Management of Quality

7/21/2019 Management of Quality

http://slidepdf.com/reader/full/management-of-quality-56e0cd662c396 1/21

UNIT

9

MANAGEMENT OF QUALITY

Structure

9.1 Introduction

Objectives

9.2

ConceptsA3asics of Quality

9 2 1 Importance of Quality and Need for

Improvement

9 2 2 Controlling Aspects of Quality

9.3 Some Basic Concepts of Applied Statistics

9 4 Quality Assurance and Quality Control

9.5 Acceptance Sampling

9 5 1 Sampling Plans

9 5 2 Design of Single Sampling

Plan

9 5 3 Average Outgoing Quality

9 5 4 Double ndMultiple Sanipling

9.6 Process Control

9 6 1 Variability

9 6 2 Control Charts

9 6 3 and

R

Charts

9 6 4 Other Control Charts

9 7 A Note on Variability in Manufacture

9.8 Quality Function and Its Assurance

9.9 Overview of I S 0 9000

9 10 Summary

9 11

Answers to SAQ s

9 1 INTRODUCTION

Quality of a comm odity or service refers to those features which governs its usefulness for

the intended purpose. Th e purpose for w hich a product is intended and whethe r it fulfils the

purpose is normally decided by its ultimate user i.e. customer. In practice the customer

does not usually reveal his mind: e ven he does not usually define the quality and its level

that he needs in clear terms but he wants that his requirelner~ts hould be met with.

As

a

result Ule manufacturer is forced to imgine and intelligently anticipate the needs as

well as the changing taste of users an d then he tries to provide at competitive costs.

The success story

of

Japaxese products is well know n to us. Tod ay the world m arket is

flooded with Japai ese go ods whether it is the field of electronics or of automobiles or home

appliances. Japanese could compete world market over Am erican or European goods .owing

to the unmatched quality of their products. Thus the simple logic behind this overwhelming

success is emphas ising quality backed up with productivity.

In this unit we will describe the concepts of quality its control and tilanageme nt in an

industrial orga lisation . We will develop a simple working definition of quality gradually

extending to control leading to its assurance aspects. W e will also cover som e important

aspects of statistical techniques which are adopted for quality and its control. We will finish

this unit with the modem system approach to quality with a brief overview of I S 0 9000.

Objectives

After studying this unit you should

be

able to

define quality

appreciate importance of quality and need for its improvem ent

describe controlling aspects of quality

describe inspection and their role

in

quality management area



Managerial Control

Strategies

draw single sampling plan,

colnpule Average Outgoing Quality Limit (AOQL ),

prepare control charts and perform capability analysis, and

descr ibe bas ic fea tu res o f IS0 a ~ denefits of certification.

9 2 CONCEPTS BASICS

OF QU LITY

Although qu ality is considered as a comp etitive prerequisite, ye t despite of its

acknow ledged importance, it is probably most. inis-understood concep t. It is very

unfortu nate that people use tlle term Quality very loss ely. Before we give a definition an

explana tion, it will be rather good to share your own views on quality.

In the space provided, w rite your own views and definition of quality of any conuillonly

known product, viz. car, watch, etc. Also write how it can be m easured.

Qualit y is used a variety of ways. Frankly speaking, there is no clear definitio n of quality

viz. people simply try to relate quality with cost. T o them, the best one or the q uality item i

one which is iilost costly. Som etimes, people refer it to grade for exa mple, ilulk

of

grade I

or

2

or pen of A grade o r B grade.

Many times, people try to relate it with so me special feature of the pro duct, viz. strong u~nd

tough, fire-priwf, etc.

Froill custom er's point of view quality is often associated with its value, usefulness or even

the price attached. But when the same entity, i.e. quality is viewed from m anufacturer's

point of view, it is associated with the conform ance to specifications, i.e. with producing

produc ts according to design 'md other specifications.

On e of the accepted definition s of quality by Dr. J. M. Juran is FITNESS OR USE .

According to Phil Crosly, noted Quality

Guru

Quality is Conformance to requirements .

Both these definitions dictate that Quality is a rel tive parameter and that is why it varies

from person to person. B y these practical definition s of quality we find that quality is

a

measure of how clo sely a product or service meets the specified standard s laid for it. If the

product or se rvice lives up to the expectation s, i.e. specified sta ndards, it is of higher qualit

otherwise not.

