
customer
102
– Check
• Analysing the result
• Statistical methods
will be used
– Act
• Decision is made
regarding
implementation if the
results of the check
stage are positive and
if negative alternative
plans are developed
103
– Variability reduction
and loss function
• Deming’s philosophy calls
for abandoning the idea
that everything is fine if
specifications are met
• Based on this Taguchi
formalized loss function
in 1960
• i.e. It explains that
economic loss accrue
with any deviation from
the target value
• However the loss
increase in a non linear
relationship with larger 104
deviation from target BITS Pilani, Pilani Campus
• Point 6
105
• Point 6
– Institute training
• Employee training is integral for individual performance in
the extended process setting
• When employees are hired, they should be carefully
instructed in the company’s goals and this helps them to
understand their responsibilities for meeting customer
needs
• A employee should know what is to be done and its
importance in the entire process, which will make them
take pride in their work
• It make them feel secure, have better morale and hence
improve their productivity
• Training should be ongoing
• According to Deming, it is necessary for all employee of the
organization to be trained in SQC tools106 BITS Pilani, Pilani Campus
• Point 7
107
• Point 7
– Teach and institute leadership
• Supervisors are the vital links between
management and the workers
• To be effective leader, the supervisors
must think in terms of helping workers do
a better job and manage them
• Supervisors should be trained in SQC and
should create a atmosphere, which
improves employee morale, promotes
team work and help achieve the overall
annual goal of quality improvement
108
• Point 8
– Drive out fear, create
trust. Create climate for
innovation
• Fear results in physical
or psychological
disorders, poor morale
and also results in lack of
job security and hence
reduces productivity
• An employee should feel
free to know about the
process, about his
responsibility and
should be able to suggest
something, which helps
in building trust and a
climate
109 for innovation
• Point 9
– Optimize toward the aims
and purposes of the
company, the efforts of
teams, group, staff areas
• Focus is to remove
organizational barriers –
both internal and external
• Internal barriers will result
in impeding the flow of
information /
communication between the
different level of employees
and also between
departments
• External barrier is the lack
of flow of information
between the customers,
investors
110 and vendors,
which should bePilani,
BITS eliminated
Pilani Campus
• Point 10
– Eliminate
exhortations for the
workforce
• Never set a target
arbitrarily as it has
demoralizing effect
• Set the goal, which
should be feasible and
the management
should specify the way
to achieve it
• Goals should not be
intuitive or arbitrary,
but should be based
on information from
every level
111
• Point 11
– Eliminate numerical
quotas for production.
Instead learn and
institute methods for
improvement
– Eliminate management
by objectives. Instead
learn the capabilities of
processes and how to
improve them
• Work standards should
not be fixed, because it
is based on quantity
and not on quality.
112
• Results in workers strive to meet the quota rather
than to produce acceptable goods, resulting in
inefficiency and increased costs
• Further it results in loss of pride in their work,
morale, motivation
• Generally, in work standards, allowance is made
for producing non conforming items, which is
against the philosophy of continuous
improvement
• Work standards fails to distinguish between
special causes and common causes, when
improvements in process are sought
• Common causes can be eliminated only by
management and if it can’t be eliminated if there
is no effort from management 113
• Point 12
– Remove barriers that
rob people the pride
of workmanship
• Quality is achieved in
all components of the
extended process,
when the employees
are satisfied and
motivated and when
they understand their
role in the context of
organization goals
114
– Factors that cause loss of pride
• Lack of treatment of employee with dignity
• Lack of communication of company’s mission to all
levels
• Assigning blame to the employees for not meeting
company goals though the output has problems due to
system
– Performance classification system
• Inconsistencies in performance evaluation and there
should not be categorization of employee from a similar
sample
• Teamwork should be given importance, when
conducting performance appraisal, 115
those promoting
teamwork should be placed in higher category
• Point 13
116
• Point 13
– Encourage education and self improvement for
everyone
• Management must commit resources for education and
training as educated workforce can safeguard
improvement in quality and productivity and help
companies to remain competitive
• Point 6 is about training for new employee, while this
addresses the need for ongoing and continual
education, to adopt to the latest technology
• Advantages
– Increases motivation, as they feel that company is
spending for them and hence loyalty too increases
– Keeps the employees up to date on latest technologies
and promote teamwork
117
– Helps to adopt different jobs, as they grow inPilani,
BITS companyPilani Campus
• Point 14
– Take action to accomplish the transformation
• A structure must be created and maintained for the
dissemination of the concepts
• People must be trained in statistics at all levels of the
organization
• Companies should allow free flow of information among
different levels
Deming’s deadly diseases
– Management by visible figures only
– Lack of constancy of purpose
– Performance appraisal by numbers
– A short-term orientation / short term profits
– Mobility of management – short tenure of
management and managers should be 118 avoided
Thank You
119
Reliability
Pilani Campus
Crosby’s Philosophy
• He developed Quality Management Grid

• This grid has 5 stages of maturity and 6 measurement
categories
• Stages: Uncertainty, Awakening, Enlightenment, Wisdom,
Certainty
• Categories
– Management understanding and attitude
– Quality organization status
– Problem handling
– Cost of quality as a % of sales
– Quality improvement actions
– Summary of company quality posture
• Example: Problem handling
– Uncertainty: Fire fighting, yelling and accusation
– Awakening: Team setup to attack problem
– Enlightenment: Corrective action communicated in systematic
way
– Wisdom: Problem identified early on development stage
121
– Certainty: Problems is prevented
Crosby’s Philosophy – contd..
• 4 absolutes of quality management

– Definition: Quality means conformance
to requirements
– System for achievement of quality: The
rational approach is prevention of defects
– Performance standard: Only
performance standard is zero defects
– Measurement: Performance
measurement is the cost of quality
122
• 14 step plan for
quality improvement
– Management
commitment
• Emphasis on defect
prevention should be
communicated
• Quality policy should
states that individual
performance
requirements must
match the customer
requirements
123
• 14 step plan – contd.
– Quality improvement team
• Representation from each department or division
form the team
• Responsible for bringing suggestions to actions
124
• 14 step plan for quality
improvement – contd..
– Quality measurement
• Identifies the status of
quality
• Identifies area where
corrective action is
needed and quality
improvement efforts
should be directed
• Results should be
placed in highly visible
charts
125
• 14 step plan for quality improvement – contd..
– Cost of quality evaluation
• Indicates where corrective action and quality improvement
will result in savings for the company
• It establishes a measure of managements performance
126
– Quality awareness
• Results about the cost of non quality should be shared with
all employees including service and administrative people
127
– Corrective action
• Needs open communication and action discussion of
problems
• Discussions exposes other problem not identified before
and helps to determine procedures to eliminate them
• Attempts to resolve problem should be made as they arise
128
quality improvement
– contd..
– Adhoc committee for
the zero defects
program
• Concept of zero defect
to be communicated
for all employee
• Committee is
responsible for getting
top management
commitment
129
quality improvement
– contd..
– Supervisor training
• All levels of
management must be
aware about the steps
of quality
improvement
programs
• They should be
explaining the
program to
130
employees
quality improvement
– contd..
– Zero defects day
• Philosophy of zero
defects should be
established
companywide and
should originate on
one day
• Management should
foster this type of
quality culture and
should try to motivate
the people
131
quality
improvement –
contd..
– Goal setting
• Employees along
with their
supervisors should
set specific
measurable goals,
which creates a
attitude among
people to achieve it
132
• 14 step plan for quality
improvement – contd..
– Error cause removal
• Employees are asked to
identify reasons that
prevent them meeting zero
defect goal – They should
not make suggestion, but
list the problem
• A separate functional
group should come up with
procedures for removing it
– Recognition
• Award programs should be
based on recognition
rather than money, who
have met or exceeded their
goals 133
• 14 step plan for quality improvement

– contd..
– Quality councils
• Chairpersons, team leaders and
professionals associated with quality
program meets and discusses, which create
new ideas for further improvement
– Do it over again
• Whole process of quality improvement is
continuos
134
Juran’s Philosophy
• He believes that management has to adopt a unified

approach to quality, where the focus is on needs of the
customer
• Quality is defined as “fitness for use”
• Deterrents of unified approach are
– Existence of multiple functions and they believe it to be unique
– Presence of hierarchical levels in the organization structure
which has varied background and have different level of
exposure to quality management
– A variety of product lien that differ in markets, its production
processes etc.
• Unified approach is called “Quality Trilogy”, which
includes
– Quality planning
– Quality control
– Quality improvement
135
Juran’s Philosophy
136
Juran’s Philosophy – contd..
• Quality planning
– Occurs at various levels of
organization, each having a
distinct goal
– Strategic quality
management, where broad
quality goals are set, and a
management chooses a
plan of action and allocates
resources to achieve the
goals
– Operational quality
management, where
departmental goals are set,
inline with strategic goals
– At worker level, clear
assignment of task to each
worker to contribute to
departmental goal
137
• Quality control
– Goal is to run the process
effectively, such that plans
are enacted
– Any deficiency in planning
process, results in process
operating at high level of
chronic waste
– If unusual symptoms are
sporadically detected,
quality control will try to
identify the cause behind
the variation and remedial
action taken to bring it back
to control
– Objective is to eliminate the
cause of sporadic spike and
bring process output to
zone of quality 138
• Quality improvement
– Deals with continuous
improvement of product
and process
– It requires an action on
upper and middle
management
– Deals with actions like
creating new designs,
changing methods or
procedures, new
equipment investment
etc.
– Results in reducing the
chronic waste and hence
the cost of quality
139
Quality trilogy – in detail
• Quality planning
– Identify the customer – both external and internal,
which is similar to that of Deming’s extended process
– Determine the customer needs. This decides the long
term survival of customers
– Develop product features that respond to customer
needs
– Establish quality goals that meet the needs of
customers and suppliers like and do so at a minimum
cost. Total cost from an organization point of view
should be minimum
– Develop a process that can produce the needed
product features
– Prove process capability: To establish whether the
process is adequate for making a product that 140
Quality trilogy – in detail: Contd..
• Quality control
– Choose control subjects
• Identify the product characteristics that is to be controlled in
order to make the product conform to requirements
• Selection is done by identifying the characteristics that has more
impact on customer
– Choose units of measurement
– Establish measurement
• Procedures for taking measurements are defined
• Involves identifying what equipments to be use, how to use, who
will be responsible, how to make measurements an when to
calibrate, what training is needed etc.
– Establish standards of performance
• Should be based on customer requirements
– Measure actual performance
• Concerned with measurement of actual process output; to know
the operation level of process
141
• Quality control – contd..
– Interpret the difference
• Compare actual and standard and if the process is stable,
the difference may not be significant
– Take action on the difference
• Take remedial action for the difference
• It is management’s responsibility to suggest a remedial
cause of action
• Quality Improvement
– Prove the need for improvement
• Identify problems and convert it into dollar figures to
attract management attention
• Convince the management and involve the management to
make a change in the process
– Identify specific projects for improvement
• Prioritize the problems, based on Pareto analysis as only
limited resources will be available 142
• Quality Improvement – contd..

– Organize to guide the projects
• Establish a clear organization structure
• Assign authority and responsibility at all levels of
management to facilitate this
• Establish precise responsibilities: guidance for overall
improvement program, guidance for each individual
project and diagnosis and analysis for each project
– Organize for diagnosis – for discovery of causes
• A “Diagnostic Arm”, a group of persons brought together to
determine the causes of the problem
• Organization need to enlist the right people to ensure that
the required tools and resources. This is done by “Steering
Arm”
– Find the causes
• Involves data gathering and analysis to determine the
causes
• Involves studying the symptoms, hypothesising the causes143
and validating it BITS Pilani, Pilani Campus
• Quality Improvement – contd..
– Provide remedies
• Diagnostic step is based on cause and effect relationship
• Remedies may be for problems that are controllable by
management and those that are controlled by operators
• It may involve change in methods or equipment, which may
require substantial financial investment
– Prove that the remedies are effective under operating
condition
• Check whether the process can be implemented and will it have
benefits
• Breakthrough process requires overcoming resistance to change
• Proposed procedure may require new equipment and operators
may have to be trained
• Management commitment is vital
– Provide control mechanisms to hold the gains
• Once remedial action implemented, it need to be sustained,
which needs a control system like audit in some departments 144
3 philosophies compared
Criteria Deming Crosby Juran
Definition of - Deals with - “Conformance -“Fitness for
quality predictable to requirements use”
uniformity of - Emphasis on - explicitly
product customer needs related to
- Emphasis on use implicitly meeting the
of SQC - Emphasize on needs
- Quality of product zero defects, the
depends on quality requirement
of process should be met
Management - Highly - Focussed on - Scales
commitment emphasized creating quality management
- Points 1, 2 and 14 culture which commitment at
are aimed at needs all levels
management management
commitment
145
3 philosophies compared –
contd..
Strategic - Top management - Need for - Need for
approach to involves pursuing quality quality council
quality all 13 points in a improvement and also
system structured cycle team emphasized on
problem-solving
steering and
diagnostic arms
Measurement - Depends on - Quality is free
of quality deriving a dollar and it is that un-
value for the total quality that costs
cost of quality
Never ending - Follow Plan-Do- - Repeats the Believes in
process of Check-Act cycle cycle of quality breakthrough
improvement and repeats planning, control sequence
and improvement
146
contd..
Education - Emphasizes on - Emphasizes on - His stress on
and training training of new quality culture, education is
employee and also where education implicit
re-training on is a part of it
continuous basis for
existing people
Eliminating - Categorized as - No such - Special causes

the cause of special and categorization create sporadic
problem common causes problem
- 85% of problem - Common
can be solved by causes create
management Chronic
problem
147
contd..
Goal - No numerical goals - Explains that - Recommends an
setting should be established a mutually annual quality
- Never ending quality understandable improvement
improvement process goal, which is program with
to be followed and measurable specified goals
hence no short-term - Goals should be
goals set according to
requirements of
customer
Structural - Bottom up approach: - Top down - Quality

plan Bring process to approach improvement
statistical control: - Change in through a project
lower level management by project method
- common cause culture - used for middle
removed by top mgt. management 148
Thank You
149
Quality Management –
Practices, Tools and Standards
150
151
Management Commitment
• Should have strategic plans with long-
term focus
• Should satisfy the customer and
investor by delivering quality products
• Can be accomplished by having a
‘Quality Program’ like TQE, TQM or Six
Sigma etc.
• Eg. Ford Motor Company follow Total
Quality Excellence (TQE) program
152
Total Quality Excellence
• (TQE)
A system that integrates
all aspects of the company
and its suppliers into
strategic decision making
on different functions of a
company
• It is based on Ford’s vision
of “being a low cost
producer of the highest
quality products and
services which provide
the best customer value” 153
TQE – Concepts, Functional
Areas
• Key Concepts
– Quality should be defined by customer
– Quality is achieved through prevention
of problems and not detection
– Vendor-Vendee concept, (Internal and
External Customer)
– All employees, suppliers, dealers and
part of the company (Deming’s
extended process)
– Quality improvement is never ending 154
TQE – Contd..
• Functional Areas includes
Engineering, Delivery and
Commercial Performance
• Embraces the system approach
• Relies on partnership between
manufacturer and supplier
– Ford developed ‘Q101 Preferred
Quality Award’ for suppliers
• Emphasize on Process control, Customer
feedback and Documentation including
management developed documentation 155
Q 101 Ford’s Preferred Supplier
Award
156
Basic concepts of TQM
•What are the Basic Concepts of TQM. Explain?

• A committed and involved management to provide long
term top-to-bottom organizational support
• An unwavering focus on the customer – both internal and
external
• Effective involvement and utilization of the entire work
force
• Continuous improvement of the business and production
process
• Treating suppliers as partners
• Establish performance measures for the process
157
Old vs. TQM Approach
Quality Element Previous Approach TQM Approach

Definition Product-oriented Customer-oriented
Priorities 2nd to service and Equals of service and cost
cost
Decisions Short-term Long-term
Emphasis Detection Prevention
Errors Operations System
Responsibility Quality control Everyone
Problem Solving Managers Teams
Procurement Price Life-cycle

costs,partnership
Manager’s Role Plan, assign, control, Delegate, coach, facilitate
and enforce and mentor
158
159
Total Quality Management
(TQM)
• Management Commitment
– Must exists at all levels of the company
– Creates policy at top level, implements at
middle level and quality mgt tools and
techniques are used at operation level
• Customer
– Satisfying customer needs is the driving
force
– Direct feedback using a data driven approach
is the best way to identify customer needs
– Key principle is that are customers are both
internal and external 160
TQM – Contd..
Customer surveys are used to find the
discrepancies between expectation and
satisfaction and taking measures to
eliminate is called ‘Gap Analysis’
• Process
– Vendors are part of the process
– Cross functional team, with authority to
solve problem and improve the process
– Use of technical tools and techniques and
management tools is necessary to improve
continuously 161
TQM – Contd..
• People
– Involving employees in the decision making
process (Empowerment)
– Ownership and motivation are necessary to
make employee get a sense of pride
• Communication
– Open channel of communication
– Emphasis is on Information sharing between
departments
– Change from role of management of
coordination and control to that of coaching
and caring 162
TQM – Contd..
• Culture
– Links the human element and the
company’s vision
– They are the beliefs, values, norms and
rules that prevail within an company
– Culture should embrace a ‘participative
style’ of management
• Vision
– Tells about what company wants to be
– It is general and outlines the scope and
purpose of organization
– Should be motivational and make people 163
work towards the goal BITS Pilani, Pilani Campus
TQM – Contd..
• Mission
– Derived from vision
– More focused and defines the areas of
concentration
– No mention about time frame
• Quality Policy
– Framed by senior management, it is the
company’s road map
– Indicates what to be done, but different from
procedures and instructions
– Eg. ‘quality is the basic business principle at
Xerox’
164
– Outcome is the performance standardBITS Pilani, Pilani Campus
Vision and Mission Statement
• Vision
– Be recognized as the best education provider
in Mechanical Engineering
• Mission
– The Mechanical Engineering Group is
committed
• to provide a high quality (world-class) education
in Mechanical Engineering
• to conduct strong research programs
• to foster a close partnership with industry and
– The mission is guided by a continuous
improvement philosophy 165
Vision and Mission – contd..
166
Vision and Mission – contd..
167
Six Sigma Quality
• Developed by Motorola
• Made popular by General Electric (GE)
• Some of the companies that has adopted 6 σ
are
– Wipro
– CTS
– Satyam
– General Motors
– Ford India Limited
• Uses quantitative goals, based on the measure
of process variation (σ)
• The objective is to minimize the variation from
the existing level (if it is 3 σ) to half
• The concept of 6 σ is based on the concept of 168
normal distribution BITS Pilani, Pilani Campus
Variation – A Simple Test
• I think the topic of quality initially seems to be

interesting or is it our instructor in charge
from Mechanical Engineering Group making
me feel so? But our seniors has said that it is a
boring course and the instructor always pose
inappropriate and invalid questions. He used
to conduct lot of class test and call it an
assignment making this course really tough.
Even though you study hard for this subject,
finally, he will provide you with the answer
sheet containing innumerable zeros and ones.
Hence most of the times, I will not even bother
to visit the quality class!!!
(Experience of a current student in the class)169
Solution
• I think the topic of quality initially seems to be

interesting or is it our instructor in-charge from
Mechanical Engineering Group making me feel so?
But our seniors have mentioned that it is a boring
course and the instructor always pose
inappropriate and invalid questions. He will
conduct lot of class tests and call it as assignments
making it a hectic course. Even if you study hard
for this subject, finally, he will provide you with
the answer sheet containing innumerable zeros
and ones. Hence most of the times, I will not even
bother to visit the quality class!!!
(Experience of a current student in the class)
• Total no. of ‘i’s = 49 170
A Normal distribution
171
Six Sigma Quality – contd..
• For a process output given by normal
distribution
– Mean is the measure of location of process
– The no. of non-conforming product will be
2700ppm (1350ppm on each tail)
– Assume a product with 1000 processes or
parts, then the no. of defects per product will
be 2.7!!! (only 7 out 100 will be defect free)
• For defect free product, the process spread
given by (+/- 3 σ) has to be significantly
less than that between USL and LSL
172
Thank You
173
Reliability

Pilani Campus
Quality Management –
Practices, Tools and Standards
175
176
Six Sigma Quality – contd..
• For a process output given by normal
distribution
– Mean is the measure of location of process
– The no. of non-conforming product will be
2700ppm (1350ppm on each tail)
– Assume a product with 1000 processes or
parts, then the no. of defects per product will
be 2.7!!! (only 7 out 100 will be defect free)
• For defect free product, the process spread
given by (+/- 3 σ) has to be significantly
less than that between USL and LSL
177
178
Six Sigma – Contd..
• Focus of Six sigma is to reduce the process
variation such that the spec. limits are 6 σ from
the mean
• If process is stable and remains centered, the
proportion of defective will be 0.001ppm on
each tail (Ideal)
• Even for a shifts by 1.5 σ from the mean, no. of
defects will be 3.4ppm on each tail
• For the example above there will be 0.0034
defect per product and the yield will be 99.66%
• Requires a fundamental change in management
philosophy and culture 179
Quality Function
Deployment (QFD)
• A planning tool that focus on designing
quality into a product
• A system that involves cross functional
team and takes care of complete cycle
of product development
• Helps to translate customer needs into
technical requirements of products
• Called as House of Quality, Matrix
product planning, Customer drive
180
Questions answered by QFD
• What do customers want?
• Are all preferences equally important?
• Will delivering perceived needs deliver a
competitive advantage?
• How can we change the product?
• How do engineering characteristics influence
customer perceived quality?
• How does one engineering attribute affect
another?
• What are the appropriate targets for the
engineering characteristics?
181
QFD “house of quality”
182
QFD - Process
• It is based on a structure called House of