9 2 1

Importance of Quality and Need for Improvement

Perception towards quality has undergone a se a change. Now-a-days the quality is not

treated from product an gle alone nor it is supposed to be the baby of quality control

department rather it permeates in every thing that an organisation does. It applies to all type

of organisa tions irrespective of their size, irrespective of their activities be

it m alufacturing

coilstruction or catering service. Now-a-days quality ha s turned in to the total experien ce of

the custom er with the organisatioil at all levels. Some of the ma jor ways that quality affects

an organisation are as follows

Reputation and image of organisation

Liability

Productivity

Costs

W e are well familiar with the productslservices rendered by Tata Grou p of Indusuies,

watch es by H.M.T., autom obiles by Bajaj or toileteries by Go drej and Hindus tan Lever. It

all due to quality of their products and its assurance. Reputation and ilxlage oT

org'misation is associated with quality and the cost of the gocdslservic e that it offers.



Liability of the organisations towards the damages or injuries due to poor quality

Management of Q~~ality

productslservices to the eustomer is the second issue. With the increasing awareness and

role of consumer forum s or courts, this issue is also very important.

Next issue is that of productivity. Quality an d productivity are often closely related to each

other. According to the motto of Japanese companies Quality is the easiest way to improve

productivity , one can understand the importance of quality. For the survival of a compan y,

quality backed-up with productivity is a must.

Remem ber that it is the poor prcxluctivity and not the sup erior quality that makes pro duc t

expensive..

Poor quality increases

costs

incurred by the organisation. These include scrap and rework

cost, warranty costs. replacement and repair costs after purchase. Cost of quality is a

measure of cost to the firm for the lack of quality . It is very difficult to measure and often

cannot e found in accounts books. But by overlooking tw underestimating, we cannot lower

its importance. On e must be very careful as most of the cost elements are hidden .

Competitive pressurelcomplaints from the cu stomer, professionalism and desire to increase

sales all have a potential ro le

it

the need for improvement (Figure 9.1).

Strive for

Excellence

Survival Competition

liicreasi~ig ustomers' DRIVES FOR Complaints and

Expectations Suggestions from

Customers

Higher Profit

Globalisation of Trade

Advancemeut of

Technology

Figure

9 1

riving Forces for Improvement

The role of management is significant in coordinating the forces, shown in Figure

9.1,

and

the adoption of newer m odem and cost effective management approach.

9 2 2 Controlling Aspects of Quality

According to common m an's u sage control rne'ms making som ething to behave in a way

we want it to behave . But the technical m eaning of control is different. In fact control is

stiu~ dard f comparison fo r VERIFYIN G or CHE CKIN G the findings of a process.

But upon extending the concep t of control to quality con trol, one finds that above

definitions lack in one vital issue FEEDBA CK. Th us, a retined version would be as

quoted by Juron as, The process of measuring quality performance comparing it to the

requirements and acting on the difference .

Thus, quality coiltrol is a regulatory process through which w e m easure actual performance,

compare it with the stand ards set for it and act on the difference.

The purpose of quality control activities does not restrict to manage the q uality but ev en to

maintain the assurance of continued high quality of products and service. Like any other

control activity it also involves the measu rement, feedback, comparison with the standards

and correction where necessary. The word measurement immediately tells us that there is

one more aspect of quality characteristic, i.e. qualitative aspect.

Again qualitative aspect of quality itself leads to following two different categories

(i) Attributes, and

(ii) Variables.

Attributes are actually

countables

Situations like present or not present (a dichotomous

classification). For exam ple, materials m ay be defective or non-defective, accep ted or

rejected, go or no go.



Managerial Control

Strategies

For attributes data, discrete distributions such as Binom ial and Poisson are used ttrlnake

inferences about the po pulation characteristic being controlled.

Variables are actually measurable Examples of m easurable quaitities are din~ ension s.

weights, temperature and times. For variable data,

we

use continuous distributions such as

nonnal distribution.

Inspection

Before going into the details of methods to control quality it will be good if we fanliliarise

ourselves with the basic issues of condu cting inspection. The sole purpose of inspection is to

provide infornlation concerning the degree to which items conform to a stx idar d.

Some im portant issues pertaining to inspection are as follows

1 ) How much to? inspect

?

2) At what points the process inspection should be carried ou t

3)

Wh ether to inspect attributes or variables

?

I )

The inspection may range from no or zero inspection to rigorous inspection inspecting

each iten1 many times as is the case with sp ace vehicles, rock ets, etc. But ino st of the

industrial situations lie in between these two ex tremes. i.e, zero inspectioil and 100

intensive inspection.

No d oubt, nlost of situations require some inspection but neither it is possible nor

econom ically feasible to critically examine each and every part or activity.

The amo unt of inspection needed is govenied by costs

of

inspection and expected costs

of passing defective iten s.

I Optimum mount of inspection

F i r e 9 2 Cost vs Amount of Inspection raph

From the graph given in F igure 9.2, it is very mu ch ob vious ,that the goal ofcourse is to

minimise the sum of two costs.