Quality (HOQ)
• Objective
– Delineates the scope of the project
– Only one task is specified in the objective
• What
– Voice of the customer, which express their
needs and wants
• Rating
– Scale of 1 to 5 is used to give prioritize the
wants (1 – Least and 5 – Most Important)
183
Customer Rating
Voice of the Customer
184
QFD – Process (Contd..)
• Technical Descriptors (Hows)

– All possible technical / engineering
requirements are found through
Brainstorming
– Done by a cross-functional team
• Target goals
– May or may not be used
– Uses symbols
185
Target goals
Voice of the Designer
186
• Co-relationship matrix
– Forms the roof of HOQ
– 4 levels of relationship – strong
positive, positive, negative, strong
negative
– Used for determining the trade-off
between ‘Hows’
187
x = Design Trade-offs
188
• Customer competitive assessment

– To identify the relative position of the
organization w.r.t competitors for each
needs
– Helps to identify the strength and
weakness of the company and thereby
close the gap
– Called as Product Benchmarking
189
Benchmarking
190
• Technical Competitive Assessment
– Similar to that of customer assessment of
competitor
– Done by a technical staff of the company
– Used to set objective values for the
technical descriptors
191
Technical Competitive
Assessment
192
• Relationship matrix
– A mechanism to analyse how each
technical descriptors will help in
achieving ‘What’
– Represented by 0 to 5, where 0 is no
relationship 1 – low, 3 – medium and 5 -
High
193
194
• Absolute and relative scores are

calculated, based on which
prioritization of needs is done
• In next phase, the ‘Hows’ will
become ‘What’
195
U – Us
M – Maytag
W – Whirlpool
G – GE
F – Frigidaire
A - Amana
196
QFD – The real meaning
• It is used to take the ‘voice of customer’
and translate it into a set of product and
process parameters and can be deployed
through a 4-phase process
– Product planning
– Part / Subsystem deployment
– Process planning
– Production planning
197
QFD – The real meaning – contd.
technical
requirements
component
characteristics
process
operations
quality plan
198
QFD - Advantages
• Evaluates competitor from 2

perspectives – Customer and
Technical
• Used to identify the degree of
improvements needed in the product
and process
• Reduces the product development
time as it reduces product redesign
• Improves the communication
between different department 199
QFD - Disadvantages
• Enormous amount of information must be
gathered
• Matrix is too large
– Prioritize on important attributes
– Analyzes independent subsystems independently
• Customer priorities not clear
– Consider segmenting market
• QFD process is implemented in phases and
hence time consuming
• QFD is messy
– Not QFD, but rather the interaction between diverse
groups is cause
200
– Stick with it, the results are worthwhile
Innovative Adaptation &
Performance Evaluation
• Innovation • Continuous
– Cut costs, reduce Improvement
lead time, improve
– Any one such
productivity, save
capital and human improvement takes
resources and place
hence increased – Happens very
revenue frequently
– Do not happen very – Leads to slow and
often steady increase in
– Result in a steep quality
increase in quality
– Eg. Benchmarking
– High risk strategy 201
Innovation & Cont. Improvement
202
Benchmarking (BM)
• Practice of identifying best practices in

industry and thereby setting goals to
emulate them
• Facilitates continuous improvement
• Started by Xerox in the year 1970 by R.C.
Camp
• Steps for BM differs from company to
company due to different operating
concerns
203
Objective of Benchmarking
•The main objective of any benchmarking activity is:

– To identify key performance measures for each function of
a company’s operations
– To measure the internal performance levels of the company
as well as that of competitors
– To ensure that its business process goals are set to exceed
the best quantitative results achieved by leading
organizations
– To incorporate best practices throughout the organization
business process
– To reach a level of maturity where benchmarking is an
ongoing part of management system in all areas of the
company
204
Benchmarking Process
205
Benchmarking - Requirements
• Not just merely identifying best practices,

it should be how practices can be adapted
• It needs Management commitment and
should motivate employees with rewards
• Training is a must on process mapping,
flowcharting
• All resources should be made available
• Hard and Soft systems should mesh
206
Why Benchmarking?
• Thorough understanding of company’s own
processes, products etc.
• Identify Non Value Added activities
• Provides an effective way to manage change
and also a roadmap to adapt best practices
• Impact of technological development
• Reduce cycle time, which in turn related to
PM
• Enable comparison of performance measures
(PM) in different dimensions with ‘spider
chart’ 207
Spider Chart
208
Quality Audit
• To examine the effectiveness of management

control programs and also to identify
problems
• 3 parties are involved – Client, Auditor and
Auditee
• Auditor – Internal and External
• Purpose
– Suitability Quality Audit: Evaluation of Quality
program against reference standard
– Conformity Quality Audit: Evaluation of
operations, activities within the quality system 209
with respect to procedures BITS Pilani, Pilani Campus
Quality Audit – Contd..
• Types
– System Audit
• Most extensive
• Evaluation of the quality program
documentation with respect to a reference
standard
• Includes evaluation of activities and
operations that are followed to meet the
desired quality objectives
210
• Types – contd..
– Process Audit
• In depth evaluation of one or more
processes
• Takes less time
• More focussed and less costly
• Used generally in process industries like
chemical industries
• Done when there is need for process
improvement or when there is unexpected
output
211
• Types – contd..
– Product Audit
• Assessment of final product or service on its
ability to meet customer needs
• Involves conducting periodic test on products
or obtain info. from customers
• Separate from decision on product
acceptance or rejection and hence not part of
the inspection system
212
• Audit Quality
– Influenced by independence and objectivity of the
auditor
– Depends on being a external or internal auditor
– For suitability audit, external is most preferable
• Methods
– Location oriented
• All functions are audited in that particular location
• Examines the action and interaction of the elements in the
quality program at that location and may be used to identify
discrepancies between location
– Function oriented
• Examine and evaluate activities related to a particular element
or function within a quality program at all locations
• Successive visits to each location is necessary to complete the
audit
– Utility of quality audit is derived only when remedial
actions are taken 213
Tools for Continuous
Quality Improvement
• For solving the problems and involves
data collection, analysis, hypothesis and
validation
• Called as 7 QC tools
– Check sheets
– Histogram
– Flow chart
– Pareto chart
– Cause and effect diagram
– Control chart
214
– Scatter diagram BITS Pilani, Pilani Campus
What are tools and techniques?
• Tools
– Tools and techniques are practical methods, skills,
means or mechanisms that can be applied to particular
tasks
– Used to facilitate positive change and improvements
– Narrow in focus and is usually used on its own
– Examples of tools are 7 QC tools
• Technique
– It has a wider application than a tool.
– Needs more thought, skill and training to use
– It can be thought of as a collection of tools
– Eg., statistical process control (SPC) employs a
variety of tools such as charts, graphs and histograms
215
Why to use tools and
techniques?
• Plays a key role in a company-wide
approach to continuous improvement. They
allow:
– Processes to be monitored and evaluated;
– Everyone to become involved in the
improvement process;
– People to solve their own problems;
– A mindset of continuous improvement to be
developed
216
Requirement for tools and
techniques
• Requires attention to be paid to a number of
“critical success factors” to make their use and
application effective and efficient. Some of
these are:
– Full management support and commitment;
– Effective, timely and planned training;
– A genuine need to use the tool or technique;
– Defined aims and objective for use;
– A co-operative environment;
– Backup and support from improvement facilitators.
217
7 QC Tools – contd.
• Check sheets
– A systematic record
keeping or data
collection
– Observations are
recorded as they
happen and reveals
pattern or trends
218
7 QC Tools – Contd..
• Histograms
– Displays large amount
of data that are difficult
to interpret in raw
form
– Provides summary of
data and also reveals
whether the process is
centered, the degree of
variation etc.
– Used to identify
process capability
relative to customer
requirements
219
• Flow charts
– Shows the sequence
of events
– Mostly used in
manufacturing and
service operations
to describe working
procedures
– Valuable process
information can be
obtained in addition
to identifying
problematic areas
220
• Pareto Diagram
– helps to prioritize
the problem by
arranging them in
decreasing order of
importance
221
• Cause and Effect diagram
– Called as Ishikawa diagram for fish
bone diagram
– Explores possible causes of
problem due to men, machine,
material and method
– To identify root cause, each cause
may be further broken down
222
223
• Control charts
– Distinguishes special (assignable) causes of
variation from common causes
– Used to monitor and control a process on a
ongoing basis
– Plots a selected quality characteristic, found
from subgroups of observations, as a
function of sample number
– Central line on the control chart is the
average value of the characteristic
– Two limits, UCL and LCL are used to detect
whether the process goes out of control
224
225
• Scatter Plots
– Shows the
relationship
between two
variables
– Used as a follow
up to cause and
effect analysis to
find whether a
stated cause has
impact on quality
226
Chapter - 4
Fundamentals of Statistical
Concepts & Techniques in Quality
Control and Improvement
227
Basic Terminologies
• Population • Sample
– Set of all items that – A subset of population
possess a certain – Eg. Selecting 200 plastic
characteristic of interest cups from the week 23
– Eg. Average thickness of output
the plastic cups produced
• Statistic
in week no. 23 (10,000)
– A characteristic of a
• Parameter
sample, which is used to
– Is a characteristic of a make inferences on the
population, which population parameters
describes it that are unknown
– Eg. Average thickness of – Eg. Average thickness of
10,000 cups
200 plastic cups is 1mm
228
Basics of Probability
• Probability of an event describes the
chance of occurrence of that event
• A probability function is bound by 0 and
1
– 0 for non-occurrence, 1 for occurrence
• Set of all outcomes of an experiment is
called ‘sample space’ (S)
• If each outcome in sample space is likely
to happen, then the prob. of event A is
given by P(A) = na / N and probability
associated with sample space is P(S) = 1 229
Example
A company makes plastic storage bags for the

food industry. Out of the hourly production of
2000 500-g bags, 40 were found to be
nonconforming. If the inspector chooses a bag
randomly from the hour's production, what is
the probability of it being nonconforming?
230
Basics of Probability – Contd..
• Events
– Simple events cannot be broken into other events
– Compound events are made up of two or more
simple events
– Complementary of an event, say A, implies the
occurrence of everything except A. i.e. P(Ac) = 1 –
P(A)
• Laws
– Additive law defines the probability of the union
of 2 or more events (say A & B), i.e. implies A may
happen, B may happen or both
– P(A u B) = P(A) + P(B) – P(A n B)
231
Suppose that an inspector is sampling
transistors from an assembly line and
identifying them as acceptable or not.
Suppose the inspector chooses two
transistors. What are the simple events? Give
an example of a compound event. Find the
probability of finding at least one acceptable
transistor.
232
• Laws – contd..
– Multiplicative law defines the probability of the
intersection of 2 or more events (say A & B),
i.e. implies all the events in the group occurs
– P(A n B) = P(A).P(B | A) = P(B).P(A | B)
– P(B | A) represents conditional probability,
(i.e., probability that B occurs if A has)
• Independence
– Two events A & B are said to be independent, if
the outcome of one has no influence on
outcome of other
– P(B | A) = P(B) and hence P(A n B) = P(A).P(B)
233
• Mutually Exclusive
– Two events A & B are said to be mutually
exclusive, if they cannot happen
simultaneously.
– Probability of Intersection P(A n B) = 0
and probability of union P(A u B) = P(A)
+ P(B)
– For mutually exclusive, the events A & B
are dependent. If A & B are
independent, the additive rule will be
P(A or B or both) = P(A) + P(B) –
P(A).P(B) 234
In the production of metal plates for an assembly, it
is known from past experience that 5% of the plates
do not meet the length requirement. Also, from
historical records, 3% of the plates do not meet the
width requirement. Assume that there are no
dependencies between the processes that make the
length and those that trim the width. What is the
probability of producing a plate that meets both the
length and width requirements?
235
Thank You
236
Reliability

Pilani Campus
In the production of metal plates for an assembly, it
is known from past experience that 5% of the plates
do not meet the length requirement. Also, from
historical records, 3% of the plates do not meet the
width requirement. Assume that there are no
dependencies between the processes that make the
length and those that trim the width. What is the
probability of producing a plate that meets both the
length and width requirements?
238
• What proportion of the parts will not meet
at least one of the requirements?
• What proportion of parts will meet neither
length nor width requirements?
239
Suppose the operations that produce the length and
the width are not independent. If the length does not
satisfy the requirement, it causes an improper
positioning of the part during the width trimming
and thereby increases the chances of nonconforming
width. From experience, it is estimated that if the
length does not conform to the requirement, the
chance of producing nonconforming widths is 60%.
Find the proportion of parts that will neither
conform to the length nor the width requirements.
240
Statistics
• Statistics is the science that deals with the collection,
classification, analysis and making of inferences from
data
• Descriptive Statistics
– Describes the characteristics of product or process
using information collected on it
• Inferential Statistics
– Draws conclusion on unknown process parameters
based on sample information
• Data Collection
– Direct observation
– Indirect observation (Questionnaires)
• No control over data and Chances of Error is more
– It can be described by random variable – continuous 241
or discrete BITS Pilani, Pilani Campus
Statistics – Contd..
• Continuous variable
– Variable that can assume any value on a continuous scale
within a ‘range’ Eg. Viscosity of a resin
• Discrete variable
– Variable that can assume a finite number of values are called
discrete Eg. No. of defect in a shirt
– They are classified as acceptable or not
– Continuous characteristic can also be viewed as discrete. Eg.
Diameter of a hub in a tire
• Accuracy
– Refers to the degree of uniformity of the observations around
a desired value, such that on average, target is realized
• Precision
– Refers to the degree of variability of observation
242
Measures of Scale
• Nominal Scale
– Data variables are simply labels to identify an attribute
– Eg. Critical / Major / Minor
• Ordinal Scale
– Data has the properties of nominal data
– Data ranks or orders the observation
– Eg. Grades, 1- outstanding, 5 – poor
• Interval Scale
– Data has the properties of ordinal data and a fixed unit of
measure describes the interval between observations
– Eg. Temp. of water in diff stages of cooling during 2 hrs
interval
• Ratio Scale
– Data has the properties of Interval data and a natural zero
exists for measurement scale Eg. Wt. of cement bag: 100
kg, 100.2kg. 243
Basic Statistical Concepts
www.phdcomics.com

Distribution
How excited are you about learning QCAR concepts?

1 2 3 4 5 6 7
Comatose Hyperventilating
1 2 2 3 4 4 5 6 7
7 Types: 1,2,3,4,5,6,7 9 Tokens: 1,2,2,3,4,4,5,6,7

Properties of a
Distribution
• Shape
– symmetric vs. skewed
– unimodal vs. multimodal
• Central Tendency
– where most of the data are
– mean, median, and mode
• Variability (spread)
– how similar the scores are
– range, variance, and standard deviation

Representing a
Distribution
• Often it is helpful to visually represent
distributions in various ways
• Graphs
– continuous variables (histogram, line graph)
– categorical variables (pie chart, bar chart)
• Tables
– frequency distribution table

Distribution
• What if we collected 500 observations

instead of only 9?
…
Distribution
N = 500
50
40
30
20
10
1 2 3 4 5 6 7

Central Tendency
• Mode: the most frequent score

– good for nominal scales (eye color)
– a must for multimodal distributions
• Median: the middle score

– separates the bottom 50% and the top 50%
of the distribution
– good for skewed distributions (net worth)

Central Tendency
• Mean: the arithmetic average

– add all of the scores and divide by total number
of scores
– This the preferred measure of central tendency
(takes all of the scores into account)
X X
 X
N n
population sample

Computing a Mean
10 scores: 8, 4, 5, 2, 9, 13, 3, 7, 8, 5
ξΧ = 64
ξΧ/n = 6.4

Central Tendency
• Is the mean always the best measure of

central tendency?
• No, skew pulls the mean in the direction of

the skew

Central Tendency and Skew
Mode
Median
Mean
Central Tendency and Skew
Mode
Median
Mean

Distribution
• So, central tendency simplifies and

describes our distribution by providing a
representative score
• What about the difference between the

individual scores and the mean?
(variability)

Variability
• Range: maximum value – minimum value

– only takes two scores from the distribution into
account
– easily influenced by extreme high or low scores
• Standard Deviation/Variance
– the average deviation of scores from the mean of the
distribution
– takes all scores into account
– less influenced by extreme values

Standard Deviation
• most popular and important measure of

variability
• a measure of how far all of the individual
scores in the distribution are from a
standard (mean)

Standard Deviation
m ean m ean
low variability high variability
small SD large SD

Computing a Standard Deviation
10 scores: 8, 4, 5, 2, 9, 13, 3, 7, 8, 5 ξΧ/n = 6.4
8 – 6.4 = 1.6 2.56

SS = 96.4
4 – 6.4 = - 2.4 5.76
5 – 6.4 = - 1.4 1.96
2 – 6.4 = - 4.4 19.36 variance = 2 = SS/N
9 – 6.4 = 2.6 6.76
10.71
13 – 6.4 = 6.6 43.56
3 – 6.4 = - 3.4 11.56
2  X   2
7 – 6.4 = 0.6 0.36 standard deviation =  =  

N
8 – 6.4 = 1.6 2.56 3.27
5 – 6.4 = - 1.4 1.96
 BITS Pilani, Pilani Campus

Standard Deviation
• In a perfectly symmetrical (i.e. normal)

distribution 2/3 of the scores will fall
within +/- 1 standard deviation
-1 +1
3.13 6.4 9.67

Variance vs. SD
• So, SD simplifies and describes the

distribution by providing a measure of
the variability of scores
• If we only ever report SD, then why
would variance be considered a separate
measure of variability?
• Variance will be an important value in
many calculations in inferential statistics

Measures of Central Tendency
• Mean
– Simple average of the observations in a dataset
– Sample Mean, Population Mean (formulae)
• Median
– Is the value in the middle, when observations are ranked
– It is more robust than mean, as it is not influenced by
extreme values in dataset
• Mode
– Is the value that occurs more frequently in a dataset
– A dataset having more than one mode is called Multi-
modal
• Trimmed Mean
– Obtained by calculating the mean of data, that remain
after a proportion of high and low values being deleted 263
(a% trimmed)
The time taken for car tune-ups (in minutes)
is observed for 20 randomly selected cars. The
data values are as follows:
15 10 12 20 16 18 30 14 16 15
18 40 20 19 17 15 22 20 19 22
find the 5% trimmed mean [i.e., T(0.05)]
264
Measures of Dispersion
• Provides information on the variability or scatter of

the observations around a given value
• Range
– Difference between largest and smallest value in the
dataset
– R = XL - XS
• Variance
– Measures the fluctuation of the observations around the
mean
– Population variance, Sample Variance (formulae)
• Why n-1 in sample variance?
– To satisfy the property of unbiasedness i.e. average of
sample variance (keeps varying between sample) should
be equal to population variance which is constant 265
Measures of Dispersion
• Standard Deviation
– Mostly used measure of dispersion and has the same unit as the
observation
– Measures the variability of the observation around the mean
– Population Standard Deviation, Sample Standard Deviation
• Inter Quartile Range
– Lower / First quartile (Q1) is the value such that 1/4th of the
observations fall below it and 3/4th fall above it (Q1 = 0.25 (n+1))
– Vice Versa for Third Quartile (Q3) (Q1 = 0.75 (n+1))
– Difference between 3rd quartile and 1st quartile (IQR = Q3 – Q1)
– Larger the value of IQR, greater the spread of data
– To find IQR, the data are ranked in ascending order and then Q1
and Q3 are calculated
266
A random sample of 10 observations of the
output voltage of transformers is taken. The
values (in volts, V) are as follows:
9.2, 8.9, 8.7, 9.5, 9.0, 9.3, 9.4, 9.5, 9.0, 9.1
Calculate the variance and standard deviation.
267
A random sample of 20 observations on the
welding time (in minutes) of an operation
gives the following values:
2.2 2.5 1.8 2.0 2.1 1.7 1.9 2.6 1.8 2.3
2.0 2.1 2.6 1.9 2.0 1.8 1.7 2.2 2.4 2.2
Calculate the IQR.
268
Thank You
269
Reliability