(2) Wh ere to inspect The right answer for this question varies from industry to industry

as it depends on type of prod ucts, types of processes used, m achines used, and the level

of human inv olvement.

Still in practice the following are typical inspection points

a) Raw materials and purchased parts.

b) Finished products.

c) Before a costly operation is carried out.

d) Before an irreversible process.

e) Before a covering process, viz. painting, heat treatment, assembly that may hide

defects.

(3)

Control proced ures will differ depending o n the characteristic of a prod uct or service

that is to be controlled, i.e., attribute data or variable one. For exam ple, actual weight

of cement bag can be measured w hereas numb er of broken bottles in a ship ir~ent an be

counted.



9.3 SOME

B ~ CONCEPTS OF APPLIED STATISTICS

Stat~st~css

the

collection. analysis, inlerpretation ar~d resentation of data.

9.3.1

Methods of Summarising Data

Practical

neth hods

of sumnn~arising ata stress simplicity. Sometimes one method provides a

useful ar~d omplete sumnarisatlon. In ntlier cases. two or even three methods are needed

lor conlplete clarity. Tllree key methods art: t l~cieyuency distribution, histogram, and

measures of central tendency

nd

dispersion.

The Frequency Distribution

The freyuency distribution 1s a statistical tool for presenting numerous facts in a form

which m'akes more clear the central tendency amid di spers~on long the scale of

measurement.

Table 9.1 below shows raw data'' representing the time taken hy a gang of

lnairite~iarlcemen on hundred different occasions for overhauling a particular type

of

machine. practitioner scarllling these

100

facts has difficulty in grasping their

meaning.

Table

9.1

Time in Minutes) for Overhauling a Particular Type

of Machine on

100

Occasions

Table 9.2 shows the same data after tabulation. Note how

to

tallies in the column

Tahulation make more evident where i s the central tendency and what is the

dispersion. The column Frequency is merely a recorded count by these same

tallies. The column Cumulative Frequency shows the number of mcasic>nswith

overhauling time equal to or lesser tlian the associated overliauling time.

Table

9.2

Tally of Overhauling Times

Overhauling

Time

Minutes)

7

Tabulation Frequency

Cumulative

Frequency



Managerial Control

Strategies

Table 9.2exhibits a range of values from 27 to

44

or

17

intervals of

1

minute each.

When it is desired to reduce the number of such intervals, the data are grouped into

cells . Table

9 3

shows the same

data

grouped into a frequency of only six cells,

each 3 minutes wide. Grouping the

4ata

cells simplifies presentation and study of the

distribution but loses some of the details. (How ever, one can alway s

go

back to the

original data, if necessary.)

Overhauling

Time

Minutes)

36

Table 9.3 Frequency Distribution of Overhauling Times

Tabulation

W W ll

The follow ing are the steps taken to cons truct a frequency distribution :

1)

Decide on the number

of

cells :Table 9.4 provides guidelines which are

adequate for most cases encountered. These g uidelines are not rigid and should

b

adjusted when necessary.

Table

9.4 Number

of

Celts

in Frequency Distribution

Frequency

I 3

Overhaul Time

Cumulative

Frequency

77

Frequency

5

Boundaries

26.5

- 29 5

Over JOOO

1 -20

2)

Calculate the approximate,cell fnteroalI

:The cell interval equals the largest

.

observation m inus

the

smallest

observation

divided

by

the number

of

1s.

Round jhis result to some conven ient number.

> - __

Cumulative

Frequency

3

Mid Points

28

Recommended

umber

of

Cells

6

SI.

Na

Number

of

Observations

20

- 50



(3)

onstruct the cells by listing cell boundaries As an aid to later calculation

Manageluento

Quality

(a) The cell bounclaries should

be

to one more decimal place

than

the actual

data and should end in a 5.

The cell interval should constant throughout the entire frequency

distribution.

(4) Tally each observation into the appropriate cell and then list the total frequency

for each cell.

The Histogram

There are several ways of showing a frequency distribution in graphic form. The

most popular is the frequency histogram. Figure 9 3shows the overhauling time data

of Table

9.3

depicted in histogram form.

The

diagram is so easy

to

construct and

interpret that it is widely used in elementary analys is of data.

t

Figure

9 3

lsbgmrn

of

OverhaulT h e

One example of wide, effective use of frequency histograms is comp arison of process

capabilities with tolerance lim its. The histogram of Figure

9.4

shows a pro ess which

is

mherently capable of ho lding the tolerances drawn on the same figure. l ' k h i g h

degree of defectives k i n g produced is the result of running this process at a setting

which does not locate its central tendency near the mid point of the Q l e r m e range.

Lower

tolerance

l im i t

Upper

tolerance

l m t

F i r e

9 4

: istogram

Process

Measures of entral Tendency

Most frequency distributions exhibit a central tendency , i.e. a shape such that the

bulk of the observations pile up in the area betweell the two extremes . The measure

of this central tendency is one of the two mos t fundamental measures in all statistica l

analysis.

There are following three principal measures of central tendency

1) ritbmttlc

Mean

(the ordinaty average ), used for ~ t r i ~r ne r

symmetrical distributions, or for distributions which lock

a

clear dominant

srtlgle

peak.



Managerial ontrol

Strategies

defective etc. that control c h d s have been devised to t~nalyse nd keep track

of

it. Such control charts give the earliest, obtainable wanling of signific:u~t

changes in the central value.

The mean is calculated by adding t l~e bservations and dividing by the number

of observations.

(2)

Median (the niddle value where the figures ;Ire arranged according to slze),

used for reducing the effects of extreme values, or for data which call be ranked

but are not economically measurable (shades of colour, visual appearance, odd)

or for special testing situations. If, for example, the average

of

five parts tested

is used to decide whether a life test requirement has been met. then the life time

of the third part to fail can sometimes serve to predict the average ol all five,

and thereby the decision of the test can be made much sooner.

3)

Mode (value which occurs most often in data), used for severely skewed

distributions. describing

an

irregular situation where two peaks are ti)und, or for

eliminating the effects of extreme values.

Measures of Dispersion

Data are always scattered around the zone of central tendency, and the extent of this

scatter is called dispersion or variation. Measures of dispersion is the second of the

two most fundamental measures in all statistical analysis.

There are several measures of dispersion. The simplest is the range which

is

the

difference between the maxinlum and minimum values in the data. As the range is

based on only two values, it is most useful when the number of observations is small

(about 10or less).

The inost important measure of variation is the standard deviation The definition of

the standard deviation is a following forinula :

where x

it

observation, x arithmetic mean,

= number of observations, and

sample standard deviation.

For calculation purposes, following equivalent formula is used :

The square of the standard deviation is called the variance with data in frequency

distribution form, short cut calculations can simplify finding the average and the

standard deviation. This is illustrated in Table

9 5

To start, an arb~trary rigin is

assumed as 37

Table 9 5 Calculation of Average and Standard Deviation

Mid

Point Frequency V

d

f

d fd



A Lero is arbitrarily placed on tllis line in the 6 colunm.

Thc

other figures in this

Mai~akc ln i~ t~uality

colu~rulndicate how lnany cells the entry is above or below the arbitrary zcro. Minus

signs are attached when the entry is srnaller than thc,assumed value,

37.

The,fil values

in column (4) are found by multiplying together the erltries in columns 2) and

3).

~imilarly.,fld)~s found by multiplyirlg the.figures in colulnns (3) arid (4). Note Urat

the tolals in colurnrls (4) and

6 )

are

identified

in the formulas as

fil

;uid ~ f ( d ) ~

respectively and

I

is the cell interval. As the multiplications are small enough to he

carried out nientally, the complete table can be made quickly.

For snnple sizes of about

10

or fewer observations, the standard deviation can be

R

approximated from the range by calculating

-

where the value of

(12

can be

62

obtained fr on ~able 9.6. For example, suppose thc first colunm of values in

Table

9.1

represents a sample of

10,

The range 1s (40

29)

or 11 From Table

Y.6

d = 3

078.

hc estimate of standard deviation 1s therefore 1113.078

=

3.6. This is

nluch simpler than calculating the standard deviallon directly.

Table 9 6 d Factor for Estimating Standard Deviation

5

Obserrations

-F-T--

1 128

1.693 2 059 2 '326 2

5 34 .704

2

847

2 970

3.078

SAQ

1

(a)

Why quality is gaining importance day-by-day Do you think that quality

skrouldbe treated as a separate functicx~ r it shnuld be under production ptt.

(b)

Dzscribe some of the major areas of

n

organisation w h~c h re being influenced

by quality and how ?

(c)

Dcscrihe various methods of sumnarising data

9 4 QUALITY ASSURANCE AND

QUALITY

CONTROL

--

There are two distlnct quality functions within t h~ t as kf making a product or delivering a

service. The first is quality assurance. This is concerned with the determination

of

methodslprocedures to

be

used and the type

and

frcquency of checks to be made within the

system in order to ensure that the system is meeting the specifications embodied in the

design. The second side of coin is qual~ty ontrol, which concerns with tlie actual checking

or completion of tlle various tasks involved in the quality procedures which have been

determined within t he qual~ty ssurance function. Whilst the latter is often best completed as

a

centralised function due to the test equipment investment and the level of slulled staff

required, the former is normally best undertaken within the operations function and at the

time the job is completed. Developments and the place of quality control within the job of

work can be easily understood through Figure

9.5.