Pilani Campus
Measures of Skewness &
Kurtosis
• Skewness coefficient (V1)
– Describes the asymmetry of the dataset about the mean or
indicates the degree to which distribution deviates from
symmetry (formulae)
– Negatively skewed: V1= -ve, Mean < Median
– Positively skewed: V1= +ve, Mean > Median
– Not skewed: V1= 0, Mean = Median
• Kurtosis coefficient (V2)
– Is a measure of peakness of the dataset (formulae)
– Is also a measure of heaviness of the tails of distribution
– For normal distribution (Mesokurtic), V2 = 3
– Leptokurtic, More peaked, V2 > 3
– Platykurtic, Less peaked, V2 < 3
271
272
A sample of 50 coils to be used in an electrical circuit is
randomly selected, and the resistance of each is
measured (in ohms). Calculate the skewness coefficient
and kurtosis coefficient.
273
Probability Distribution
• Sample data can be described with histograms, while population data are
described by probability distribution
• For discrete random variables, the probability distribution shows the value
that the variable can take and their corresponding probabilities
– P(xi) ≥ 1 for all i, P(xi) = P(X = xi); i = 1, 2,…
– Sum of all P(xi) = 1
– Some examples of discrete random variables are the number of defects in
an assembly, the number of customers served over a period of time, and
the number of acceptable compressors.
• Continuous random variable can take a infinite number of values and hence
probability distribution is expressed by Mathematical function
– f(x) >= 0 for all x, where P(a ≤ x ≤ b) = b∫af(x)dx
– Integration from minus infinity to plus infinity is one
– Almost all variables for which numerical measurements can be obtained
are continuous in nature: for example, the length of a pin, the diameter of
a bolt, the tensile strength of a cable, or the specific gravity of a liquid.
274
Variables and Random Variables
Earlier, we saw that just like in algebra, we can use a variable to represent some
quantity, such as height.
i.e. It is just like a variable in
statistics, except each outcome has
now been assigned a probability.
1 2 3 4 5 6
0.3 0.2 0.1 0.25 0.05 0.1
“The probability
that…
275
Is it a discrete random variable?
The height of a person 

No 
Yes
randomly chosen. This is a continuous random
variable.
The number of cars that 

No 
Yes
pass in the next hour.
The number of countries in 

No 
Yes
the world. It does not vary, so is not a variable!
276
Probability Distributions vs Probability Functions
There are two ways to write the mapping from outcomes to probabilities.
The “{“ means we have a ‘piecewise

function’. This just simply means we choose
the function from a list depending on the
Probability Functions input.
Advantages of probability function:

Can have a rule/expression based on the outcome.
Particularly for continuous random variables (in S2), it
Probability Distribution
?
would be impossible to list the probability for every
outcome. More compact.
1 2 3 4
0.1 0.2 ?0.3 0.4
The table form that you know and love.
Advantages of distribution:
Probability for each outcome more explicit.
? 277
Example
Underlying Probability Distribution

Sample Space
0 1 2 3
{ HHH, ?
HHT,
HTT, Probability Function
HTH,
THH,?
THT,
TTH, ?
TTT }
278
Cumulative Distribution Function (CDF)
If X is the number of heads

thrown in 2 throws...
? 0 1 2
? ?
0.25 0.5? ?
0.25
0 1 2
0.25
? 0.75
? 1 ?
279
CDF
F(x)
p(x)
Shoe Size (x) Shoe Size (x)
It’s just like how we’d turn a frequency graph

into a cumulative frequency graph. 280
Cumulative Distributive function
• For a discrete random variable, F(x) = ∑ all i p(xi) for

xi ≤ x
• For a continuous random variable, F(x) = b∫a f(t)dt
• F(x) is a non decreasing function of x such that for
limit, x tending to infinity, it is 1 and it is 0 for
minus infinity
• Expected value
– µ = E(x) = ∑ all i xi p(xi) , if x is discrete
– µ = E(x) = b∫a x f(x)dx, if x is continuous
• Variance of a random variable is given by
– Var(X) = E[(X - µ)2] = E(X2) – [E(X)]2
281
Let X denote a random variable that represents
the number of defective solders in a printed
circuit board. The probability distribution of the
discrete random variable X may be given by
the mean or expected value E(X), variance and

standard deviation.
282
Discrete Distributions
• The discrete class of probability distributions deals with those random variables
that can take on a finite or countably infinite number of values.
• Hyper geometric distribution

– Useful in sampling from a ‘finite’ population without
replacement, where the outcomes are success or failure
– If we consider, getting a nonconforming item as success,
the probability distribution of nonconforming item (x)
is given by P(x) = Dcx . (N-D)c(n-x) / Ncn
• D: no. of defects in population, x: no. of defects in
sample
• N: Size of population, n: Size of sample
– Mean µ = E(x) = nD/N
– Variance σ2 = Var(x) = nD/N(1 – D/N)((N-D)/N-1))
283
A lot of 20 chips contains 5 nonconforming
ones. If an inspector randomly samples 4
items, find the probability of 3 nonconforming
chips.
284
• Binomial distribution
– Useful in sampling from a ‘large’ population
without replacement, or to sample with
replacement from a finite population
– Probability of success (p) on any trial is
assumed to be a constant
– Let x denote the no. of successes, if n trials are
conducted, probability of x successes is given
by
• P(x) = ncx . px (1-p)n-x , x = 0,1, 2..
– Mean µ = E(x) = np
– Variance σ2 = Var(x) = np(1 – p)
285
Example
286
A manufacturing process is estimated to
produce 5% nonconforming items. If a
random sample of five items is chosen, find
the probability of getting two nonconforming
items.
287
• Poisson distribution
– Used to model the no. of events that happen
within a product unit, space or volume or time
period. Eg. No. of machine breakdown per
month
– Probability distribution function of the no. of
events (x) is given by
• P(x) = e-λ . λx / x! , x = 0,1, 2..
– Mean or average no. of events is given by λ
– Mean µ = Variance σ2 = λ
– It is used as an approximation to the binomial,
when ‘n’ is large and ‘p’ is small, such that np = λ
is a constant or average no. of defects per unit is
constant 288
It is estimated that the average number of
surface defects in 20 m2 of paper produced by
a process is 3. What is the probability of
finding no more than 2 defects in 40 m2 of
paper through random selection?
289
Discrete vs Continuous Distributions
You all know the distinction between discrete and continuous variables:
• Discrete: hair colour, shoe size, IQ, cars passing in the next hour, …
• Continuous: height, weight, time, …
Random variables are discrete or continuous when the outcomes are discrete or
continuous.
Discrete Random Variables Continuous Random Variables
Examples: Poisson Distributed Examples: Normally Distributed

BinomiallyExamples
Distributed Examples
Discrete Uniform Distribution
General Properties ? General Properties ?

Let’s go into more detail on
290
this… BITS Pilani, Pilani Campus
Probability Density Function (p.d.f.) !
For continuous variables, it doesn’t

make sense to talk about the
frequency of a specific value, e.g.
no one will have an exact height of
1.7m. We would instead use
‘frequency density’.
Earlier we saw that for histograms,

the value of the frequency density is
not particularly useful in isolation.
But when we find the area under a
bar, we get the frequency.
291
Cumulative Distribution Function
This works in exactly the same way as a discrete distribution:
292
Continuous Distribution
• Normal Distribution
– Most widely used
– Probability density function of a normal random variable is
f(x) = 1/√2πσ exp[-(x - µ)2/ 2σ2], -∞ < x < ∞
– µ = population mean, σ = population std. deviation
– Change in mean changes the location of distribution. As µ
increases, distribution shifts right and vice versa
– As variance increases, the spread about mean increases
– Normal distribution is symmetric and Mean = Median = Mode
– Proportion of population values that fall in the range of µ +/-
σ is 68.26%, µ +/- 2σ is 95.44%, µ +/- 3σ is 99.74%,
– Shape of the density function changes for diff. values of µ & σ
and hence it needs standardization
293
294
Continuous Distribution –
Contd..
• Normal Distribution
– Area within certain limits can be found by
looking up tabulated values for a ‘Std. normal
distribution’
– Standardized normal random variable z = x - µ /
σ and z is the no. of standard deviations a raw
value x is from mean and has mean of 0 and
variance of 1
– Z-value can be positive or negative and at mean
it is zero
– Density function (formula)
– Cumulative density function (formula)
– Sample problem 295
The length of a machined part is known to have
a normal distribution with a mean of 100 mm
and a standard deviation of 2 mm.
• What proportion of the parts will be above 103.3 mm?
• What proportion of the output will be between 98.5 and
102.0 mm?
• What proportion of the parts will be shorter than 96.5 mm?
• It is important that not many of the parts exceed the desired
length. If a manager stipulates that no more than 5% of the
parts should be oversized, what specification limit should be
recommended?
296
The length of a machined part is known to have
a normal distribution with a mean of 100 mm
and a standard deviation of 2 mm.
• What proportion of the parts will be above 103.3 mm?
• What proportion of the output will be between 98.5 and
102.0 mm?
• What proportion of the parts will be shorter than 96.5 mm?
• It is important that not many of the parts exceed the desired
length. If a manager stipulates that no more than 5% of the
parts should be oversized, what specification limit should be
recommended?
297
298
299
Thank You
300
Reliability

Pilani Campus
Inferential Statistics
• Talks about statistical procedures to make

inferences about a population on the basis of
sample data
• 2 main procedures: Estimation and Hypothesis
testing
• Sampling Distribution
– An estimator or statistic is used to make inferences on its
corresponding population parameter
– Studying the behavior of estimator through repeated
sampling is used to draw conclusions about the parameter
– This behavior is called sampling distribution, expressed as
the probability distribution of statistic
• http://davidmlane.com/hyperstat/index.html
302
303
Inferential Statistics – Contd..
– Sample mean is most widely used estimator and
hence it is important to know the sampling
distribution of mean, which is given by Central
Limit Theorem (CLT)
– Suppose we have a population with mean ‘µ’ and
standard deviation ‘σ’. If random sample of size
‘n’ are selected from this, and if sample size is
large, following holds good: (CLT)
• Sampling distribution of the sample mean will be
approximately normal
• Mean of this sampling distribution of mean ‘µx’ will be
equal to population mean ‘µ’
• Std. deviation is σx = σ / √n 304
305
Inferential Statistics – Contd..
– The degree to which a sampling distribution of a
sample mean approximates a normal distribution
becomes greater as sample size ‘n’ becomes larger
– For population distribution that is symmetric and
uni-modal, sample size as small as 4 or 5 yield
sample means that are approximately normally
distributed
– Variability of the sample mean, as measured by
standard deviation decreases as sample size
increases
306
The tuft bind strength of a synthetic material
used to make carpets is known to have a mean
of 50 kg and a standard deviation of 10 kg. If a
sample of size 40 is randomly selected, what
is the probability that the sample mean will be
less than 52.5 kg?
307
Estimation of Product and Process
Parameters
• Can be done using Point Estimation and
Interval Estimation
308
Estimation – contd.
• Point Estimation
– Consists of a single numerical value that is used to
make an inference about an unknown product or
process parameter
– Eg. Estimate the length of a shaft produced in
certain month (25mm)
– A point estimator is said to be ‘unbiased’ if the
expected value or mean of sampling distribution
is equal to the parameter
– A point estimator is said to have minimum
variance, if its variance is smaller than that of any
other point estimator for the parameter under
consideration
309
Estimation – Contd..
• Interval Estimation
– A range of interval is determined, such that
there is some desired level of probability that
the true parameter value is contained within it
– Called as confidence interval
– Say 2 end points (L, U), probability of
parameter ‘γ’ being contained in the interval
is some value (1 – α)
– P(L ≤ γ ≤ U) = 1 – α, which represents a 2-
sided confidence interval
310
Estimation – Contd..
• Interval Estimation
– If a large number of such confidence intervals
were constructed from large no. of independent
samples, then 100(1 – α)% of these intervals
would be expected to contain the parameter
value ‘γ’
– It can also be one sided. An interval of the type
L <= γ, such that P(L <= γ) = 1 – α is a one sided
lower 100(1 – α)% confidence interval for γ
– Context of the situation influence the type of
confidence interval selected
311
312
Estimation of Parameters
• Confidence interval about the mean
(Unknown)
– Case 1: Variance Known (Population)
• Based on Central Limit Theorem, the
sampling distribution of this mean is
approximately normal. A 100(1 – α)%
two-sided confidence interval for µ is
given by
X  z / 2  n
 X  z
2

n
313
314
The output voltage of a power source is
known to have a standard deviation of 10 V.
Fifty readings are randomly selected, yielding
an average of 118 V. Find a 95% confidence
interval for the population mean voltage.
315
• Confidence interval about the mean
(Unknown)
– Case 2: Variance Unknown (Population)
• A random sample of size ‘n’ is chosen and
the sample mean X and sample variance is
calculated. The sampling distribution of
X - µ / (s / √n) follows ‘t’ distribution with
(n-1) degree of freedom, which is similar to
normal distribution.
s s
X  t( / 2),( n  1)   X  t( / 2),( n  1)
n n
• Sample problem
316
A new process has been developed that transforms ordinary
iron into a kind of super iron called metallic glass. This new
product is stronger than steel alloys and is much more
corrosion-resistant than steel. However, it has a tendency to
become brittle at high temperatures. It is desired to estimate
the mean temperature at which it becomes brittle. A random
sample of 20 pieces of metallic glass is selected. The
temperature at which brittleness is first detected is recorded
for each piece. The summary results give a sample mean X of
600 °C and a sample standard deviation s of 15 °C. Find a 90%
confidence interval for the mean temperature at which
metallic glass becomes brittle.
317
Estimation of Parameters –
Contd..
• Confidence interval for the difference

between two mean (Unknown)
– Case 1: Variances of two populations
Known
• ‘n1’ is the sample size from 1st population
with mean x1, while ‘n2’ and x2 is for 2nd
population. A 100(1 – α)% two-sided
confidence interval is given by
 12  22  12  22
 X1  X 2   z 2  1   2  X 1  X 2   Z 
n1 n2 2 n1 n2
318
Estimation of Parameters – Contd..
• Confidence interval for the difference
between two mean (Unknown)
– Case 2: Variances of two populations Unknown
• Two unknown variances are equal (σ12 = σ22)
1 1
( X 1  X 2 )  t / 2,(n1  n2  2) . s p .   
 n1 n2 
 (1   2 ) 
1 1
( X 1  X 2 )  t / 2,(n1  n2  2) . s p .   
 n1 n2 
 ( n  1) s 2
 ( n  1) s2 
2
– Sample problem s p 
2 1 1 2

 ( n1  n 
2 3192) 
Estimation of Parameters – Contd..
• Confidence interval for the difference between two
mean (Unknown)
– Case 2: Variances of two populations Unknown
• Two unknown variances are not equal (σ12 ≠ σ22)
 s12 s2 2 
( X 1  X 2 )  t / 2, .   
 n1 n2 
 (1   2 ) 
2
 s12 s2 2  s 2
s2 2
( X 1  X 2 )  t / 2, .    n n 
1
 n1 n2    12 2  2
ν = no. of degrees of freedom  s12   s2 2 
n  n 
 1   2 
n1  1 n2  1
320
Two operators perform the same machining operation. Their
supervisor wants to estimate the difference in the mean
machining times between them. No assumption can be made as
to whether the variabilities of machining time are the same for
both operators. It can be assumed, however, that the
distribution of machining times is normal for each operator. A
random sample of 10 from the first operator gives an average
machining time of 4.2 minutes with a standard deviation of 0.5
minutes. A random sample of 6 from the second operator yields
an average machining time of 5.1 minutes with a standard
deviation of 0.8 minutes. Find a 95% confidence interval for the
difference in the mean machining times between the two
operators.
321
Estimation of Parameters –
Contd..
• Confidence interval for a proportion
– Let ‘p’ the proportion of success in binomial
distribution i.e. ‘p’ is the proportion of non-
conforming items in a process or lot. ‘p*’ is the
sample proportion (p* = x / n). The confidence
interval is given by
ˆ 1  p
p ˆ ˆ 1  p
p ˆ
ˆ  z
p ˆ  z
p p
2 n 2 n
322
– Contd..
• Confidence interval for difference between 2
proportion
– A sample size of n1 selected with parameter p1
and a sample size of n2 selected with parameter
p(2pˆ  pˆ )  z  pˆ1 1  p
ˆ1  pˆ 2 1  p
ˆ2 
1 2    
2  n1 n2 
p1  p2 
ˆ1 1  p
 p ˆ1  pˆ 2 1  p
ˆ2 
ˆ1  p
(p ˆ 2 )  z   
2  n1 n 2 
323
– Contd..
• Confidence Interval for Variance
– A random variable x from a normal distribution with mean
µ and variance σ2 (both unknown)
– An estimator of σ2 is the sample variance s2
– Sampling distribution of s2 follows a ‘chi-square’
distribution with (n-1) degrees of freedom given by
(n-1)s2 / σ2 = χ2n-1
– A chi square distribution is skewed to the right and a
100%(1-n 1 s 2for populationvariance
α) interval n  1issgiven
2 by
 2 
2 / 2, n  1 12  / 2, n  1
– Sample problem
324
The time to process customer orders is known
to be normally distributed. A random sample
of 20 orders is selected. The average
processing time X is found to be 3.5 days with
a standard deviation s of 0.5 day. Find a 90%
confidence interval for the variance ó2 of the
order processing times.
325
– Contd..
• Confidence Interval for ratio of two variances
– Random variable x1, x2 from a normal distribution with mean
µ1, µ2 and variance σ12, σ22 (both unknown)
– Random sample of size n1, n2 are chosen and the variance was
found to be s12, s22
– The ratio of these 2 variances will follow an ‘F’ distribution
with (n1-1) degrees of freedom in numerator and (n2 – 1) d.o.f
in denominator s 2given
/ by2
1 1
~ F( n1  1),( n2  1)
s /
2
2 2
2
– A ‘F’ distribution is skewed to the right and a 100%(1-α)

interval for this ratio is given by 2
s 2
1 1 s1 2
.  2  2 F / 2 , v 2 , v1
2
s 2 F / 2 , v1 , v 2 2 s2 326
The chassis assembly time for a television set is observed for
two operators. A random sample of 10 assemblies from the
first operator gives an average assembly time of 22 minutes
with a standard deviation of 3.5 minutes. A random sample of
8 assemblies from the second operator gives an average
assembly time of 20.4 minutes with a standard deviation of
2.2 minutes. Find a 95% confidence interval for the ratio of
the variances of the operators‘ assembly times.
327
328
Thank You
329
Hypothesis Testing

Pilani Campus
Hypothesis Testing
• To determine whether claims on product or process

parameters are valid
• Based on sample data
• A sample statistic used to test hypothesis is known as
test statistic
• Rather than a point estimate, a standardized quantity
based on the point estimate is found and used as test
statistic
• If a population standard deviation is known, the
normalized value of the sample mean is the z-statistic
given by x 
Z 
 / n 331
Hypothesis Testing – Contd..
• If a population standard deviation is unknown,
the standardized value of the sample mean is
the t-statistic given by
x 
t 
s/ n
• 2 types of hypothesis
– Null Hypothesis (Ho): Represents the status quo or
the circumstance being tested
– Alternative Hypothesis (Ha): Represents what we
want to prove. It is formulated to contradict null
hypothesis
• The null hypothesis is assumed to be true unless 332
proven otherwise BITS Pilani, Pilani Campus
• For example, Mean diameter of the rod is 25mm and

we are interested in determining whether for the
week 25, the mean diameter differs from 25mm
– Null Hypothesis (Ho): Represents the status quo or the
circumstance being tested (µ = 25mm)
– Alternative Hypothesis (Ha): Represents what we want to
prove. It is formulated to contradict null hypothesis (µ 
25mm)
• If we want to test, mean diameter exceeds 25mm
– Null Hypothesis (Ho): Represents the status quo or the
circumstance being tested (µ  25mm)
– Alternative Hypothesis (Ha): Represents what we want to
prove. It is formulated to contradict null hypothesis (µ >
25mm)
333
• How to reject Null Hypothesis?
– To determine this, we need a cut-off point – i.e. how
small the sample mean (diameter) to be for us to
conclude that mean diameter is less than 25mm
– There is a critical value, in this case on the left tail,
such that if the sample mean falls below it, null
hypothesis is rejected. This is called ‘Rejection
Region’
– If the statistic does not fall in this region, we do not
have significant evidence to reject null hypothesis
– It is dependent on the ‘Level of significance’ i.e. the
rejection region is chosen such that if the null
hypothesis is true, the probability of test statistic
falling in that region is small.
– Level of significance is denoted by ‘α’ 334
• Hypothesis testing for mean

– Case 1: Variance known (Sample problem)
• Assumption: Sample size is large and population
distribution is normal
• Hypothesis
– Ho :  = 0 Ho :   0 Ho :   0
– Ha :   0 Ha :  > 0 Ha :  < 0
• Rejection region
– | z0 | > z/2, z0 > z  z0 < -z
• Test statistic: x  0
Z0 
/ n
335
• Hypothesis testing for mean

– Case 2: Variance unknown (Sample problem)
• Assumption: Sample size is large (n > 30) and
population distribution is normal, while sample
follows ‘t’ distribution
• Hypothesis
– Ho :  = 0 Ho :   0 Ho :   0
– Ha :   0 Ha :  > 0 Ha :  < 0
– | t0 | > t/2 , (n-1) t0 > t , n-1 to < -t, n-1
x  0
Test statistic  to 
s/ n
336
A new process has been developed that transforms
ordinary iron into a kind of super iron called metallic
glass. This new product is stronger than steel alloys and is
much more corrosion-resistant than steel. However, it has
a tendency to become brittle at high temperatures. It is
desired to estimate the mean temperature at which it
becomes brittle. A random sample of 20 pieces of metallic
glass is selected. The temperature at which brittleness is
first detected is recorded for each piece. The summary
results give a sample mean X of 600 °C and a sample
standard deviation s of 15 °C.
Determine whether this mean temperature exceeds 595
°C. A random sample of 20 is taken, yielding a sample
mean X of 600 °C and a sample standard deviation s of 15
°C. Use a level of significance a of 0.05.
337
338
• Hypothesis testing for the difference between two means