9 5

ACCEPTANCE SAMPLING

ManagerialControl

ne

of the powerful statistical techniques of quality control is Acceptance Sampling. This

technique is generally used in those situations where items ate inspected in batches,

generally known as lots. For example, you may receive a shipment of 10 000electric bulbs

and you may have

to

decide whether

to

accept

the

shipment or return it back to

the

supplie

The acceptability will depend on

the

acceptable quality of the lot, which in

turn

depends on

the use

nd

the price you are willing

to

pay for this quality. Suppose you decide to accept if

the average fraction defective is less than

5

percent. Then to ascertain the actual quality you

may decide to inspect each and every bulb. Such a strategy of 100 percent Inspection.

however, may often be expensive and impractical.

In

such cases a more intelligent way is

to

use

the

concept of SamplingInspection.

sb tcgies

-

The

dea of sampljng inspection is to inspect d y stnall portion of the lot and infer the

quality

of

the

lot,

based

on the quality of

the

sample. Acceptance is based on the

inference

msde from the sample

d hence

the technique is known

s

Acceptance Sampling.

Typid ly a lot is specified

by

its size N) nd the fraction V of defectives that are expected

tobe

present

(at th most) in

the lot. The principles of statistics are used in the inference

process.

orrTprrTs

Interestingly

ih

conceptof acceptancesampiinb is no different from the strategy adopted

by a typical housewife who decides whether

or

not a pot-ful of rice is cooked by inspecting

just a qm nf ul

of

grains.

Two

things must

be

kept

in

mind.

n

order that sampling inspection might work, the sample

must

be

npmitnlrtlve of

the

lot. Typically this is ensured by choosing the sample at

random

so that

every portion

of

the lot har equal representation

in the

sampie. Such a

sampling is

known

asRqndom Sampling.

Second, a sample is

only

representative and not

identical

(in

chatocteristics) with the

lot.

In the inference process, therefore, a few

good

lot

wili be rejected nd a few bad lots will

be

accepted. We can control such samplingerrors,

but they cannot

be

eliminated.

n

fact in the design of

sampling plans

we will ensure that

the errors

are kept

below

certain

acceptable levels.

ControlQlvts

Activity

T

Think of a common situation

where 100

percent inspection is

(a) imptdcal

Sampliag

Figure 9 5

S Q

2

D~tferent ia te

etween quahty

assurance

and

quallty

control

b)

impossible.

\



9 5 1 SamplingPlans

We will first consider a single sampling plan in which acccptlreject decisions are based on

the results of a single sample of n items from the lot of items. Each of the sample items

is inspected and categorised

as

either

acceptaMe or defective.

Such

a

plan is known as

Samplingby Attributes.

(We will not discuss Sampling by Variables in this unit. The

interested rhders may refer to the Further Reading Section given at the end of the block.) If

the number of defective items in the sample exceeds a pre-specified cut-off level, c, the8

entire batch is rejected. (Depending on costs, a rejected lot may

be

saapped, 100 per cent

inspected

or

returned back to

the

manufacturer.) Since a finding of

c

or fewer defective

items in the sample implies accepting the batch, c i s often referred to as the acceptance

lever.A sa~npling lan is specified by the values of n aid

c.

The sampling plan is supposed to separate good lots from bad lots. As mentioned earlier

there are bound to be sampling errors. We will now study the probabilities of such error

graphically, using an

Operating Characteristic Curve.

TheOperating Characteristic Curve

It is useful to have a si

le picture that allows us to compare sampling plans as to

how they will react to

;P

fferent lots with

unknown varying

fraction defective. Such

a comparison is provided by the Operating Characteristic Curve (OCC) which

displays

the

probability of accepting

a

lot with

any

fraction defective.

C

lncominq ~ r a c 6on r f r c t i v r

Figure9.6 : peratingCl~aracteristic brve

Figure 9.6 shows OCC for twasingle sampling plans A and B with

=

3 5 c 1 hd

rt

i

150,

c=6,

respectively.

For

example, suppose that a lot wit

F

= 10 percent

defectives is considered to be a bad lot and a lot with =

2

percent defectives is

considered to be a good lot. From Figure

9.6,

it is clear that sampling plan A would

s W a 14 percent chance

of

accepting a bad lot. The same unfortunate error can

occur with the sampling plan B with larger sample size

also

but the probability of

error is much smaller. In fact,

it

is only 1percent. The sanipling plan B is also better

,

- at not rejecting goads lots

(f= 2

percent). Sampling planA has 1

6

percent chance of

rejecting

a

good lot whereas plan B h'as only 3 percent chance of rejecting a good lot.

It is not surprising that a larger sample does a better

job

of discriminating between

good and bad lots. It has more information. However, the price

for

increased

accuracy is higher inspection costs. The design of a sampling plan has o optimally

trade off cost with discrimination.

Vie values of thiordinates of the Operating Characteristics Curve are determined

from tlie Poisson Disbibution. The actualdetails canbe found in the advanced texts

listed in the reference. -

'At

this

moment,

pause

for

a

while and check for yoarself whether

you

ha;e und~stood

OCC.

Do the following Activity.

Activity

(a)

ow

will the

OCC

change shape

s

(i) n is increased, keeping c comfant

(ii) c-is increased, keepilig constant.

nnagement of Quality



Mallagerial C ontrr~l

Strategies

(b ) What will be tile linlitillg shape of the OCC

c ) Can you interpret your iulswers to

3)

i u ~ db) ?

9.5.2

esign of Single Sampling Plan

You ha ve to design a sampling pl m (n,

c

hat has an OCC that meets certaiil pre-specified

requirements. Generally, the design is based on the following criteria that are related tu the

probability of ma king eiU~ er f the following errors : accepting a bad lot (P) ai d rejecting a

good lot ( a) . The criteria are established subjectively and ultimately should reflect the cost

of accepting a bad lot or rejecting a good lot. Needless to say, before and values can be

,

specified. one has to decide what is a go td lot and what is a bad lot. Invariably, this is done

by specifying tl ~ eowerlupper linlits of fraction defective 0 s illustrated below :

AQL (Acceptable Quality Level) the fraction defective n that the user conhiders

acceptable. Thus if a batch were know n t o have a fraction defective equal

t o AUL

it

should not be rejected.

LTPD (Lot Tolerance Percent Defective) the fraction defective that defines I bad lot

or (me that should be rejected. Ofcourse AOL must be less than LTPD

Producers Risk ( a ) the largest allowable probability of rejecting a good lot (due to

statistical error). Note that a go od lot has fraction defective less than or eclual to AQL

(generally, 5 percent).

Consum ers Risk ( P) the largest allowable probability of acc epting a had lo1 (due to

statistical error). Note

that a bad lot as fraction defective greater than or equal to

LTPD (general]y 10 percent).

Example 9.1

Consider a manufaclurirlg situation with the following va lues

AQL = 0:02

= 0.05

LTPD = 0.10 = 0.01

Solution

From Figure 9.6, you can verify that sampling plan A

n=

35,

=

1) has a probability

of acceptance of 84 percent for a fraction defective of 0.0 2 (AQL ). In other words,

this plan has a 16 percent chance

of

rejecting a good lot. Similarly, it has a

probability o t 0.14 of rejecting a bad lot with

f

=0.10 (LTPD). Since both tl ~e alues

are l~igbe rhan the allowed va lues of 0.05 and 0.01 , respectively, this sanipling plan

is not acceptable. Only larger values of

n

can yield better discrimirlatiilg power.

The sampling plan

B

n

=

150, 6) has the probability of acceptlilg a good lot by

97 percent and probability of rejecting a bad lot by percent. In other words,

i t

has a

value of 0.03 and fi value of 0.01. This discriminating power is eve11 nlore than

what is needed. Tlie plan is acceptable but it inay be possible to get the required

discriillination with smaller sample size and in turn with a lower inspectioil

ost

One w ay to decide the optimal sampling plau is to search through several sainpling plans

with

n

values between 35 and 150 and select the one th at matches

a

and

values more

closely. An easier way is to use ThorndikeChart (Table 9.7):This chart call he used for

(a) Plotting )CC , and

(b) Designing a Sampling Plan.

We will illwtrate the design of sampling plan using this chart. Before we nlovt. t o this to pi^

ensure that Plan A doe s not nleet the requirement and Plan B mee ts m ore tll;u~ lltl

requirements, by following the arguments given earlier.

T o design tlle

sampling

plan follow the ~nstn rcti ol~t the bottom

of

the chart. Note that we

have read off p = 0.01 and 1 a )

=

0.95 rows o nly. We first need to filld a

c

value lor

which

LTPD

&-

a AUL



For

this

problem,

LTPD 0.10

- z - = 5 .

AQL

; 2

Management

of Quality

Starting with

c 0

we read off

pg

4.6052 and

pa

0.0513 and so

5 is not

P a

14.5706

satisfied. Continuing with c 1,2 ,3 ,4 ,5we f;nd that for c

6,

5.