– Case 1: Variance known
• Assumption
– 2 samples are independent
– Population have the same variance and normally distributed
• Hypothesis
– Ho : 1 - 2 = 0 Ho : 1 - 2  0 H0 : 1 - 2  0
– Ha : 1 - 2  0 Ha : 1 - 2 > 0 Ha : 1 - 2 < 0
– | z0 | > z/2, z 0 > z z0 < -z
• Test statistic:
Z0 
 X 1  X 2   0
• Sample problem  12 / n1   22 / n2
339
The owner of a local logging operation wants to examine the
average unloading time of logs. Two methods are used for
unloading. A random sample of size 40 for the first method
gives an average unloading time X\ of 20.5 minutes. A
random sample of size 50 for the second method yields an
average unloading time X2 of 17.6 minutes. We know that the
variance of the unloading times using the first method is 3,
while that for the second method is 4. At a significance level a
of 0.05, can we conclude that there is a difference in the mean
unloading times for the two methods?
340
• Hypothesis testing for the difference between two
means
– Case 2: Variance unknown
• Assumption
– Population have the same variance and normally distributed
• Hypothesis
– Ho : 1 - 2 = 0 Ho : 1 - 2  0 H0 : 1 - 2  0
– Ha : 1 - 2  0 Ha : 1 - 2 > 0 Ha : 1 - 2 < 0
– | t0 | > t/2, (n1 + n2 –2) t0 > t, (n1 + n2 –2) t0 < -t, (n1 + n2 –2)
• Test statistic:
t0 
 X1  X 2   0 
s p 
2 ( n1  1) s 1
2
 ( n2  1) s 2
2

( n  n  2) 
1 1  1 2 
sP 
n1 n2 341
A large corporation is interested in determining whether the average
days of sick leave taken annually is more for night-shift employees
than for day-shift employees. It is assumed that the distribution of
the days of sick leave is normal for both shifts and that the
variances of sick leave taken are equal for both shifts. A random
sample of 12 employees from the night shift yields an average sick
leave X\ of 16.4 days with a standard deviation s, of 2.2 days. A
random sample of 15 employees from the day shift yields an
average sick leave Xi of 12.3 days with a standard deviation s2 of
3.5 days. At a level of significance a of 0.05, can we conclude that
the average sick leave for the night shift exceeds that in the day
shift?
342
Solution

• Hypothesis testing for the difference between two means
– Case 3: Variance unknown
– Population have different variance and normally distributed
• Hypothesis
– Ho : 1 - 2 = 0 Ho : 1 - 2  0 H0 : 1 - 2  0
– Ha : 1 - 2  0 Ha : 1 - 2 > 0 Ha : 1 - 2 < 0
– | t0 | > t/2, γ t0 > t, γ t0 < -t, γ
• Test statistic: 2
s 2
s2  2
t0
X 
 1
X 2   0  n
1

n 
s12 s2 2   12 2  2
  s12   s2 2 
n1 n2  n   n 
 1   2 
• Sample problem n1  1 n2  1
344
• Hypothesis testing for proportion

• Assumptions
– No. of trials ‘n’ is a binomial experiment is large, that np  5
• Hypothesis
– Ho : p = po Ho : p  po Ho : p  po
– Ha : p  po Ha : p > po Ha : p < po
– | z0 | > z/2 z0 > z  z0 < -z
• Test statistic: pˆ  po
Zo 
po 1  po 
n
• Sample problem
345
The timeliness with which due dates are met is
an important factor in maintaining customer
satisfaction. A medium-sized organization
wants to test whether the proportion of times
that it does not meet due dates is less than 6%.
Based on a random sample of 100 customer
orders, they found that they missed the due
date five times. What is your conclusion? Test
at a level of significance a of 0.05.
346
Problem

• Hypothesis testing for the difference between two
binomial proportions
• Assumptions
– Difference between the 2 proportions is zero
• Hypothesis
– Ho : p1 - p2 = 0 Ho : p1 - p2  0 Ho : p1 - p2  0
– Ha : p1 - p2  0 Ha : p1 - p2 > 0 Ha : p1 - p2 < 0
– | z0 | > z/2 z0 > z  z0 < -z
• Test statistic:
pˆ1  pˆ 2
Zo   n1 pˆ1  n2 pˆ 2 
1 1
pˆ 1  pˆ    pˆ  
 n1 n2  n
 1 2   n
• Sample problem 348
A company is interested in determining whether the
proportion of nonconforming items is different for two of its
vendors. A random sample of 100 items from the first vendor
revealed 4 nonconforming items. A random sample of 200
items from the second vendor showed 10 nonconforming
items. What is your conclusion? Test at a level of significance
a of 0.05.
349
Problem

• Hypothesis testing for the variance

•
Hypothesis H o : 2  02 H o :  2  02 H o :  2  02
H a :  2  02 H a :  2   02 H a :  2   02
 02  2 / 2,n  1 or  02  12   / 2,n  1  02   2 , n  1  02   21  , n  1
• Test statistic:
 n  1 s 2
 02 
 02
• Sample problem
351
The variability of the time to be admitted in a
health care facility is of concern. A random
sample of 15 patients shows a mean time to
admission X^ of 2.2 hours with a standard
deviation s of 0.2 hours. Can we conclude that
the variance of time to admission is less is
than 0.06? Use a level of significance a of 0.01.
352
Problem

• Hypothesis testing for the ratio of 2 variances

• Hypothesis
H o :  12  22 H o : 12  22 H o :  12 12
H a :  12  22 H a :  12   22 H a :  12  22
• Rejection Region
F0  F / 2, 1 , 2 or F0  F1  / 2, 1, 2 F0  F ,  1, 2 F0  F1  , 1, 2
• Test statistic: s12
F0  2
s2
• Sample problem
354
The variabilities of the service times of two bank tellers are of
interest. Their supervisor wants to determine whether the
variance of service time for the first teller is greater than that
for the second. A random sample of 8 observations from the
first teller yields a sample average X1 of 3.4 minutes with a
standard deviation s1 of 1.8 minutes. A random sample of 10
observations from the second teller yields a sample average
X2 of 2.5 minutes with a standard deviation of 0.9 minutes.
Can we conclude that the variance of the service time is
greater for the first teller than for the second? Use a level of
significance a of 0.05.
355
Thank You
357
Reliability

Pilani Campus
Chapter - 6
Statistical Process Control using

Control Charts
359
Introduction to Control Charts
• Emphasis is on process control and improvement

• A Control chart is a
– Graphical tool
– For monitoring the activity of an ongoing process
– It is called as ‘Shewhart’ chart
– Values of quality characteristic is plotted along ‘ordinate’
– Sample number or subgroups (in order of time) is plotted
along ‘abscissa’
– Example for quality characteristic includes Average length,
Average tensile strength, Average service time etc.
– Quality characteristic can be categorized as
• Variable: Numerical values can be obtained
• Attribute: No numerical values, only conforming or non

conforming
360

Control Chart – Contd..
361
More about Control Charts –
Contd..
• A Control chart has
– A centre line, represents the average value of the characteristic
– It indicates, where the process is centred
– UCL and LCL, used to make decision regarding the process
– If the points plot within the control limits and do not exhibit
any identifiable pattern, then process is in statistical control
– If a point plots outside the control limits or if an identifiable
random pattern exists process is out of statistical control
– Benefits
• Indicates when something may be wrong, so that corrective
action can be taken
• Pattern of plot can help in diagnosing the possible causes and
hence possible remedial action
• Helps in estimating the process capability of the process
362

Causes of Variation
• Variability is part of any process
• Causes can be men, methods, machine, materials, environment
• Can be categorized in to two
• Special causes • Common causes
– Assignable causes or – Chance causes or
sporadic problem chronic problem
– Not inherent in the – Inherent in the process
process – Affects all items, as it is
part of process design
– Will not affect all items
– Exists as long as the
– Can be due to use of a process is not changed
wrong tool, improper – Can be due to vibration
raw material, operator of a machine, fluctuation
error etc. etc.
363
Causes of Variation – Contd..
• Special causes • Common causes

– If an observation falls – Variation is the effect of
outside the control limits or many small causes and
a non-random pattern exists,
special causes exists
cannot be totally
eliminated
– Control chart is used to
detect the presence of – When this variation is
special causes and should be constant, called as stable
eliminated system of common causes
– 15% of all problems are due – Management is alone
to special causes responsible for common
– Action is required from both causes
management and workers – 85% of all problems are
due to common causes
364
Statistical Basis for Control
Charts
• Centre line is based on the mean of a process
and is found by taking the average of the values
in sample
• Can also be desirable target or standard value
• Role of Normal distribution
– Value of the statistic plotted on a control chart are
assumed to have an approximately normal
distribution
– Any sample from population distribution that is
uni-modal and symmetric, the central limit theorem
states that if the plotted statistic is a sample
average, it tend to have a normal distribution
365

Control Chart – Contd..
366
Statistical Basis – Contd..
• Why 3σ limits?
– Chosen in such a way that probability of the sample points
falling between them is ‘1’, if the process is in statistical
control
– Normal distribution theory states that a sample statistic will
fall within the limits 99.74% of the time, if the process is in
control
• Most common basis for deciding whether a process is out
of control is “the presence of a sample statistic outside the
control limits” and also it depends on other rules
• A control chart is a means of online process control
• If the control limits are calculated from current data, then
it tells whether the process is presently in control or not
• If the control limits are calculated from previous data,
based on a process that was in control, it helps to find
whether the process has drifted out of control
• Control chart can also estimate process parameters such
as mean, standard deviation and fall outs 367

Statistical Basis for Control
Charts – Contd..
• It can be used to determine the capability of process
i.e, ability of the process to produce within desirable
specifications
– Process capability studies help in making major
management decisions like make or buy, investment in
machines, vendor selection etc.
• Issues in construction of control chart are:
– Selection of control limits, No. of items in the sample,
Frequency of sampling, Minimizing error in inferences,
Interpretation of control charts etc.
368

Selection of Control Limits
• Let ‘θ’ is the mean diameter of parts produced by the

process and θ* would be sample mean diameter of parts
chosen from the process
– E(θ*) is the mean or expected value and σ(θ*) be the standard
deviation of the estimator θ*
• Then the centre line and control limits are given by
• CL = E(θ*)
• UCL = E(θ*) + k σ(θ*)
• LCL = E(θ*) - k σ(θ*)
– Where k = no. of standard deviations of the
sample statistic that the control limits are
placed from the centre line
– Generally value of k is chosen as 3
369

Selection of Control Limits –
Contd..
• Sample statistic is assumed to have an normal
distribution
• Value of k = 3 implies that there is a probability of
only 0.0026 of a sample statistic falling outside the
control limits, if the process in in control
• k is also chosen, based on the desired probability of
sample statistic falling outside the control limits,
when process is in control. These limits are called
“Probability limits”
• Choice of k is influenced by the Error consideration
(Type I error and Type II error)
370

Errors in making inferences
• Making inferences from a control chart is similar to testing a

hypothesis
• Control limits are the critical points that separate the rejection and
acceptance region
• If a sample values fall above UCL or below LCL, then we reject null
hypothesis or say that process is out of control
• Type I error
– Inferring that a process is out of control, when it is actually in control
– Probability of Type I error is given by ‘α’
• Type II error
– Inferring that a process is in control, when it is really out of control
– All points may fall within limit, but process mean has changed
371

Example
A control chart is to be constructed for the average breaking

strength of nylon fibers. Samples of size 5 are randomly chosen
from the process. The process mean and standard deviation are
estimated to be 120 kg and 8 kg, respectively.
(a) If the control limits are placed 3 standard deviations from
the process mean, what is the probability of a type I error?
(b) If the process mean shifts to 125 kg, what is the probability
of concluding that the process is in control and hence making a
type II error on the first sample plotted after the shift?
(c) What is the probability of detecting the shift by the second
sample plotted after the shift if the samples are chosen
independently?
372

373
374
375
376
(1 -0.0548)(0.0548) = 0.0518 0.0548 + 0.0518 = 0.1066.
377
Operating Characteristics
curve (OC Curve)
• An OC curve is a measure of goodness of a control chart’s
ability to detect changes in process parameters
• A plot between the probability of the Type II error versus the
shifting of a process parameter value from its in-control value
• Helps us to determine the chances of not detecting a shift of a
certain magnitude in a process parameter on a control chart
• Shape of the curve is similar to inverted ‘S’
• For small shifts in the process mean, the probability of non-
detection is high
• If the change in process mean is high, the probability of
detection increases
• The ability of the control chart to detect changes quickly is
indicated by the steepness of the OC-curve and the quickness
with which the probability of non-detection reaches zero
378

OC Curve
379

Effect of control limits on
errors in inference making
• Choice of the control limits influence the occurrence of
Type I error and Type II error
• As control limit are placed farther apart, the probability
of Type I error decreases
– For 3σ limit, the probability of Type I error is 0.0026
while for 2σ limit, it is 0.0124
• But probability of making Type II error increases,
showing that both these errors are inversely related
• If all other process parameters held fixed, probability of
Type II error decreases with an increase in sample size
– As n increases, standard deviation decreases and hence
control limit will be nearer, and reduces probability of
Type II error.
380

Effect of control limits on
errors in inference making –
Contd..
• It involves making a judicious choice of control limit
selection
• Generally, control limits are placed at 3σ distance from
centre line, restricting the occurrence of ‘false alarm’
• For critical process, it is often necessary to detect small
changes as early as possible, hence tighter control limits
are preferable
• Warning limits
– Are usually placed at 2σ limits from the centre line
– When a point falls outside the warning limits, but within
control limits, process ‘may’ go out of control
– For a normally distributed statistic, the probability of falling in
a band between the warning and control limit is 4.3%
381

Effect of sample size on
control limits
• Sample size affects the standard deviation of the sample statistic
plotted
• In a control chart, for the sample mean standard deviation is given by
σx = σ / √n
• Standard deviation and sample size are inversely related
• n increases, σ decreases, i.e. it provides more information and hence
less variability, which reduces the frequency of inference error
occurrence
• Average run length
– An alternate measure of performance of control chart
– Denotes the no. of samples, on an average is required to detect the out
of control signal
– Let Pd denote the probability of observation plotting outside the
control limit, then run length is 1, with probability of P d and run length
is 2, with probability of Pd(1 – Pd)
– ARL = Pd * ∑ j (1–Pd) j – 1 = 1 / Pd (Based on expansion of infinite series)
382

Average Run Length –
Contd..
• For a process in control, Pd = α
• For a 3σ control chart, detection of out of control condition is given
by ARL = 1 / Pd = 1 / 0.0026 = 385, i.e. an observation will plot
outside the control limit every 385 samples on average
• For a process in control, the ARL to be large, so that observation
plotting outside control limit is less or it does not create a false
alarm
• For an out of control process, the ARL should be small, so that out
of control can be easily detected. In this case, Pd = 1 – β and hence
ARL = 1 / 1 – β
• ARL curves can also be constructed for the control chart based on
the shift in mean with respect to ‘σ’ and ARL
• ARL can be expressed in terms of expected number of individual
unit sampled I. It is given by I = n (ARL)
383

Selection of Rational samples
• Shewhart described the fundamental criteria for the

selection of rational sub group
• Rational sample is chosen in such a manner that the
variations within sample is considered to be due to
common causes alone
• Samples are selected such that special causes if present,
will occur between samples
• i.e. the difference between sample will be ‘maximized’
while difference within sample is ‘minimized’
• It can be done by the instant-of-time method i.e.,
Observation are selected at approximately the same time
for the population under condition
384

Selection of Rational samples – Contd..
• Sample size
– It is dependent on the expected degree of shift in the
process parameter
– For detecting slight change in the process parameter, a
large sample size is needed
– If we can tolerate smaller changes in the process
parameters, then a small sample size might suffice
• Frequency of Sampling
– It must be decided before constructing control charts
– Options are: choosing small sample size at frequent
intervals or choosing large sample size at infrequent
intervals
– Generally small sample size at frequent intervals is
adopted
385

Selection of Rational samples –
Contd..
• Other factors that influence sample size
and frequency are
– Type of inspection for getting data –
destructive or non-destructive
– Current state of process
– Cost of sampling and inspection per unit
– Loss incurred due to non-conforming
items etc.
386

Analysis of patterns in control
chart
• Criteria other than a plotted point falling outside the
control limits are also used to determine whether a process
is out of control
• Plot patterns often indicate whether the process is in
control or not
• Analysing these patterns is more difficult than plotting the
chart
• Identifying the causes of non-random patterns require a
knowledge of the process, equipment and operation
conditions as well as their impact of characteristic of
interest
387

ARL
388

Rules for Control chart
389

Analysis of control chart –
contd..
• Rules
– A process is assumed to be out of control, if a single point plots outside
the control limits
390

contd..
• Rules – contd.
– A process is assumed to be out of control, if two out of three
consecutive points fall outside the 2σ warning limits on the same side
of the centre line
391

contd..
– A process is assumed to be out of control, if four out of five consecutive
points fall beyond the 1σ unit on the same side of the centre line
392

contd..
– A process is assumed to be out of control, if 9 or more consecutive
points fall to one side of the centre line
393

Analysis of control chart – contd..
– A process is assumed to be out of control, if there is a run of
six or more consecutive points, steadily increasing or
decreasing
394

contd..
– For a process in control, roughly equal number
of points should be above or below the centre
line with no systematic patterns
395

Interpretation of plots
• Five rules for determining out of control conditions are not
all used simultaneously
• Rule 1 is routinely used along with a couple of other rules
• Reason for not using all rules is that it increases the chance
of Type – I error, i.e., Probability of false alarm increases
• Suppose a number of independent rules are used for out of
control criteria is ‘k’. Let αi be the probability of type I error
of rule ‘I’.
• Then the overall probability of type I error is α = 1 – product
(1 – αi)
• This is under the assumption that rules are independent, but
generally rules are not independent and hence it is
approximate only
• Even if all rules are applied, the process may be out of
control, for which the experience, judgement and
interpretive skills are needed
396

Other important considerations
• Determination of causes of out of control points

– Task of control chart is not just identification of out-of control points,
but should identify the causes behind that
– Requires a thorough knowledge of the process and also the sensitivity
of the output quality characteristic to process parameter
– Determination of causes is a team effort
– Generally cause and effect diagram is used
• Maintenance of control chart
– Should be done on continuous basis, as it involves changing the centre
line and limits and when process goes out of control
– Should be easily accessible to the person, who is responsible for that
quality characteristic
– Should be placed in shop floor so that all can know about the current
situation
397

Thank You
398
Reliability

Pilani Campus
The diameter of cotter pins produced by an automatic machine is a characteristic of interest.
Based on historical data, the process average diameter is 15 mm with a process standard
deviation of 0.8 mm. If samples of size 4 are randomly selected from the process:
(a) Find the 1ó and 2ó control limits.
(b) Find the 3ó control limits for the average diameter.
(c) What is the probability of a false alarm?
(d) If the process mean shifts to 14.5 mm, what is the probability of not detecting this
shift on the first sample plotted after the shift? What is the ARL?
(e) What is the probability of failing to detect the shift by the second sample plotted
after the shift?
(f) Construct the OC curve for this control chart.
(g) Construct the ARL curve for this control chart.
400

401
402
403
404
The length of industrial filters is a quality characteristic of interest. Thirty samples, each of
size 5, are chosen from the process. The data yields an average length of 110 mm, with the
process standard deviation estimated to be 4 mm.
(a) Find the warning limits for a control chart for the average length.
(b) Find the 3ó control limits. What is the probability of a type I error?
(c) If the process mean shifts to 112 mm, what are the chances of detecting this shift by the
third sample drawn after the shift?
(d) What is the chance of detecting the shift for the first time on the second sample point
drawn after the shift?
(e) What is the ARL for a shift in the process mean to 112 mm? How many samples, on
average, would it take to detect a change in the process mean to 116 mm?
405

406
407
408
A health care facility is monitoring daily expenditures for a certain diagnosis-
related group (DRG). Individual observations are selected. After 50 samples, the
average and standard deviation of daily expenditures (in hundreds of dollars) are
estimated to be 15 and 2, respectively.
(a) Find the 3ó control limits.
(b) Suppose that Rules 1 and 2 are used simultaneously for the detection of outof-
control conditions. Assuming independence of the rules, what is the overall
probability of a type I error? Explain the meaning of a type I error in this context.
(c) Suppose that the average daily expenditures for the same DRG increases to
$1750. What is the chance of detecting this shift by the second sample drawn after
the shift?
(d) What is the ARL?
409