Pa 3 2853

Hence, we choose

c

6.

Table 9 7 :Thorndike Chart for Single Sampling Plan

1 )

To plot an OCC curve for a given sample plan (n, c) :

(a) Find the ccllumn for your

c

value.

(b) Divide each number in that column

n

The results are thef values for the horizontal axis.

(c) The

P

values, for the vertical axis, are in the

f r

left column.

Acceptance

Probability

p

rows, entries

d e n o t e d ~ ~

I

-a ows,

entries

denoted pa

(2) To find a single sampling plan :

LTF D

(a) Find

c

for which

QL.

Acceptance Number c

(b) Then choose any

n

between

p

Cb

X , d n , = -

LTF D AQL

(3) To find the acceptance probability for a given n, c andf :

(a, Multiply n) f.

(b) In the appropriate c column, fmd values above and below nf:

(c) In the Pn column read upper and lower limits for P in two rows from step (b) (interpolate, if

you wish).

To get the value of nithe limits are

np =

AL md

na

TPD

Pa

Reading off the table,

AQL

Po

0.010

0.025

0.050

0.100

0.200

0.500

0.800

0.900

0950

0.975

0.990

we get,

2

Hence, sampling plans with size in the range of 146 to 165 will satisfy the requirement. The

exact values of and for

any

sampling plan can be determined using Thomdike chart

again. The exercise at the end of the unit will give you

an

opportunity to design any other

sampling plans and decide the exact values of Consumer s Risk and Producer s Risk.

p

nf

expected number defective in the sample

9 5 3

Average Ou tgoing Quality

The inspection process rejects lots with high fraction defectives. After rejection either you

may stop,

or

you may continue the inspection of all the items in the rejected lot and all

defective items are replaced with good items. Such a policy is known as

Rectifying

Inspection

3

4.6052

3.6889

2.9957

2.3026

1.6094

0.6931

0.2231

0.1054

0.0513

0.0253

0.0101

6.6383

5.5716

4.7439

3.8897

2.9943

1.6783

0.8244

0.53 18

0.3554

0.2422

0.1486

5

8.4059

7.2247

6.2958

5.3223

4.2790

2.6741

1.5350

1 I021

0.8177

0.6187

0.4360

6

10.0450

8.7672

7.7537

6.6808

5.5150

3.6721

2.2968

1.7448

1.3663

1.0899

0.8233

7 8

1

11.6046

10.2416

9.1535

7.9936

6.7210

4.6709

3.0895

2.4326

1.9701

1.6235

1.2791

13.1085

11.6683

10.5 130

9.2747

7.9060

5.6702

3.9037

3.15 19

2.6130

2.2019

1.7853

14.5706

13.0595

1.8424

10.5321

9.0754

6.6696

4.7337

3.8948

3.2853

2.8144

2.3302

16.0000 17.4027

18.7831 20.1447

14.4227 15.7632

17.0848 18.3904

13.1481 14.4346 15.7052 16.9622

11.7709 12.9947 14.2060 15.4066

10.2325 1 1.3798

12.5 188 12.6507

7.6692 8.6690 9.6687 10.6685

5.5761

6.4285 7.2892

8.1570

4.6561 3.4325

6.2213 7.0208

3.9808 4.6952

5.4254 6.1690

3.4538 4.1 154

4.7954 5.4912

2.9061 3.5075 4.1302 4.7712



Managerial C ontrcrl

Strategies

In rectifying inspection, all outgoing lots consist of

N

itellis either accepted ones or rejected

ones. Suppose ?a lot has inconung fraction defe ctiv ef. If it is accepted N n tems remain

uninspected. We. therefore, expect,f N n) defectives in the accepted lots ;lssunling that

the defectives found in the sample are replaced with good ones). In contrast,

if

it is rejected

and hence 10 0 percent rectified and inspected) there are n o defectives. Thus

if P,,

is the

probability that the sanlpling plan will accept the lot,

Outgoing fraction defective =

Pa) f l N n 1 P,,) 0)

N

f

N - n

P , ) m (

N

=

P , I V ,

A plot

of

ongoing fraction defective against incoming fraction defec tive

f )

is

generally

called the Average lutgo ing Quality AO Q) curve. Figure

9.7

shows the curve

tor

silrnpfing

plan B n = 150, = 6) o f the earlier exam ples. This curve h as a surprisiilg property th;rl as , f

increases, there conles a point at wh iih the outgoing fraction defective actually begins to

improve . The reason being that the sanlpling plan rejects m ost bad lots and they

are

rectified

through 100percent inspection.