BITS Pilani
Pilani Campus
Control Charts for Variables
410
Introduction
• Variables are those quality characteristics that are measurable

on a numerical scale, e.g. length, viscosity
• It is necessary to control the mean value of a quality
characteristic as well as its variability
• Mean gives an indication of the central tendency of a process
• Variability provides an idea of the process dispersion
• When there is a change in process mean or process standard
deviation, the proportion of parts that do not meet specification
increases
• Control chart aid in detecting such changes in process parameters
• Variables provide more information than attributes
• Attributes deal with qualitative information, whether an item is
non-conforming or not ok, but didn’t show the degree to which a
quality characteristic is non conforming
• Eg. 40 ± 0.5. Both 40.6 and 42 are nonconforming
411
Introduction – Contd..
• Cost of obtaining variable data is usually higher than for

attributes because
– Attribute data is collected by means such as go / no-go gages,
which are easier to use and hence less costly also
– Total cost of data collection is the sum of two costs – fixed and
variable cost
– Fixed cost include cost of inspection equipment, while variable
cost include cost of inspecting
– More the parts inspected, higher the variable costs and fixed
cost is unaffected
– With the use of automated devices for measurement, the
differences in variable unit cost between variable and attribute
may not be much, but fixed cost may increase
– Attribute chart for the proportion of non-conforming may
represent the general operational level of the plant but not the
variable chart
412
Selection of Characteristics of
Investigation
• A single component has usually several quality
characteristic
• Number of quality characteristic that affect a product
will be large
• Not feasible to maintain control chart for all as decision-
making becomes difficult
• Involves selecting a few vital characteristics and
selection is based on those that cause more non-
conforming items and increase cost
• Pareto analysis plays a major role in selection
• It is desirable to measure process characteristic that
have a casual relationship to product characteristic
413
Preliminary decisions
• Selection of Rational samples

– Sampling method should maximize differences
between samples and minimize differences within
sample
– Lot from which samples are chosen should be
homogeneous
– For detecting shift in process parameter, samples
should be made up of item produced at same time
– For non-conformance of items produced since the
previous sample, then samples should be selected
from items produced since that time
414
Preliminary decisions – contd..
• Sample size
– Size is normally between 4 and 10 and in industry it
will be 4 to 5
– Larger the sample size, better the chance of detecting
small shifts
– Based on factors like cost of inspection or cost of
shipping a nonconforming item to the customer etc.
• Frequency of sampling
– Depends on the cost of obtaining information
compared to the cost of non detecting a non
conforming item
– As process is brought to control, frequency of
sampling decreases
415
Preliminary decisions – Contd..
• Choice of measuring instruments

– Accuracy of the instruments directly influences the
quality of the data calculated
– Should be calibrated and tested
– Characteristic being controlled, the desired degree of
measurement precision, both impacts the selection
• Design of data recording forms
– Recording forms should be designed based on the
control chart to be used
– Includes sample no., date, time, raw values of data, part
name, lot number, operator name, machine, gauge, unit
of measurement, specification etc.
416
Control chart format
417
Control chart for Mean and Range
• Development of chart
– Using a pre-selected scheme and sample size record
measurements of the selected quality characteristic
– For each sample, calculate the sample mean and range
Ri = xmax - xmin x1  x 2    x n
x
n
– Obtain and draw the centre line and trial control limit
• Find the mean of all sample mean (Formula)
• Find the mean range of all samples (Formula)
– 3σ control limits for Mean chart is given by
 ˆ
 Z / 2 x  Z / 2 X  3 X  X  3
n n
– For normally distributed population, the distribution of the
statistic’s relative range (w) = R / σ and it is dependent on
sample size ‘n’
418
Control Chart for Mean (Xbar)
419
Control chart for Mean and Range –
Contd..
– Mean of w is given by d2
– Estimate of the process standard deviation is
ˆ R /d 2
R
– (UCL, LCL) = X 3  X A2 R
d2 . n
R 3 R
– Control limits for R chart is given by
– R = σ.w and hence σR = σ. σw
– Mean of σ
wˆ is
Rgiven
/d 2 as d3 and we Rknow that
also
 R  .d 3
Hence d2
D3 R D4 R
– UCL = LCL =
420
Control Chart for Range (R)
421
Thank You
422
Reliability

Pilani Campus
BITS Pilani
Pilani Campus
Control Charts for Variables
424
Contd..
• Plot the values of the range on the chart and find
whether points are in statistical control
– An R chart is analysed before X-bar chart to determine
out of control situations, as R chart reflect process
variability, which should be brought to control.
– If R chart shows out of control, then the X-bar chart is
meaningless
• Delete the out of control points for which remedial
action has been taken to remove special causes and
the remaining samples are used to obtain revised
limits
• A point of interest is about the point that falls
below the LCL, when LCL is greater than zero
425
Contd..
• These points are desirable because they indicate
unusually small variability, within the sample
and might be due to special causes
• This condition helps in further reducing our
process variability
• Implement the control chart
Why two charts?
• X bar chart monitors the mean between sample
values
• R chart monitors the variation within sample.
426
Example
427
Example – contd..
428
Example – contd..
429
Example – contd..
In this case, D3 = 0 for n = 5

and hence LCL = 0
430
Example – contd..
431
Example – contd..
432
433
Standardized control chart
• When the sample size varies, it results in fluctuating

control limits and hence we need standardized control
chart
• Standardized values represent a deviation from the
mean in units of standard deviation
• They are dimensionless and have a mean of zero
• Control limits for this are ±3 and are constant
• Easier to interpret the shift in the process
• Sample size for sample i is ni and let Xi* and si denotes
g
its mean and standard deviation
 ni X i
• Mean of sample average X  i 1g
n i
434i 1
Standardized control chart – Contd..
• An estimate of the process standard deviation σ* is

given by g
 (n
2
i  1) si
ˆ  i 1
g
 (n
i 1
i  1)
• For sample i, standardized value for the mean zi is
Xi  Xi
zi 
ˆ
ni
• A plot of the Zi values on a control chart with the centre
line at ‘0’ , UCL at 3 and LCL at -3 is standardized X-bar
chart
435
Standardized control chart – Contd..
– To standardize a Range chart, the range Ri for sample ‘i’

is first divided by the estimate
Ri of the process standard
deviation ri 
ˆ
ri  d 2
ki 
– Standardized value of Range is given by d3
– Value of ki are plotted with a centre line at ‘0’ and UCL =

‘3’ and LCL = ‘-3’
• Control Limits for a given target or Standard
– Management may specify values for the process mean or
standard deviation
– These values are the goals or desirable standards or
target values, which helps to determine process able to
meet standard 436
Control limits for a given target or
Standard - Contd..
– Let X o and  o represent the target values of the
process mean and standard deviation respectively
and the centre line and limits are given by
CL X X 0
0
UCLX X 0  3 .  X 0  A . 0
n
0
LCL X X 0  3 .  X 0  A . 0
n
– For R - chart
R
CLR d 2 . 0 (ˆ  )
d2
UCLR R  3 . R  d 2 . 0  3 . d 3 0 D2 . 0
LCLR R  3 . R  d 2 . 0  3 . d 3 0 D1. 0
437
Interpretation and Inferences
• Caution
– Sample plots may fall outside the limit, even
though no special causes are present
– Reason being that desirable standards may
not be consistent with the process conditions
– It is easy to meet a desirable target value for
process mean than it is for process
variability (Range)
438
Interpretation and Inferences –
Contd..
• Interpretation
– Difficult and needs thorough knowledge
about different process parameters on
quality characteristic
– When R-chart is brought to control, many
special causes for the Xbar chart are
eliminated as well
– Xbar chart monitors the centering of the
process and a jump indicates process
average has jumped
439
Interpretation and Inferences – Contd..
– Increasing trend indicates the process centre is

gradually increasing and requires adjustments in
machine settings or such controllable parameters as
proper tool, proper depth of cut, feed etc.
– Reducing process variability in order to allow an R-
chart to exhibit control is difficult and needs “quality
improvement”
– Process capability can be estimated from the process
standard deviation, and helps to determine how the
process performs with respect to specification limits
and also in determining the proportion of output that
is non conforming
440
Control chart patterns and corrective
action
• A ‘non-random identifiable’ pattern in the
plot of a chart might provide reason to look
for special cause in a process
• There are about 15 typical patterns
identified by Western Electric company and
9 of them have been discussed here
• Natural Patterns
– No identifiable arrangement of the plotted point
exists
– No point fall outside the control limits
– Majority of the points are near the centre line and
few points close to control limits
– Demonstrates the presence of stable system of
common causes 441
Control chart patterns – contd..
442
Control chart patterns – Contd.
• Sudden shifts in the level

– Can occur because of changes intentional or
otherwise in process settings (temperature, depth of
cut etc.)
– New operators, new equipment, new vendors, new
methods are the reasons for sudden shift
443
Control chart patterns– Contd..
• Gradual shifts in the level

– Occurs when a process parameter changes gradually over a period
of time
– X-bar chart might exhibit such a shift because of incoming quality
of raw materials that would have changed with time
– Change in style of supervision, maintenance program etc.
– R-chart might exhibit such a shift because of new operator,
decrease in worker skill due to fatigue etc.
444
Control chart patterns – Contd..
• Trending pattern
– Differs from gradual shift in level, that trends do not stabilize or settle
down
– Represents changes that steadily increase or decrease
– For X-bar chart, can be due to tool wear, deterioration of equipment,
build up of debris on jigs and fixtures, change in temperature etc.
– For R-chart, it may be due to improvement in operator skill due to on
job training, decrease due to fatigue etc.
445
• Cyclic patterns
– Characterized by repetitive periodic behaviour in the system
– Cycles of low and high points will appear on the control chart
– X-bar chart may exhibit because of rotation of operators, periodic change of
temperature or humidity, seasonal variation in incoming components
– R-chart may exhibit this pattern because of operator fatigue and getting
energized in subsequent breaks, a difference between shifts, periodic
maintenance of equipments etc.
– If samples are taken so infrequently, only the high and low points will be
represented
446
• Wild patterns
– Can be classified as Bunches and Freaks
– Cluster of several observation that are different
from other points and special causes are associated
with these points
– Freaks
• are caused by external disturbances that influence one or
more samples
• They are points that are too small or large with respect to
control limits and fall outside the control limits and
hence easy to identify
• Care should be taken that no measurement or recording
error is associated with that freak point
• Some special causes may be sudden, very short-lived
power failures, use of new tool for a brief test period etc.
447
448
action – Contd..
• Wild patterns – Contd.
– Bunches
• Cluster of several observation that are different from other
points
• Possible causes may be use of new vendor, use of a different
machine, use of new operator etc., for a short time period.
449
action – Contd..
• Mixture patterns
– Effect of two or more population in the sample
– Characterized by points that fall near the control limits, with
absence of points near the centre line
– Might be due to material from two different vendors, different
production method, two or more machine being represented
450
action – Contd..
• Stratification patterns
– Is also due to presence of two or more population distribution
– Output is combined or mixed and samples are selected from it
– Majority of the points fall close to centre line, with very few points
near the control limits
– Can be misinterpreted as indicating unusually good control
451
action – Contd..
• Interaction patterns
– Occurs when the level of one variable affects the
behaviour of other variables associated with the quality
characteristic
– Interaction pattern can be detected by changing the
scheme for rational sampling
– Example, low pressure and low temperature may
produce a desirable effect on output characteristic
– Effective sampling method would involve controlling
the temperature at several high values and then
determining the effect of pressure on output
characteristic for each temperature value
– If the R-chart shows the sample range to be small, then
information regarding the interaction could be used to
establish desirable process parameter settings
452
• Interaction patterns
453
454
Control chart for Mean and Standard
deviation (X-bar and s chart)
• R-chart is easy to construct and use, but a

standard deviation chart (s-chart) is preferable
for large sample sizes (> 10)
• Reason is that range accounts only for the
maximum and minimum sample values and is
less effective for large samples, where as sample
standard deviation serves as a better measure of
process variability
• Population distribution of a quality
characteristic is assumed to be normal with a
population standard deviation denoted by σ
455
deviation (X-bar and s chart)
• Mean and standard deviation of the sample

standard deviation is given by
E (s) c4
 s  . 1  c4 2 where c4
depends on the sample size given by
1  n  2   !
 2    2 
2 
c4    n  3  
 n  1     !
  2  
456
deviation (X-bar and s chart) – Contd..
g
• No Given Standards s i
– Centre line of a s-chart is CLs  s  i 1
g
s
where g is the no. of samples and ˆ  c
4
si is the standard deviation of the ith sample
– UCL s = s  3 .  s  s  3 .  . 1  c 4
2
– Estimate of a population standard deviation is

– Hence control limits are given by
s 2
UCLs  s  3. . 1  c4  B4 s
c4
s 2
LCLs  s  3. . 1  c4  B3 s
c4 457
• No Given Standards
• X-bar chart
– Centre line is similar to that of X-bar chart
– Control limits are given by
s
UCLX  X  3.  X  A3 s
c4 n
s
LCLX  X  3.  X  A3 s
c4 n
• S-chart is constructed first as the standard deviation of
X-bar is dependent on ‘s’ and if the s-chart is not in
control, any estimate of the standard deviation of X-bar
chart is unreliable
458
• Given Standard
– If the target standard deviation is given as σ0 then, centre line of
a s-chart is CLs  c4 0
– Hence control limits are given by
2
UCLs  c4 0  3 s  c4 0  3. 0 . 1  c4  B6 0
2
LCLs  c4 0  3 s  c4 0  3. 0 . 1  c4  B5 0
– X-bar Chart
• Target value is specified as Xo then control limits are given
by CL X X 0
3
UCLX  X 0  A 0 , where A 
n
LCLX  X 0  A 0 459
Example
The thickness of the magnetic coating on audio tapes is an

important
characteristic. Random samples of size 4 are selected, and the
thickness is measured using an optical instrument. Table 7-3 shows
the mean Xand standard deviation s for 20 samples. The
specifications are 38 ±4.5 micrometers. If a coating thickness is less
than the specifications call for, that tape can be used for a different
purpose by running it through another coating operation.(Sum
X=743.5, Sum s=95.8)
460
(a) Find the trial control limits for an X- and an s-chart.
(b)Assuming special causes for the out-of-control points,

determine the revised control limits.
(c) Assuming the thickness of the coating to be normally

distributed, what proportion of the product will not meet
specifications?
(d)Comment on the ability of the process to produce items

that meet specifications.
(e) If the process average shifts to 37.8 micrometre, what

proportion of the product will be acceptable?
461
462
463
464
Thank You
465
Reliability

Pilani Campus
Control chart for individual units – MR
Chart
• Chart for sample size is 1, i.e. for individual units
• Reasons for sample size to be ‘1’
– The rate of production is low
– Testing process may be destructive and the cost of the item is
very high
– If every manufactured unit is inspected
• The value of the quality characteristic is expressed as ‘X’
• Variability of the process is estimated from the ‘Moving
Range’, that are found from successive observations
– Moving Range of 2 observations is simply the difference
between them
– Moving Range are co-related, because they use common rather
than individual values and hence the pattern of MR chart must
be interpreted carefully
467
Chart – Contd..
• Also, it can’t be assumed that X-values for the individual
units are normally distributed as in other control charts
• Hence checking of the distribution of individual values
is done first, by using frequency histogram to verify the
shape, skewness, kurtosis of the distribution
• No Given Standards
– An estimate of the process standard deviation is
– Where MR is the average of moving range of successive
observation MR
ˆ 
– If there are ‘g’ observations, there will be (g -1) moving rangesd 2
– Centre line is given by
CL  MR
MR
468
Chart – Contd..
MR
UCLMR  MR  3.  D4 MR
d2
MR
LCLMR  MR  3.  D3 MR
d2
– For n = 2, D4 = 3.267, D3 = 0 and hence UCL = 3.267MR and LCL
=0
• X-Chart
– Centre line of X-chart is given by CL X  X
MR
UCLX  CL  3  X  3.
d2
MR
LCLX  CL  3  X  3.
d2
469
Example
470
Example
471
Example
472
Example
473
Chart – Contd..
• Given Standards
– If standard values are specified as Xobar and σ0, then,
CLX  X o
UCLX  X o  3. o
LCLX  X o  3. o
– Assuming n = 2, (difference between 2 values only)
CLMR  d 2 o 1.128 o
UCLMR  D4 d 2 o  3.685 o
LCLMR  D3 d 2 o  0
474
Example
Table shows the Brinell hardness numbers of 20 individual steel

fasteners and the moving ranges. The testing process dents the
parts so that they cannot be used for their intended purpose.
Construct the X-chart and MR-chart based on two successive
observations.
Specification limits are 32 ± 7.
475
476
Chapter - 8
Control Charts for

Attributes
477

Introduction
• Attribute
– It is a quality characteristic for which a numerical value is not
specified
– It is measured on a nominal scale, i.e., it does or does not meet
certain guidelines
– It is categorized according to the scheme or levels
• Nonconformity
– A quality characteristic that does not meet certain standards or
specification is said to be non conformity
– Eg. Length of a bar 40 ± 0.5. Both 40.6 and 42 are nonconformity
• Nonconforming item
– A product with one or more non-conformities such that it is unable
to meet the intended standards and is unable to function as
required
– It is possible to have several non-conformities on a product without
being classified as a nonconforming item
478

Introduction – Contd..
• Types of charts
– Based on binomial distribution
• Proportion of non conforming items (p-chart)
• No. of non conforming items (np-chart)
– Based on Poisson distribution
• Total number of non conformities items (c-chart)
• Non conformities per unit chart (u-chart)
– Used in situation where the size of the sample varies
from sample to sample
– Based on weights
• Chart for Demerit per unit (U-chart)
– Deals with combining nonconformities on a weighted
basis, based on the severity of the non conformity 479

Advantages and Disadvantages
of Attribute charts
• A variable quality characteristic can be measured as an
attribute, because of limited time, money, worker
availability etc., if attribute information is alone sufficient
• Advantages
– For a given product, if variable charts are used for each of its
quality characteristic, no. of charts constructed will be very high
and in such case, attribute control chart provides overall quality
information
– Can be used to summarize information about several
components of a product
– Attributes are encountered at all levels of an organization
(company, plant, department, operator level etc), but variable
charts are used at lower only (machine level)
– Provides effective information to top management
– Helps in going from the general to more focussed level
480

Advantages and Disadvantages of
Attribute charts – Contd..
• Disadvantages
– Attribute information does not indicate the degree to which
specification are met or not met, but variable provide more
information a process and performance
– Variable plots provide more information on the potential
causes and hence makes the identification of remedial
action easier for out of control situation
– Variable chart can forewarn, when the process goes out of
control, while attribute does not detect a lack of control,
until the process parameters are changed
– When specification limit are equal or tighter than the
inherent variability of the process, attribute chart indicates
out of control process
– Attribute chart requires larger sample size, to ensure
adequate protection against the certain level of process
changes
• If no historical information is available, attribute
control charts are used first and once problem area is 481
identified, it is replaced by a variable control chart
Preliminary decisions
• Level
– Decide on the level at which it has to used – plant, department,
operator etc.
• No. of quality characteristic
– Chart for single quality characteristic, multi quality characteristic
or for a product
– Depends on no. of potential quality characteristic, no. of products
produced, cost and time required for inspection
• Sample size
– Should be large enough to allow non conformities or non
conforming items to be observed in the sample
– Eg. Non conformance rate is 2.5%, while sample size is 25,
average no. of non conforming per sample is 0.625
– Small sample size leads to misleading inference
• Nonconformity
– What will be considered a nonconformity should be defined
– It depends on product, functional use and customer needs
482

Preliminary decisions – Contd..
• Frequency of sampling
– Depends on production rate, cost of sampling etc.
• Choice of measuring instruments
– Accuracy of the instruments directly influences the quality of
the data calculated
– Should be calibrated and tested
– Characteristic being controlled, the desired degree of
measurement precision, both impacts the selection
• Design of data recording forms
– Includes part name, sample number, date, time, raw values of
data, no. of non conforming items, operator name, machine,
gauge, comments etc.
• One, two or three-sigma zones and rules related to it are
not used because the underlying theory in non-normal
483

Control chart for proportion
nonconforming (p-chart)
 p-chart is one of the most versatile chart
 Used to control the acceptability of single quality
characteristic or a whole product
 Also used to measure the quality of an operator,
machine, department or entire plant and also the
performance of a CEO
 Provides fair indication about the general state of the
process based on the average quality level of the
proportion non conforming
 It can act as a source of locating problems and provides
information for improving product quality
 Based on binomial distribution
484

p-chart – Contd..
Assumptions
– Probability of obtaining a non conforming item
must be constant
– Sample must be identical
– Samples are independent, but is not valid for
samples manufactured in groups (Eg. Steel ingots
– batch production)
– For a given sample, proportion of non conforming
is given by x
ˆ
p
n
where x = no .of nonconforming item in sample, n

= sample size 485

p-chart – Contd..
 The distribution of no. of nonconforming item in a
sample is given by binomial distribution
n!
p x 1  p 
n x
P( X  x) 
n  x ! x!
where x 0,1,2,.....n
 p = probability of getting a non conforming item
on each trial
 Mean and variance of sample proportion non
conforming is given by
E ( pˆ ) p
p (1  p )
Var ( pˆ ) 
n 486

p-chart – Contd..
No standard given
– Proportion of non conforming is specified, it must
be estimated from the sample information
x
pˆ 
n
– Average of these individual sample proportions non
conforming is used as the centre line g g
g = no. of samples  pˆ i  xi
CL p p  i 1  i 1
xi = no. of non conforming items g ng
– Variance is calculated from
p (1  p )
Var ( pˆ ) 
n 487

p-chart – Contd..
No standard given – contd..
– True value of ‘p’ is not known, hence value of
p is used as an estimate
p (1  p )
UCLp p  3 .
n
p (1  p )
LCL p p  3 .
n
488

p chart – contd..
No. of defectives
489

p chart – contd..
g g
 pˆ  x
i i
CL p p  i 1  i 1 0.129
g ng
p (1  p ) 0.129 +
UCLp p  3. 3 [0.129(1 – 0.129)/50]
n
= 0.271
p (1  p )
LCL p p  3 .
n 0.129 -
3 [0.129(1 – 0.129)/50]
=0 490
p chart – contd..
491

p-chart – Contd..
Standard given
– If the target value of the proportion of non
conforming item is known, the centre line is
selected as that target value, i.e.
– Control limits are given by CL p p0
p0 (1  p0 )
UCLp p0  3 .
n
p0 (1  p0 )
LCL p p0  3 .
n
– If lower limit turns out to be negative, then
LCL is taken as zero
492

p chart – contd..
493

Problem
Twenty-five samples of size 50 are chosen
from a plastic-injection molding machine
producing small containers. The number
of nonconforming containers for each
sample is shown in Table 8-2, as is the
proportion nonconforming for each
sample,
494

p-chart – Contd..
495

Problem
Management has decided to set a standard of
3% for the proportion of nonconforming test
tubes produced in a plant. Data collected
from 20 samples of size 100 are shown in
Table 8-3, as is the proportion of
nonconforming test tubes for each sample.
496

p-chart – Contd..
497

p-chart – Contd..
Unique features
– If the p-chart shows continuously increase in the
average proportion non conforming, management
should investigate the reasons behind the
increase rather than revising the limit
– Causes can be lower incoming quality from
vendor or tightening of specification limit
– Downward trend in the proportion
nonconforming is most desirable
Control limits for representative sample
sizes
– Based on those sizes that occur more frequently
– Should span the range of sample sizes
498

p-chart – Contd..
Variable sample size
– Size varies due to rate of production, lack of inspection
personnel, change in unit cost of inspection
– Change in sample size results in change in control
limits though the centre line remains constant
– Sample size increases, control limits become narrower
– Control limits for individual sample sizes are given by
p (1  p )
UCL p  3 .
ni
p (1  p )
LCL p  3 .
– If lower limit turns out to nbe
i negative, then LCL is
taken as zero
499

p chart – contd..
500

p chart – contd..
501

p chart – contd..
Day 1 2 3 4 5 6 7 8 9 10
n 500 550 700 625 700 550 450 600 475 650
D 5 6 8 9 7 8 16 6 9 6
p 0.010 0.011 0.011 0.014 0.010 0.015 0.036 0.010 0.019 0.009
pbar 0.014 0.014 0.014 0.014 0.014 0.014 0.014 0.014 0.014 0.014
UCL 0.029 0.028 0.027 0.028 0.027 0.028 0.030 0.028 0.030 0.027
LCL 0 0 0 0 0 0 0 0 0 0
502

p chart – contd..
p-chart
0.060
0.050
p
0.040
pbar
0.030
p
0.020 UCL
0.010 LCL
0.000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Day
503

Problem
Twenty random samples are selected from a

process that makes vinyl tiles.
The sample size as well as the number of
nonconforming tiles are shown in Table 8-4.
504

505

506

507

Control chart for proportion
nonconforming (p-chart) – Contd..
Control limits based on average sample
g
size n i
– Mean of these sample size is n i 1
g g
g
– Centre line is given by
 pˆ ii x
CL p p  i 1  i 1
g = no. of samples g ng
xi = no. of non conforming items
p (1  p )
UCL p  3 .
– Control limits are given by n
p (1  p )
LCL p  3 .
– Certain precautions should be taken while makingn
inferences and individual control limits have to be
calculated to allow a proper decision to be made
regarding the particular sub group 508

p chart – contd..
509

p-chart – Contd..
Standardized control chart
– Used for varying sample size
– Value of the proportion non conforming of a sample is expressed as
the sample’s deviation from the average proportion non
conforming in units of standard deviation
– They have a mean of zero and control limits are ±3 and are
constant
ispˆ given
 by

p 1 p 
– Mean is given Eby( pˆ ) p and standard deviation
n
p = true process non-conformance rate n = sample size
– Standardized value of proportion
pˆ i  p non-conforming for the ith
sample is Z 
p 1  p 
ni
510

511

512

p chart – contd..
513

Special considerations for p-chart
Observations below the LCL
– Points falling below LCL is desirable as
it indicates an improvement in process
– Care should be taken such that error
measurement has not occurred
Comparison with a specified standard
– When standard is given and if points
fall outside the UCL and no special
causes can be assumed, then existing
process is not capable
514

Special considerations for p-chart
 Impact of design specifications
– Non conforming product refers to quality
characteristics that do not meet ‘certain
specifications’ and hence even if process is in
control, average proportion of non conforming is
high
– Requires change in the design or specification of
the product and tolerance should be loosened with
no deviation from customer needs
 Information about overall quality level
– Used for aggregating information and can be used
at plant with many product lines, machine etc and
will provide overall product non conformance rate
515

Chart for No. of nonconforming
items (np-chart)
 Based on the count of no. of nonconforming items in
sample
 Operator may feel easy to use this chart
 Assumptions used in p-chart is applicable for this chart
also
 Drawback of np-chart
– If sample size changes, the centre line and control limit
g
changes, and hence inference becomes difficult
 No standard given
xi
CLnp np  i 1
– Centre line is given by g
– xi = no. of non-conforming for ith sample  np  np (1  p )
– g = no. of samples
– n = sample size
– Standard deviation of number of nonconforming is 516
np-chart – contd..
 Control limits are given by
UCLnp np  3 . np (1  p )
LCLnp np  3. np (1  p )
 If LCL is negative, then it is taken as zero
 Standard given
– If standard is specified for number of non conforming
items is given as npo, then the limits are given as
CLnp npo
UCLnp npo  3 . npo (1  po )
LCLnp npo  3 . npo (1  po ) 517

np chart – contd..
No. of defectives
518

Problem
Data for the number of dissatisfied customers

in a department store observed for 25
samples of size 300 are shown in Table 8-6.
Construct an np-chart for the number of
dissatisfied customers.
519

520

521

Control chart for the number of
nonconformities (c-chart)
Nonconformity
– Defined as the quality characteristic that does not meet some
specification
– Non conforming item has one or more nonconformities that
make it non functional
Introduction
– A c-chart is used to track the total no. of non conformities in a
sample of constant size
– When the sample size varies, u-chart is used to track the no. of
non conformities per unit
– Size of the sample is also called as “Area of Opportunity”
– Area of opportunity may be a single or multiple units of a
product or it can also be continuous
– If average no. of nonconformities per unit and area of
opportunity are small, then most observations show zero non
conformities, which is misleading and hence it is beneficial to
choose large area of opportunity
522

c-chart – Contd..
 Introduction – contd..
– Occurrence of nonconformities is assumed to follow a ‘poisson
distribution’
– Poisson distribution is well suited to model the number of
events that happen over a specified amount of time, space or
volume
 Assumptions
– Opportunity for the occurrence of non conforming should be
large
– Average no. of non conformities per unit must be small
– Occurrences of non conformities must be independent of each
other
– Each sample should have an equal likelihood of the occurrence
of non conformities
– Eg. Different screw drivers are used to tighten the screws in a
product, the opportunity for change in torque will vary and
hence should not be used 523

c-chart – Contd..
If ‘x’ represents the no. of
nonconformities in the sample unit and
‘c’ is the mean, then the poisson
distribution is given as e cc x
p( x) 
x!
 p(x) is the probability of observing ‘x’ non conformities
– Mean and variance are equal for poisson distribution
– Average no. of nonconformities per sample is found
from the sample observation, denoted by ‘c-bar’ and
the limits are CLc  c 524
UCLc  c  3 . c
LCLc c  3 . c BITS Pilani, Pilani Campus
c-chart – Contd..
Standard given
– Let the specified goal for the no. of non
conformities per sample unit be co
CLc  co
UCLc co  3 . co
LCLc co  3 . co
525

c chart – contd..
526

Problem
Samples of fabric from a textile mill, each 100

m2, are selected, and the number of
occurrences of foreign matter are recorded.
Data for 25 samples are shown in
Table 8-7. Construct a c-chart for the number
of nonconformities.
527

528

nonconformities per unit (u-chart)
Introduction
– Used, when the area of opportunity changes
– Used, when companies inspect all items produced or
services rendered, for the presence of non conformities
– Generally output per production run vary, because of
fluctuating supplies of labour, machines, raw material
and hence no. of units inspected changes, thus resulting
in varying sample size
– The control limits change as the sample size varies, but
centre line remains same
– When sample size varies, the no. of non conformities per
unit for the ith sample is given by ui = ci / ni
– ci is the no. of non conformities in the ith sample
– ni is the size of ith sample and it may not be a integer
529

nonconformities per unit (u-chart) – Contd..
 No standard given – contd..
– Average no. of nonconformities per unit and the control
limits are given by
g u
c UCLu  u  3 .
i 1
i
ni
CLu  u  g
u
n i LCLu u  3 .
i 1 ni
 As the sample sizes increases, control limits draws
closer
 If n = 1, all formula will be equal to that of c - chart
i
530

u-chart – Contd..
In a textile finishing plant, dyed cloth is
inspected for the occurrence of defects
per 50 square meters. The data on ten
rolls of cloth are shown below. Use these
data to set up a control chart for
nonconformities per unit. The center line
of the chart should be the average
number of non-conformities per
inspection unit--that is, the average
number of nonconformities per 50
square meters.
531

u chart – contd..
Roll Number Number of Number of Number of Number of UCL LCL
Square Noncon Inspecti Noncon
Meters formitie on formitie
s Units in s per
Roll, n Inspecti
on Unit
1 500 14 10 1.4
2 400 12
3 650 20
4 500 11
5 475 7
6 500 10
7 600 21
8 525 16
9 600 19
10 625 23 532

u chart – contd..
533

Problem
The number of nonconformities in carpets
is determined for 20 samples, but the
amount of carpet inspected for each sample
varies. Results of the inspection are shown
in
Table 8-8. Construct a control chart for the
number of nonconformities per 100 m2
534

535

Charts for Demerits per unit
(U-chart)
 ‘c’ and ‘u’ chart treat all types of non conformities
equally, regardless of their degree of severity
– Eg. Monitor A has trouble in colour while Monitor B has 5
scratch marks on its surfaces – here Monitor A defect is
more serious than Monitor B’s
 An alternative approach is to assign “weights” to non
conformities according to their relative degree of
severity
 The quality rating system, which rates demerits /
unit is called U-chart and is often helpful in service
applications
 Classification of non conformities
– Classification is based on degree of seriousness, ANSI
standard A3 classifies in the following manner:
• Class 1 defects: Very serious and defects lead directly to
‘severe’ injury or to catastrophic economic losses
536

(U-chart) – Contd..
 Classification of non conformities – contd..
• Class 2 defects: serious and defects that lead to
‘significant’ injury or to significant economic losses
• Class 3 defects: major and defects that cause ‘major’
problems with normal use of product or service rendered
• Class 4 defects: minor and defects that cause ‘minor’
problems with normal use of product or service rendered
 Once classification of defects or non conformities has
been established, demerits per unit are assigned to each
class
 Definition of classes and no. of classes are not rigid and
varies with respect to organization and the problem
environment
 Assigned weights for defects is also user dependent, while
ANSI standard uses 100, 50, 10, 1 for the above
mentioned classes 537

Construction
– Let us have 4 categories of non conformities
– Let the sample size be ‘n’
– Let c1, c2, c3 and c4 denote the total no. of non
conformities in the sample for 4 categories
– Let w1, w2, w3 and w4 denote the weights
assigned to each category
Assumption
– Non conformities in each category is
independent of defect in other categories
– Occurrence of non conformities in each
category is given by Poisson distribution
 For a sample size of ‘n’, total no. of demerits is given by D538
= w1c1 + w2c2 + w3c3 + w4c4 BITS Pilani, Pilani Campus
 Demerits per unit of the sample are given by
D w1c1  w2 c2  w3c3  w4 c4
U 
n n
 U is the linear combination of independent
poisson random variable
 Centre line of the U chart is U  w u  w u  w u  w u
1 1 2 2 3 3 4 4
– u1 ,....u4 average no. of non conformities
per unit in respective classes and are
g
calculated as  c i
i 1
u  g
n
i 1
i 539

 Estimated standard deviation of U is given by
2 2 2 2
w1 u1  w2 u2  w3 u3  w4 u4
ˆU 
n
 Control limit of the U chart is UCLU U  3 .ˆ U
LCLU U  3 .ˆ U
 If LCL is negative, then take LCL as zero
540

Problem
541

U chart – contd..
542

U chart – contd..
543

Thank You
544
Reliability

Pilani Campus
BITS Pilani
Pilani Campus
Process Capability
546
Process Capability
Process capability represents the performance of a process in a
state of statistical control. It is determined by the total
variability that exists because of all common causes present in
the system.

SPECIFICATION LIMITS AND
CONTROL LIMITS
• Tolerance limits are generally preferred in evaluating
manufacturing or service requirements, whereas specification
limits are more appropriate for categorizing materials,
products, or services in terms of their stated requirements.
• In general, tolerances are a subset of specifications.
• Remember that the control limits on an X bar-chart are a
measure of variability of the sample means. Specification
limits represent the acceptable bounds of variability for
individual items.

example
The diameter of a part has to fit an assembly. The
specifications for the diameter are 5 ± 0.015 cm. The
samples taken from the process in control yield a sample
mean X of 4.99 cm and a sample standard deviation sigma
of 0.004 cm. Find the natural tolerance limits of the
process. Would you consider adjusting the process center?

SPECIFICATIONS AND PROCESS CAPABILITY

Process Capability
Process capability is the ability of the process to meet the
design specifications for a service or product.
Nominal value is a target for design specifications.
Tolerance is an allowance above or below the nominal
value.

Process Capability
Nominal
value
Process distribution
Lower Upper
specification specification
20 25 30
Process is capable

Process Capability
Nominal
value
Process distribution
Lower Upper
specification specification
20 25 30
Process is not capable

Process Capability Ratio, Cp
Process capability ratio, Cp, is the tolerance width divided by 6 standard

deviations (process variability).
Upper specification - Lower specification

Cp =
6

Example
The relative humidity in a greenhouse is expected to be
between 65 and 85 %. Random samples taken over a span
of one week yield the following values: 60,78,70,84,81, 80,
85, 60, 88, 75. Find and interpret the process capability
index.

Process Capability Index, Cpk
Process Capability Index, Cpk, is an index that measures the potential

for a process to generate defective outputs relative to either upper
or lower specifications.
=
Cpk = Minimum of
x – Lower specification , Upper specification – x=
3s 3s
We take the minimum of the two ratios because it gives the worst-
case situation.

Intensive Care Lab
Example
The intensive care unit lab process has an average turnaround time
of 26.2 minutes and a standard deviation of 1.35 minutes.
The nominal value for this service is 25 minutes with an upper
specification limit of 30 minutes and a lower specification limit of 20
minutes.
The administrator of the lab wants to have three-sigma performance
for her lab. Is the lab process capable of this level of performance?
Upper specification = 30 minutes

Lower specification = 20 minutes
Average service = 26.2 minutes
 = 1.35 minutes

Intensive Care Lab
Assessing Process Capability
Example Upper specification = 30 minutes
Lower specification = 20 minutes
 = 1.35 minutes
= =
Cpk = Minimum of x – Lower specification , Upper specification – x
3s 3s
Cpk = 26.2 – 20.0 30.0 – 26.2

Minimum of
3(1.35) , 3(1.35)
Process
Cpk = Minimum of 1.53, 0.94 = 0.94 Capability
Index

Intensive Care Lab
Assessing Process Capability
Example
Upper specification - Lower specification
Cp =
6
30 - 20
Cp = = 1.23 Process Capability Ratio
6(1.35)
Does not meet 3 (1.00 Cpk) target due to a shift in mean

(Note variability is ok since Cp is over 1.0)
Before Process Modification

Upper specification = 30.0 minutes Lower specification = 20.0
minutes
 = 1.35 minutes Cpk = 0.94 Cp = 1.23

Example
In an electrical circuit, the capacitance of a component
should be between 25 and 40 picofarads (pF). A sample of
25 components yields a mean of 30 pF and a standard
deviation of 3 pF. Calculate the process capability index
Cpk, and comment on the process performance. If the
process is not capable, what proportion of the product is
nonconforming, assuming a normal distribution of the
characteristic?

569
Example
In an assembly operation in a semiconductor manufacturing
company, the lower and upper specification limits are given by
4.8 and 5.2 seconds. A random sample of 25 completion times
gave a mean and standard deviation of 5.12 and 0.06 seconds,
respectively. Can we conclude that the Cp index for this
operation exceeds 1, so as to be considered acceptable by the
customer? Test at a significance level of 0.05.
570
Thank You
571
Quality Control Assurance
Reliability

Pilani Campus
Acceptance Sampling
573
Definition
Let us take a sample of size n randomly from a batch. If the

number of defective articles in the sample is not greater than
a certain number c, then the batch is accepted; otherwise, it
is rejected.
This is how we define acceptance sampling plan.

Probability of Acceptance
Let us assume that the proportion of defective articles in the

batch is p. Then, when a single article is randomly chosen
from a batch, the probability that it will be defective is p.
Further assume that the batch size is sufficiently larger than
the sample size n so that this probability is the same for each
article in the sample. Thus, the probability of finding exactly r
number of defective articles in a sample of size n is
n!
P r   p r 1  p 
n r
r !n  r !
Now the batch will be accepted if r ≤ c, i.e., if r=0 or 1 or 2

or….or c. Then, according to the addition rule of probability,
the probability of accepting the batch is

Operating Characteristic
r c
n!
Pa  p  P r 0   P r 1  P r 2     P r c   p r 1  p 
n r
r 0 r !n  r !
Pa  p 
This tells that once n and Ideal OC Curve
c are known, the 1

probability of accepting a
batch depends only on the
Practical OC Curve
proportion of defectives in
the batch. Thus, Pa(p) is a
function of p. This
function is known as
p
Operating
Characteristic (OC) of
the sampling plan.
Construction of OC curve
we assume that the sample is chosen from an isolated lot of finite size.
we assume that a stream of lots is produced

If the lot size is large and the
by the process and that the lot size is large
probability of a nonconforming item is
(at least 10 times) compared to the sample
small
size.
Construct an OC curve for a single sampling plan where the lot size is 2000,
the sample size is 50, and the acceptance number is 2.

Acceptable Quality Level (AQL) p=p1
This represents the poorest

level of quality for the Pa  p 
producer’s process that the
consumer would consider to 1 
be acceptable as process 1 
average. Ideally, the producer

should try to produce lots of
quality better than p1.
Assume there is a high
probability, say 1-, of
accepting a batch of quality p1
p
p1. Then, the probability of

rejecting a batch of quality p1
is , pwhich
 p1 Pa  p1  1  as
is known .
producer’s risk. BITS Pilani, Pilani Campus
Lot Tolerance Proportion Defectives (LTPD) p=p2>p1
This represents the poorest

level of quality that the Pa  p 
consumer is willing to accept
1
in an individual lot. Below this 1 

level, it is unacceptable to the
consumer. In spite of this, let
there will be a small chance
(probability)  of accepting

such a bad batch by the

consumer,  is known as p
p1 p2
consumer’s risk.
pp 2 Pa  p2  .
When then
LTPD is also known as
rejectable quality level (RQL). BITS Pilani, Pilani Campus
Producer and consumer risk can also be demonstrated through the
OC curve. Suppose that our numerical definition of good quality
(indicated by the AQL) is 0.01 and that of poor quality (indicated
by the LQL) is 0.11. From the OC curve, the producer's risk alpha is
1 - 0.986 = 0.014. We consider batches that are 1 % nonconforming
to be good. If our sampling plan is used, such batches will be
rejected about 1.4% of the time. Batches that are 11%
nonconforming, on the other hand, will be accepted 8.8% of the
time. The consumer's
risk is therefore 8.8%.

Some Key points
• Number of acceptance will be change if

proportion of nonconformity varies
• Changing sample size causes drop in

acceptance probability
• probability of acceptance decreases for a given

lot quality as the acceptance number c
decreases.

In a single sampling plan, the information obtained from one sample
is used to make a decision to accept or reject a lot. There are two
parameters in this sampling plan: the sample size n and the acceptance
number c.
A double sampling plan involves making a decision to accept the lot,

reject the lot, or take a second sample. If the inference from the first
sample is that the lot quality is quite good, the lot is accepted. If the
inference is poor lot quality, the lot is rejected. If the first sample
gives an inference of neither good nor poor quality, a second sample is
taken.

Rectifying inspection as it applies to lots that are rejected
through sampling plans. Usually, such lots go through 100% inspection,
known as screening, where nonconforming items are replaced with
conforming ones.
The average outgoing quality (AOQ) is the average quality level of a series
of batches that leave the inspection station, assuming rectifying inspection,
after coming in for inspection at a certain quality level p.
Taking N as the lot size, n as the sample size, p as the incoming lot quality,
and Pa as the probability of accepting the lot using the given sampling plan,
the average, outgoing quality is given by

Single sampling plar N = 2000, n = 50, c = 2. Suppose that the
incoming quality of batches is 2% nonconforming.

The average outgoing quality limit (AOQL) is the maximum value, or peak, of the
AOQ curve. It represents the worst average quality that would leave the inspection
station, assuming rectification, regardless of the incoming lot quality.
If rectifying inspection is conducted for lots rejected by the sampling plan, another
evaluation measure is the average total inspection (ATI). The ATI represents the
average number of items inspected per lot.
For single sampling plans, the average total inspection per lot for lots with an
incoming quality level p is given by
For a double sampling plan, the ATI is given by

Construct the ATI curve for the sampling plan where N = 2000, n = 50, c = 2.
Consider the calculations for a given value of the lot quality p of 0.02. The probability of
accepting such a lot using the sampling plan is Pa = 0.920. The ATI for this value of p is
ATI = 50 + (1 - 0.920)(2000 - 50) = 206

Find a single sampling plan that satisfies a producer's risk of 5% for lots that
are 1.5% nonconforming.

Find a single sampling plan that will satisfy a consumer's risk of 10% for
lots that are 8% nonconforming.

Find a single sampling plan that satisfies a producer's risk of 5% for lots
that are 1.8% nonconforming, and a consumer's risk of 10% for lots that are 9%
nonconforming.

Construct the Average Total Inspection (ATI) curve for the
sampling plan where N = 2000, n = 50, c = 2.
Given value of the lot quality p of 0.02. The probability of
accepting such a lot using the sampling plan is Pa = 0.920. The ATI
for this value of p is
ATI = 50 + (1 - 0.920)(2000 - 50) = 206

Thank You
601
Reliability

Pilani Campus
Reliability
603
Introduction
• Why to study about Reliability?

– “Reliability is a measure of the quality of the product
over the long run.”
– We expect the product will function according to
certain expectations over a stipulated period of time
– To ensure customer satisfaction during the
performance phase we must address measures to
improve reliability in the design phase itself.
– The complex nature of products requires many
components in their construction, and its necessary to
have a good reliability for the whole system itself.
This is called as “System Reliability.”
– With the customer and warranty costs in mind, one
must know the chances of successful operation of the
product for at least a certain stipulated period of time
– Such information helps the manufacturer to select the
parameters of a warranty policy
604
Definition of Reliability
• What is Reliability?
– “Reliability is the probability of a product performing its
intended function for a stated period of time under certain
specified conditions”
• Four aspects of Reliability
– Reliability is a probability related concept
– Functional performance of the product has to meet certain
stipulations
• Product design will ensure development of product that meet
or exceed stipulated requirements
• Eg. Strength of cable = 1000Kg, operation it should be more
– Reliability implies successful operation over a certain
period of time
– Operating or environmental condition under which product
use takes place are specified.
• Example: Reliability of a cable is given as having a
probability of successful performance of 0.90 in
withstanding 1000Kg of load for 2 years under dry
condition 605
Life cycle curve
• Most product goes through three distinct phases from
product creation to wear out
• A life cycle curve is a plot between failure rate λ and time
• Also called as ‘Bathtub curve’ (Figure)
• Consists of three phases, namely: Debugging or infant
mortality, Chance failure / useful life time, Wear out / aging
• Debugging phase exhibits a drop in the failure rate as initial
problems identified during prototype testing are ironed out
• In chance failure phase, the failure rate is constant and here
failure occurs randomly and independently. Also, it is called
as useful period
• In wear out phase, an increase in failure rate is observed, as
the product approaches its end of their useful life as parts
age and wear out
606
Life cycle curve – contd..
607
Probability distributions
• Used to model the failure rate

• Exponential distribution
– During the chance failure phase, the failure rate is constant
– Hence ‘exponential distribution’ can be used to describe
the time to failure of the product for this phase
– Probability density function is given by
t
f (t ) e t 0
 is failure rate
– Mean time to failure (MTTF) for exponential
1 distribution is
MTTF 
given by 
– If the failure rate is constant, MTTF is the reciprocal of the

failure rate 608
Probability distributions – contd..
• Exponential distribution – contd..

– For repairable equipment MTTF = mean time
between failure MTBF
– There will be a difference between MTBF and
MTTF only if there is a significant repair or
replacement time upon failure of the product.
– “Reliability at time t is R(t): is the probability of
the product lasting up to at least time t” and it is
given by
• R(t) = 1 – F(t)
F(t) represents the cumulative distribution
function at time ‘t’
– Cumulative distribution function for the 609
exponential distribution is shown (Figure)
• Exponential distribution – contd..
– Reliability function R(t) for the exponential distribution is
shown in the figure
– At time 0, reliability is 1 and it decreases exponentially
with time t
1   e  t dt e  t
0
– “Failure-rate function r(t) is the ratio of the time to failure
probability density function to reliability function”
f (t )
r (t ) 
R(t )
– For exponential distribution, implying a constant failure
rate, then
e  t 
r (t )   t

e
610
Problem
• An amplifier has an exponential time-

to-failure distribution with a failure
rate of 8% per 1000 hours. What is the
reliability of the amplifier at
5000hours? Find the mean time to
failure.
611
Problem
• What is the highest failure rate for a

product if it is to have a probability of
survival (i.e., successful operation) of
95% at 4000 hours? Assume that the
time to failure follows an exponential
distribution.
612
• Weibull distribution
– “It is used to model the time to failure of
products that have a varying failure rate
– Hence a candidate to model the debugging or
wear out phase
– It is a three parameter distribution where
density function is given by
1 
 t   t  
f (t )    e     , t 
        
– The parameters are:
• Location parameter is given by       
• Scale parameter is    0 
• Shape parameter is
   0 
613
• Weibull distribution – contd..
– The probability density function varies for
different values of these parameters (Figure)
– Weibull distribution reduces to exponential
distribution, when  = 0 and β = 1
– For reliability modelling, the location parameter
=0
– For α = 1 and β = 1, the failure rate decreases
with time and can therefore be used to model
components in debugging phase”
– For α = 1 and β = 3.5, the failure rate increases
with time and so can be used to model products
in the wear out phase”. In this case, the weibull
distribution approximates normal
614
• Weibull distribution – contd..
– Reliability function for the Weibull distribution is
by 
R (t ) exp 

t  

 

– Mean time to failure is given by

1 
MTTF   1
 
– Failure rate function is given by
f (t ) t   1
r (t )   
R (t ) 
– The failure rate function for the weibull failure
density function, for values of the parameter β =
0.5, 1, and 3.5 (Figure)
615
616
Capacitors in an electrical circuit have a
time-to-failure distribution that can be
modeled by the Weibull distribution with
a scale parameter of 400 hours and a
shape parameter of 0.2. What is the
reliability of the capacitor after 600
hours of operation? Find the mean time
to failure. Is the failure rate increasing or
decreasing with time?
617
System Reliability
• System reliability
– “Reliability of the product (made up of a number of
components) is determined by the reliability of each
component and also by the configuration of the system
consisting of these components”
– Product design, manufacture, maintenance influence
reliability, but design has a major role
– One common approach for increasing the reliability of the
system is through “redundancy in design”, which is usually
achieved by placing components in parallel.
– As long as one component operates, the system operates
• Systems with components in series
– For the system to operate, each component must operate
– It is assumed that the components operate independently
of each other (Failure of one component has no influence
on the failure of any other component)
A B C
618
Series System Example
619
System Reliability – contd..
• Systems with components in series – contd..
– If there are ‘n’ components in series, then system
reliability is given by Rs = R1 x R2 x - - - - - - Rn
– System reliability decreases as the number of
components in series increases
– Manufacturing capability and resource
limitations restrict the maximum reliability of
any given component
– Product redesign that reduces the no. of
components in series is the viable alternative
620
A module of a satellite monitoring system
has 500 components in series.
The reliability of each component is
0.999. Find the reliability of the module.
If the number of components in series is
reduced to 200, what is the reliability of
the module?
621
– Use of the Exponential Model
• If the system is in chance failure phase, a constant
failure rate could be justified based on which we can
calculate failure rate, mean time to failure and system
reliability
• Suppose the system has ‘n’ components in series
• Each component has exponentially distributed time-to-
failure with failure rates given by
• The system reliability is given by
1 , 2     n
Rs e  1t X e  2t X e  3t X          e  nt

  n  
    i  t 
  i 1  
e 622
– Use of the Exponential Model
• Thus if each component that fails is replaced
immediately with another that has the same failure rate,
the mean time to failure for the system is given by
1
MTTF  n
i  constant

i 1
i
• When all components have same failure rate, If

then
1
MTTF 
n
623
• Systems with components in parallel

– System reliability can be improved by
placing components in parallel as system
will operate as long as at least one of the
components operates.
– The only time the system fails is when all
the parallel components fail
– All components are assumed to operate
simultaneously.
– A system having ‘n’ components in parallel,
with the reliability of the ith component
denoted by Ri, i=1, 2, ----- n.
– Also assume that the components operate
randomly and independently of each other.
624
Parallel System Example
625
• Systems with components in parallel –
contd..
– The probability of failure of each component is
given by Fi = 1-Ri.
– System fails only if all the components fail and
hence the probability of system failure is
n
Fs 1  R1 1  R2    1  Rn  (1  Ri )
i 1
– Reliability of the system is the complement of Fs
and is given by
n
Rs 1  Fs  1  1  Ri
i 1
626
contd..
– Use of Exponential model
• If the time to failure of each component can be
modelled by the exponential distribution,
each with a constant failure rate λi, then the
system reliability, assuming independence of
component operation is
n
Rs 1   1  e   i t
i 1
• Time to failure of the system is not
exponentially distributed
627
contd..
– Use of Exponential model
• In the special case, where all the components have the
same failure rate the system reliability is
 
n
Rs 1  1  e 
i 1
 i t
Rs 1  1  e  t n
• Mean time to failure for a system of n components in

parallel is given by
1 1 1 1
MTTF         
 1 2 3 n 
628
• Complex system
– A complex system is one which has
components that are both in series and in
parallel
– Assumption
• Components operate independently
• Time to failure of each component is assumed
to be exponentially distributed
– The above described methods are used for
calculating the reliability and failure rate
629
Complex system – Example
630
631
632
System Reliability – Home work
• A company assembles a product from four major
components arranged as follows
B
A C D
• The components are purchased from different
vendors, who have supplied the following
reliability data:
Vendor
Component 1 2 3
A 0.94 0.95 0.92
B 0.86 0.80 0.90
C 0.90 0.93 0.95
D 0.93 0.95 0.95
633
• System with standby components
– “In a stand by configuration one or more
parallel components wait to take over
operation upon failure of the currently
operating component”
– It is assumed that only one component in
parallel configuration is operating at any
given time
– Hence the system reliability is higher than
for comparable systems with components
in parallel
– In parallel systems, all components are
assumed to be operating simultaneously
634
Standby system – Example
635
• Standby components – contd..
– A standby system with a basic component and
two standby components in parallel (Figure) is
shown
Am
As
As
– Typically a failure sensing mechanism triggers
the operations of a stand by component when
the currently operating component fails
636
• System with standby components – contd..
– Use of Exponential Model – contd..
 If the time to failure of the components is assumed to
be exponential with failure rate , the number of failure
in a certain time ‘t’ adheres to a Poisson distribution
with parameter t
 Hence probability of ‘x’ failures in time ‘t’ is given by
P( x) 
t  x
e t
x!
 For a system that has a basic component in parallel
with one standby component, the system will be
operational at time ‘t’ as long as there is no more than
one failure. Therefore, the system reliability would be :
Rs e  t
e  t
637
t 
• System with standby components –
contd..
– Use of Exponential Model – contd..
 For a system (stand by) with a basic
component and two standby components, the
system will be operational if the number of
failures is less than or equal to 2, then
Rs e t  e  t t  e  t
t 2
2!
 For n components on stand by, the reliability
and mean time to failure is given by

Rs e  1  t 
 t t 2

 t 3
 ....... 
t n

 MTTF 
n 1
2! 3 ! n!   638
  BITS Pilani, Pilani Campus
Problems
Exercise Problem 14
• A standby system has a basic unit with four standby
components. The time to failure of each component
has an exponential distribution with a failure rate of
0.008 per hour.
– For a 400 hour operation period, find the reliability of the
standby system.
– What is its mean time to failure of the above system?
– Suppose all five components are operating simultaneously
in parallel, what would be the system reliability?
– What would be the mean time to failure for the parallel
system?
639
Problems – contd..
Exercise Problem 14 Solution
 t


R e 1  t 
t 2

 t 3
 ....... 
t n

 MTTF 
n 1
s  2! 3! n!  

• Rs = exp[-0.008(400)] [1 + 0.008(400) +
(0.008(400))2 / 2 + (0.008(400))3 / 6 +
(0.008(400))4 / 24 = 0.7806
• MTTF = (n+1) /  = 5 / 0.008 = 625 hours
640
Problems – contd..
Exercise Problem 14 Solution
• If all five units were operating in parallel,
system reliability would be
– Rs = 1 – [1 – exp(-0.008(400))5] = 0.18786
• In that case, MTTF is
– (1/0.008) (1+ ½ + 1/3 + ¼ + 1/5) = 285.4167
hours
641
BITS Pilani
Pilani Campus
QC in Service Sector
642
• Quality Control in the
Service Sector
643
Introduction
• Service sector accounts for more than 70% of jobs and the
number continues to grow
• Major difference exists in the ‘quality characteristics’ of
manufacturing and service and accordingly both the
measurement process and management focus differ in these two
sectors
• For example, in service industry, not only the product must meet
high standards, but employee’s behavior must also be
satisfactory
• The service sector therefore works on total service concept,
which is a combination of technical and human behavioral
aspects
• Human behavioral aspects are difficult to quantify, measure and
control
• Example: Airline industry – Goal is to transport people between
cities in desirable time, depends on aircraft design for speed,
scheduling etc, that are quantifiable
• Also depends on airhostess, ticket agents and other people to
run the show 644
Introduction – contd..
• Few more examples include personal services,
education, banking, medicinal, financial, public
utilities, transportation etc.
• Service industries provide a tangible product and an
intangible component that affects customer
satisfaction
• Two parties are involved – one that assists or
provides the service (Vendor) and the party receiving
service (Vendee)
• Service functions are found in manufacturing also
done by staff personnel – provides expertise to the
operating departments
– Includes, customer services, accounting, payroll, R&D etc.
• Differences in the manufacturing and service sector
– A detailed differences are tabulated
645
Differences in manufacturing and
service sector
• Manufacturing • Service
– Product is tangible – Tangible and intangible
– Back orders are – Cannot be stored, if not
possible used
– Producer or company is – Producer and consumer
the only party involved are both involved in the
in the making of the delivery of service
product – Service cannot be resold
– Product can be resold – No formal specification
– Formal specifications given by customer
provided by customer – Customer satisfaction is
– Customer acceptance of difficult to quantify
product is quantifiable
646
Service quality characteristics
• Quality of service can be broken down into two categories:
– Effectiveness deals with meeting the desirable service
attributes that are expected by the customer.
– Efficiency concerns the time required for the service to be
rendered
• Service quality characteristics can be grouped under 4
categories
– Human factors and behavioral characteristics
• Influenced by the attitude and behaviour of the provider
• Includes eagerness to help, thoughtfulness, complacency, courtesy
etc
• Can be developed through training while some are inherent in
individual
• Hence proper screening of employees and assignment of job is
important
– Behavioral characteristics
• Attitude of the customers is beyond the organization control
• Customer mind-sets are often a function of what they expect to
receive 647
Service quality characteristics –
contd..
• Behavioral characteristics – contd..
• Customer expectations can be influenced by
advertisement and reputation
• In turn, behavioural pattern gets affected
• Measurement of attitudes and behavioural
characteristics is not as simple and well-defined
• Timeliness characteristics
• Service that is not used in a given span of time cannot be
stored
• Timeliness with which a service is performed is critical
to customer satisfaction
• Characteristics related to timeliness are categorized by
the service phase, which they are associated
• Categories might include the time to order the service,
the waiting time before the service is performed, the
time to serve, and the post service time
648
contd..
• Service nonconformity characteristics
• A nonconformity is a deviation from the ideal level
• Eg. - No. of errors per 100 vouchers, no. of data entry errors
per 1000 keystrokes, No. of complaints per 50 guests etc
• Target performance level should be zero nonconformities
• Goal is to achieve this through quality improvement measures
• Quality characteristics in this category are well defined and
more readily measured
• Facility related characteristics
• Physical characteristics of faculty associated with service and
delivery can affect customer satisfaction
• Eg. – Ambience in the restaurant, Cyber cafe in a petrol bunk,
Appearance of waiter etc.
• Characteristics are not as clearly defined and measurable but
better than behavioural
649
Examples
• Facility related
characteristics
650
Examples
• Behavioral
characteristics
651
Examples

652
Examples
• Timeliness characteristics
653
Measuring service quality
• Characteristics • Measures
– Human factors and – No. of customer complaints,
behavioural No. of compliments based
on behavioural factors of
persons in service
– Time to process a
– Timeliness
transaction, Waiting time to
receive the baggage, etc
– No. of billing errors in
– Service nonconformity mobile phone usage, No. of
errors in data entry of
marks in database etc.
– Facility related – No. of complaints due to
insufficient legroom in bus,
Lack of lab facilities in a
institute etc.
654
Measuring service quality – contd..
• In terms of ease of quantification and measurement,
categories can be ranked as service nonconformity,
timeliness, facility related and human and behavioural
factors
• Success of many service functions is determined by the
interaction between the provider and the customer
• Because of subjectiveness of quality characteristics,
measurement and evaluation of service quality is difficult,
rather defining the measurement unit itself is problematic
• Many factors influence the behavioral aspects and are
outside the influence of the company and they cannot be
predicted and may cause large performance variation
• Eg. Family life, mental outlook, unforeseen personal events
affect employee behaviour
• To counteract these performance variations in human
behavior, procedures that generate representative statistics
of performance can be devised 655
Measuring service quality – contd..
• Randomly choosing samples of performance from the time
interval under consideration is one way to eliminate the bias
• In situation where we know that behavioral patterns vary
greatly based on the time period (for instance, if error rates are
high in the first and eighth hours of an 8 hour work per day),
we can select sampling plan that adequately reflect this
(stratified sampling)
• Another difficulty is that significant difference exists between
individuals.
• Thus, even though the scheme of stratified sampling is used to
select appropriate samples that reflect the performance of an
individual, it is not obvious whether this same scheme can be
collectively applied to group of individuals
• Individuals vary in their peak performance periods, as people
work best in early morning and others at night and sampling
plan can be designed to reflect them
656
Techniques for Evaluating Service
Quality
• In the service sector, ergonomic, anthropometric, and
behavioral characteristics are important as are the physical
characteristics of the service systems and timeliness with
which the service is provided
• Descriptive statistics that provide numerical measures and the
corresponding graphical methods can be used to describe
distributions of service quality characteristics and their
summary measures
• Eg. Distribution of waiting time before seen by physician can
be shown by frequency histogram and trend chart to find busy
time
• Understanding variability in service quality characteristics is
important to the control and improvement of service quality
• In addition to variation due to equipment, process and
environment, performance variation is also due to person to
person, project to project variation etc.
• Use of Control charts are used to monitor service processes and
to determine status of statistical control
657
Quality – contd..
• Variable control chart are used to measure
quantifiable characteristics like time to serve a
customer in a restaurant, temperature maintained in
a plane etc.
• Attribute control chart are used to service
nonconformity characteristics like proportion of
billing errors, while ‘c’ and ‘u’ charts are applied to
service nonconformity, facility-related and
behavioral characteristics
• Sampling techniques in service operations
– 100% sampling
– Convenience sampling
– Judgment sampling
– Probability sampling 658
Quality – contd..
• 100% sampling
– Used when the cost of external errors or nonconformities is high
– Cost of sampling and inspection is high, but better than a
nonconformity found by a customer
• Convenience sampling
– Samples are chosen by the ease of drawing them and are
influenced by the subjectivity of the individual
– Eg. Selecting a thin file or selecting the top file during audit
• Judgment sampling
– Samples are chosen based on expert opinion
– Can also create bias and hence should be cautious in making
statistical inferences
• Probability sampling
– Has a statistical basis and is the most preferable
– Each item has an equal chance of being selected and done using
random number tables
659
A model for service quality
• Key concepts
– Customer satisfaction is a function of the perceived quality
of service, which is a measure of how actual quality
compares to expected quality.
– External and Internal factors affect customer perceptions of
quality
– External factors
• Are not directly under the control of a service organization
• Includes the social values and lifestyles of customers,
knowledge of service, and the services offered by the
competitors
– Internal factors
• Are directly under the control of service organization
• Includes the image management, client management and
service delivery system
660
A model for service quality – contd..
• Internal factors – contd..

– Image Management
• Is central to retaining and attracting new customers
• Corporate culture, products and services delivered, market
share etc, impact the reputation
• Advertisement, Annual company reports, quarterly sales
report are some techniques for image management
– Client Management
• Is important for improving the service and to meet the
changing customer needs
• Interviews, polls, surveys and making changes in the facilities
and service provided to meet these needs facilitate retaining
and expanding customer base
– Service delivery management
• Involves creating a corporate culture that motivate employees
and depends on
• Quality assurance philosophy, channels of open management,
customer participation, employee attitude etc. 661
A model for service quality – contd..
• Components of service
– Core service and peripheral service and the benefits are both
tangible and intangible
– Eg. For a travel agency providing bus service
• Core service is transport of people between cities
• Peripheral service are comfort and safety of the bus stations,
meal services, pick up and drop etc.
– Peripheral service play a major role in influencing the
customer to select one company over another
• Factors involved in service quality
– Timeliness of the service, Adequacy of the service, price of
the service
– Evidence of good service spread through mouth
– Delivery of service must change with customer expectations
662
Applications
• Discussion is about application of quality control and
improvement techniques in the service sector
• Different applications discussed are
– Administrative operations
– Banking
– Education
– Food industry
– Federal, state and local government
– Health care
– Insurance
– Public utilities
– Transportation
663
Applications – contd..
• Education
– Amount of knowledge retained is a function of the individual
receiving the service and the service delivery process
– Controlling quality in education is unique because it involves
transformation of human qualities and values into learning and
retaining knowledge
– Behavioral characteristics
• A major influence on quality is the competency and behavioral
characteristics of the teachers
• Knowledge of their subject is a must, but is not enough and
delivery process is crucial
• Enthusiasm and the ability to stimulate interest and curiosity in
students are traits of a good teacher
• There are not easily quantified, rather they are subjective
• Other measures of teacher quality can be technical background,
certification and recognition by other professional associations,
research and publication in the area of expertise etc.
664
• Education – contd..
– Facility related characteristics
• Science and technology requires current laboratory facility and
equipment
• An adequate library with holding of books and periodicals
complements the delivery process
• Modern computer facilities are especially pertinent today
– Timeliness characteristics
• Proper scheduling of classes, tests, assignments etc.
• Lead time for release of grade sheets, degree certificates
• Proper completion of syllabus, and returning of graded answer
books etc.
– Service nonconformity characteristics
• Proportion of students passing the board exams or in a class, No.
of complaints about the library, or lab facilities, No. of billing
errors in mess and tuition fee bills
665
– The model
• Customers in educational model has conflicting objectives
• Parents want quality education at lower costs
• Employers want graduates exposed to cutting-edge technology
• Board of trustees want to minimize the operational expenditure
of the university
• Others include tax payers, alumni, etc.
• Students are of course customers, but also acts as co-producers
• They play a significant role in determining the quality of end
product through the degree and intensity of their effort and
motivation
• Students are considered as “work-in-process” inventory
• Finished product is the graduate, while the vendor is the high
schools and junior colleges
• Barriers in the system are due to communication between
departments, schools and colleges and flow of information is
curtailed, which prevents the improvement of quality of service
666
– Role of quality
• In education, it is possible to achieve certain degree of
control can be achieved over the customers – students by
selectively choosing the students based on GPA/marks,
scores on standardized tests, Interview etc.
• This directly affects the degree of control on quality of
output
• Characteristics that measure the quality of output have
tangible and intangible components
• Tangible measures may be proportion of graduates
receiving job offers or average salary of graduates with job
offers etc.
• Intangible measure may be satisfaction and fulfillment that
comes from obtaining a quality education – not easily
quantifiable
• Potential of quality control and improvement667
methods are
pretty complex BITS Pilani, Pilani Campus
– Quality tools
• Evaluating the teaching performance by students – does it
really measure the ability and devotion of the teacher?
• Response of student may be motivated by his grade in the
class!!
• Solution is carefully planned evaluation form – questionnaire
• Control chart for average score and range of the score can
indicate trends or special causes
• Control charts can also be used to evaluate facility and service
delivery characteristics
• Eg. No. of students unable to register for a certain course – c
chart helps in identifying a extra section for that course
• Methods of improvement could be handled by Pareto analysis
and cause-and-effect diagram 668
• Administrative operations
– Involves a high volume of paperwork and hence opportunity for
error is high
– To improve the efficiency, a through analysis of the flow of
paperwork is necessary and should eliminate redundant steps
– Mistakes can also be reduced by redesign of form – where data are
entered, Eg., employee filling travel expense form
– Factors that affect are: Form design, Information required, sequence
of actions for processing the desired service function and the
individuals involved
– Form design should be simple and layout should be such that
individuals processing the information can conduct their operations
sequentially and also flipping of pages to be removed
– Personnel training influences efficiency and error rate
– Stability of this service functions can be checked through control
charts – ‘p’ charts and ‘c’ charts can be used
– Eg. No. of errors per 100 purchase orders, Proportion of bills that
have errors in mail order company
– Tools for improvement are Pareto analysis, cause and effect analysis,
experimental design 669
• Banking
– Involves a high volume of activity and necessary to apply quality
control and improvement procedures
– Automation helped in performing tasks quickly and timely manner,
but risk of errors are high
– Behavioral characteristics like courtesy, communication have a major
impact on customer satisfaction
– Timeliness characteristics like waiting time, transaction processing
time also of concern to customer and processing time varies based
on nature of transaction
– Perfect accuracy is expected as far as service nonconformity
characteristics and hence 10% of bank employee work on detection
and correction of errors
– Attribute sampling plan and process control techniques play a major
role
– One important activity is use of MICR which has impact on the
quality of checks
– The requirements of these magnetic characters and the process
through which they are created are very stringent
670
• Banking – contd..
– For example, the characters has to lie within certain regions of the
check area, the ink should be uniformly distributed, properties of ink
etc plays a major role
– If the reader cannot sense the characteristics, the check need to be
processed manually thereby increasing the processing cost
– Hence incoming quality of printed checks and encoding process by
which the check amount is entered requires careful control
– Sampling plans such as ANSI/ASQC Z1.4 can be used to measure the
quality of documents like checks, forms etc.
– Example
• Say MICR Checks are printed in a batch of 30000 each
• Table 10-8, sample size code is K assuming inspection level I
• Assume normal inspection and AQL of 0.4%, Table 10-9 gives sample
size is 125, with a rejection number of 2
– Process control methodology can be used for check processing and
other clerical operations – use of ‘p’ chart and ‘c’ chart
– A measure of process capability is obtained from the center line and
the control limits of the attribute charts
671
• Food industry
– Several quality characteristics differentiate unit of food products and
impact product acceptability
– More clearly an attribute is defined and measured, the easier it is to
judge the quality of the food
– Some quality characteristics are readily measurable – proportion of
fat in milk, % of foreign material in a cereal, etc.
– Other characteristics are not as easily quantifiable – smell of a ripe
fruit, taste of a particular brand of strawberry etc.
– Requirements for the level of quality are set by customer –
wholesaler, distributor etc.
– Specifications and requirements are also influenced by federal
agencies like Food and Drug Administration, FPI, BIS etc.
– Total quality systems approach is required in which the level of
quality is monitored at each step from raw materials to processing to
packing and delivery
– For perishable foods - packaging, delivery and storage of the food
items are important
– Variety of quality attributes exists in food industry
672
• Food industry – contd..
– A consumer may combine specific attributes to arrive at a single
measure of quality, generally referred to as sensory quality
– Sensory quality includes attributes as appearance, flavor and
kinesthetic associated with the product
– Appearance may be color, size, shape while flavor refers to smell and
taste. Kinesthetic or muscle sense includes attribute as texture and
viscosity
– Amount of force required to initiate a saliva flow is a measure of these
two attributes
– Requirements for the level of quality are set by customer – wholesaler,
distributor etc.
– These characteristics are not usually under strict government control
and depends on customer preference
– Another class of quality characteristic are quantitative in nature and
deals with amount of particular component in the product, Eg.
Proportion of protein, fat etc.
– These characteristics are easy to measure and control chart for mean
and range can be used
673
• Food industry – contd..
– Certain quality characteristics cannot be readily judged by senses of the
consumer as they cant detect the presence of toxic items
– This is where the government plays an important role
– Tools that can be used are:
– Xbar and R charts for % of water, protein etc, while c-chart for the micro
organism or bacteria count while p chart can be used in poultry and egg to
monitor the egg shell shape and texture
• Federal, State and Local government
– Provides variety of services like recreational needs, housing,
environmental quality, small business, safety etc.
– Some government agencies provide service directly to consumer and
others are regulatory that deal with upkeep of desired quality levels for
the protection of consumer
– They generate numerous statistics and use frequency or relative frequency
histograms
– Eg. The income distribution families receiving food subsidies from the
government is obtained from relative frequency histogram
– Trend charts can be used to find the total amount allocated by the
government in budget preparation
674
• Federal, State and Local government – contd..
– Opportunities to use control charts and sampling
techniques are wide ranging
– Eg. Income tax: to determine whether tax returns to be
audited or the ratio of adjusted gross income to taxable
income can be monitored using p-chart
– In case of agriculture department, proportion of insect-
infested cocoa beans and the proportion of moldy wheat
can be monitored through p-chart in addition to know
about no. of errors in misclassification of a family,
miscounting etc can be monitored through c-chart
– For a environmental protection agency, if certain
standards are mandated by law, the degree of conformance
to those standards can be done using process capability
analysis
– Level of pesticide in the mineral water can be monitored
with Xbar and R charts 675
• Health care
– Provision of healthcare services is a national concern
– Consumer also influences the level of quality and also other groups
like Health department is highly influential in setting guidelines for
quality health care
– Unusual circumstances exist in health care. Patients are not always
in a position to judge the technical competency of the service
provided
– Neither the administrators also, but only peer reviews by physician
can judge the quality
– In other services, consumer pays for the service rendered, while in
health care its not the case
– Federal government assists certain segments of the population
through Medicare / Medicaid programs
– Program guidelines which limit the maximum amount that may be
paid affect the quality and profitability of services
– Healthcare executives set service-level policies based on providing
adequate service at a reasonable price and plan their execution
676
• Health care – contd..
– Another unusual feature is the organizational structure of hospital
administration
– Governing board and administrative officers are separate from the
medical staff
– In private hospitals, board of trustees may be the owners or
representative of the owners and not physician, while in pubic
hospitals, board members are selected from local citizenry, but may
have professional expertise, contributing to effective management of
hospital
– Executive policies set by board are carried out hospital
administration
– Physicians too have authority and they mostly serve as consultants,
who have contracts to provide specific service like surgery
– These two administrative powers must act in harmony for total
quality system, but their objectives conflict
– Best health care is not always profitable as government has used cost
cutting measures such as ‘diagnosis related group’
677
• Health care – contd..
– In this system, a patient is classified into a DRG based on the
diagnosis and other characteristics like age, length of stay in the
hospital, surgical procedures etc.
– DRG then defines the maximum amount for which hospitals will be
reimbursed by the federal government, which has direct impact on
the quality of service
– Third party insurers also have an indirect impact on the quality of
service, as patients can go to hospitals approved by the insurer
– Depending on plans, insurers also control patient’s choice of
physicians
– To summarize, hospitals are under extreme pressure to cut costs
and yet give high quality service
– Hence hospitals must come up with effective quality control and
improvement measures for both service and cost
– An error free environment is the objective
– Patients demand better levels of service – clean rooms, home like
atmosphere etc.
678
Health care – Quality characteristics and methods of
control
 Pareto chart and cause and effect diagram is used for
• Blood pressure,
quality pulse rate etc. techniques• reducing
improvement Run chartwaiting time,
• Timeadmit time
to obtain an etc.
appointment
• Response time of ambulances • Xbar and R charts
• Admit time in emergency
room services
• No. of adverse comments per
week on nurse’s performance
• No. of errors in blood tests per • ‘c’ chart
100 samples
• Proportion of cases with
inaccurate diagnosis
• Proportion of tests performed
incorrectly • ‘p’ chart
679
• Insurance
– Consumer needs insurance for protection from accidents, hazards,
and casualties
– Personal insurance:- insurance for home, automobile, life (medical
insurance) while Group insurance:- Medical insurance to
employees of an organization
– Organizations also purchase insurance from other agencies to
provide their customers a sense of security
– In these cases, service nonconformity quality characteristics are
critical and the goal is to have no errors
– It involves people and hence training and motivation are critical
– Important functions
• Data collection
• Inputting the data for calculation of casualty probabilities
• Estimating payoffs
• Underwriting policies
• Processing claims etc. which requires human operations and
interactions
680
• Insurance – contd..
– Timeliness characteristics like correcting a complaint or error,
processing a claim or obtaining damage appraisals from adjusters
should be performed within time limit
– Objective of insurance seller
• Increasing profits is the primary concern here
• Pareto analysis helps to identify vital characteristics of improving
profit
• Cause and effect diagram helps to identify the focus areas
• Cutting operations costs and avoid errors through well documented
procedures (Documentation
• Sellers are interested in reducing risks to their own company by
specifying limitations and exclusions in policies
• But this conflicts with the buyers objective of obtaining protection
against all types of risk
– Statistical techniques are widely used: Premiums are set based on
historical data, sampling methods are used to obtain data
– They are used to estimate and analyze the probability of fire,
accident, amount of loss etc. 681
• Insurance – contd..
– Control of sampling and non-sampling errors are vital to the
accuracy of the estimation process
– Generally random sampling is used along with stratified sampling
and cluster sampling techniques
– Non sampling errors include incorrect data transcription,
inaccurate data input to models used for estimation purposes,
use of inappropriate or outdated models for estimation, faulty
computer programming etc.
– Several control charts can also be used:
• P-chart monitor the proportion of unacceptable documents,
proportion of claims not processed or paid etc.
• C- chart can be used to monitor no. of incorrect entries per 100
documents and no. of underwriters per 500 customer accounts
• Xbar and R chart could apply to characteristics as processing time
for claims and the amount paid by the company for policies in
similar category etc.
– Run chart can be used to track the total amount paid on a
monthly basis 682
• Personal services
– Providers deal with individual customers on a frequent basis
– Eg. Beauty shops, hotels and motels, laundry, travel agency etc.
– Providers rely heavily on repeat business, and hence should know the
characteristics and attributes that satisfy the customers as well those
that make them hate the service
– Individuals taste and preference vary and hence businesses try to
incorporate flexibility in the design of services
– Eg. Hotel industry, some customer prefer indoor swimming pool, others
may be interested in snack bar while some would like washbasin to be
outside while some inside
– There are 2 groups of factors – those cause customer satisfaction and
those cause customer dissatisfaction
– Dissatisfied customers do not come back and also spread the word
– Reputation is everything and hence customer feedback is important,
which should be obtained by providing incentives
– Facility related quality characteristics that cause service
nonconformities fall in the customer dissatisfaction category
– Eg. Dirty carpet in a hotel room, inadequate temp. control in a saloon etc.
683
• Personal services – contd..
– Eliminating causes of customer dissatisfaction should be 1st step but it is
not enough and other factors that influence the perceived quality level to
exceed expectations is most important
– Eg. A hair dresser providing a head massage after his basic service, while
a beauty parlor providing a free mehendi design
– This added touch may lead the customers feel that they have obtained
more for their money’s worth
– Such customers act as a best source of advertisement for business and
spread the word making way for repeat business as well new customers
– Factors that cause customer satisfaction fall in human factors and
behavioral or timeliness category. Eg. Courteous behavior of the front
desk attendant in a hotel or a minimal waiting time to be seated in a
restaurant etc.
– Xbar and R chart can be used to control – time to register and check in at
a hotel, time for the suit to be dry cleaned etc.
– P-chart can be used for proportion of rooms available for occupancy or
proportion of on-time delivery of pizza etc.
– Demerit charts may be used to control the weighted quality points from
customer report cards
684
• Public utilities
– include electric power gas and telephone etc.
– Characteristics
• Unique feature is that companies providing these facilities are
nearly monopolies or oligopolies.
• Are regulated by public service commissions that set the rate
structure for the service and not the customer directly
• Indirectly, though, consumer opinions have an impact on the
negotiated rate structure approved by the public service
commission.
• The standard for quality is imposed by the regulatory agencies.
• Customer expects continuous, uninterrupted service
• Utilities have a guaranteed customer base and hence operating
scenario is different from other service industries
685
• Public utilities – contd..
– Recently competition in utility services is the talk of the day.
– Another feature of utilities is that consumer controls the
consumption and generates instantaneous demand, thereby lead
time for demand is zero and hence rely on forecasting
– Providing adequate service under these circumstances is both
science and art. The art is forecasting. (the predictive function).
– Mathematical models are developed to forecast demands.
– There is an art in deciding what model components to consider,
what measures of demand to use. What information from historical
demand to use, and how to weight them.
– These services cannot be stored. It is not feasible to over produce
and store power to meet peak demand at a later point in time while
gas can be stored
686
• Public utilities – contd..
– Service nonconformity and facility related characteristics are
critical to electric power and gas industry
– All 4 categories are important for telephone industry
– Xbar and R chart can be used for variables such as the amount
purchased from outside sources, time to restore service after
failure, waiting time to get telephone installed etc.
– Trend chart track electrical power or gas consumption on a
monthly basis for a single customer or group of customer and help
identify patterns in consumption thereby helps in developing the
model for forecasting
– No. of errors in meter readings per 100 customers, No. of customer
complaints per month, No. of human errors per month in a power
plant can be monitored by c-chart
687
• Transportation
– All 4 categories of quality characteristics play a major role
– Level of quality is largely influenced by customer and some
government agencies also have regulatory powers
– Eg, Civil Aeronautics Board in US receives customer complaints
regarding air travel and summarizes the findings to the airline
companies
– Airline companies can improve based on the remedial action for
the above complaints, but customers don’t write to CAB, rather
they switch airlines
– Other methods like rating forms should be used to find out the
level of customer satisfaction and expectation
– The response rate may be low and can be increased through free
gifts
– Important quality characteristics is safety of operations and the
fatal and nonfatal accidents should always have a goal of zero
– Prevention of accidents through equipment maintenance, checking
and training of personnel is a priority
688
• Transportation – contd..
– Other quality characteristics deal with operation and delivery of
service
– Behavioral characteristics include courtesy of check-in agent, flight
attendants or the cab driver of taxi etc.
– No. of complaints received per month can be monitored by c-chart
– Timeliness characteristics include waiting time at the check-in
counter, waiting time to retrieve baggage, flight delay in arriving at
a destination etc.
– Xbar and R chart may be used to monitor the above measures
– Service nonconformity characteristics include no. of damaged or
lost bags per month, no. of flight cancellation etc which can be
monitored through np-chart
– No. of mistake in reservation per month, no. of airport security
problems per month are monitored through c-chart
689
Thank You
690

Merged Presentation Choladeck

Uploaded by

Copyright:

Available Formats

Merged Presentation Choladeck

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Merged Presentation Choladeck

Uploaded by

Copyright:

Available Formats

Quality Control Assurance &

BITS Pilani Dr. Sudeep Kumar Pradhan, PhD.

• Egyptians demonstrated a commitment to quality in the

• Juran (1974): Fitness for use

• Defined using Quality Characteristics

• Price, Way of serving

• Can be grouped under 4 categories

Cost • Quality of design – Effect

Quality of design or designed

• Relation between QOD, QOC, QOP – contd..

• It is defined as a system that is used to maintain a

• Offline Quality Control

• Statistical Process Control

• Statistical Process Control – contd..

• Acceptance sampling plans

• Quality is not the responsibility of one person in the

• Common causes • Special causes

• Supports the principle that no deviation from a

• Three stages of quality improvement – contd..

• Three stages of quality improvement – contd..

BITS Pilani Dr. Sudeep Kumar Pradhan, PhD.

• All department right from top management to

• Elements of total quality system (ANSI/ASQC STD

• Elements of total quality system (ANSI/ASQC STD

• Elements of total quality system

• Measure of performance of the total quality system

• Internal failure costs

• External quality cost

• Internal quality cost

• Magnitude of quality cost cannot be considered

• Effect on cost – contd..

Plan Proposal - Develop quality policy

Organize Design / - Develop an org. structure

Staff Pre production - Select employee

Direct Production phase - Monitor purchased materials

Control Production and Post - Obtain quality info.

• Social responsibility of the organization is to operate and

• The BBC spends £700 million (nearly $1 billion) a year on

• BBC is the ongoing transition from analogue to

• TQEM practice for your organization

BITS Pilani Dr. Sudeep Kumar Pradhan, PhD.

• Some philosophies and their

• He developed Quality Management Grid

• 4 absolutes of quality management

• 14 step plan for quality improvement

• He believes that management has to adopt a unified

• Quality Improvement – contd..

Eliminating - Categorized as - No such - Special causes

Structural - Bottom up approach: - Top down - Quality

•What are the Basic Concepts of TQM. Explain?

Quality Element Previous Approach TQM Approach

Procurement Price Life-cycle

• I think the topic of quality initially seems to be

• I think the topic of quality initially seems to be

BITS Pilani Dr. Sudeep Kumar Pradhan, PhD.

• It is based on a structure called House of

Voice of the Customer

• Technical Descriptors (Hows)

Voice of the Designer

• Customer competitive assessment