Figure

9 7

:AOQ Curve for Sampling Plan

B N =

10,000

The

most critical incoming fraction defectlve.1 gives the worst outgoing quallty )n the

average, the value of that critlcal

f

is not ~ln po rtan t ut the correspo nding outgolng tractlo11

defective generally know11 as Average Outgoing Quality Limit AOQL) IS extremely

useful.

No

matter what the

incoming

fraction defective is, the long-run averagc ou tg o~ ~ lg

frac tio~ l efective wlll not be worse than AOQL.

You do not have to de te rl in e AOQL by plotting the AOQ. One can use the following

AOQL factor given 111Table 9.8. To get the AOQL simply use the formu1;l

AOQL AOQ L factor)

k- 1

Table

9.8

:AOQL Factor

Acceptance Number (C)

0

1

2

3

4

5

6

8

10

AOQL Factor Y )

Convince youself that you understand the mechanism of deter~niuingAOQL y doln g tlle

following activity.



Activity 3

(a)

Using OCC for sampling plan B , draw the A( )QL curve over the range of

0 10 percent fraction defective and determine AOQL.

b) Verify your answer to (a) using AO QL factor in Table 9.8 and the formula.

anagement

of

Quality

Many managers prefer to use AOQ L as a criterion for designing a sampling plan rather than

trying to decide on values of AQL,

LTPD,

and P Dodge and Rom ig (1959) present tables

that are designed for this purpose. In fact, these tab les meet the requirem ent of a specified

AOQL

and

minimise the expected number of items inspected per lot.

The OCC approach and A OQL based app roach are but two of man y other approaches that

can be used to design a samp ling plan. Choice among them is a match of personal

experience, the exact situation and the objectives of the organisation. Regardless of the

approach, all sam pling plans have both an OC curve and A OQ curve, so the principles

discussed in this unit can be used to eva luate any sampling plan .

9 5 4

Double

and-Multiple Sampling

Extensions of the single sampling plans to double and multiple sampling plans are also

available. In a double sampling plan, after the first n samples have been inspected there are

three choices depending on the number of defectives found

:

1) reject the lot,

(2) accept the lot, and

(3) draw a second sample of n tems.

If choice 3) is made the final accepthe ject decision is made on combined sample of

n

n;

items. A multiple sampling plan operates in the same way, but with more than two sam ples.

Double and multiple sampling plans reduce inspection costs because many accepdreject

decisions are made based o n the first sample which is sm aller than that of the single

sampling plan. How ever, single samp ling plan is more comm on and easy to use. Details of

multiple sampling plan are found in references.

9 6

PROCESS CON TROL

9 6 1 Vari ability

All products and services have a certain amount of natural variability because of Variations

in the input as we ll as imperfections in the process. For exam ple, different quality of raw

mate r~als ould have been used and different quantities of chem icals could have been used

in the process. This

process

variability may be measured by the process standard deviation

o hich indicates how much the products will vary even if the process is in control.

.

Noturol

P r q c q

~ p r c i t i r d

r

Level

m

Vorrob~lt ty Tolrroncr

Figure 9 8

rocess

Variability



Musagcrial :011tro1

Protlucts Ii:lve to nlecl spc cifi cd to1er:lnces lmposctl by their intendetl use. Ac zo ld i~ ld ).ht

Stratcgics

~ia turn l ;~riahility rlust he suhsta nli;~lly ni;lller th;ui the specified loler;~ ncc. his is sliown

and expl;liiictl ill Figu rc

0 X

in whicli t l i~ , e11tr;ll l i~ ies the desired average 01' thi. [~~-c c.t'sc

;uid the tliisllcd lines itre, tlie '?-Sigma IinLirs' represenling the narural procehs v;lr~i~hil~t\;.

I t can hc shown

r l ~ i l t

v a r i i ~ ~ i o n s

i

nlosc 11i;ul

?a

ronl tlic proce ss ave rage arc very unlikel)

Iri f;~ci.t is i~bo ut

.25

percclit i l- he proccss lo llows tlic no r~ na l istritwlion iu~ tiltll'ir~ilcly

less Ll~aii ;; lor

1110~1

processcx . Th e so lid line's repl-estbnl lie toler anc es spc,citiit(l hy tllr

uitencled use oi (he proJuc\ .

Wirliill lhe x pecificd lOli.r;illccs.

a

cerl ; t i~ilniount of process vi~r iahility s

1

hese x p c c t ~ ~ ~ l

However. i l is

111~

o a l o f l l ~ e ta t i s t i ca l p rocess con t ro l lo c lc ter~~i i l i chen

thi

proccss

viuiilbilily is gcttitig out of ha ~ id , o that correc tive iiclion c;ui he take n, prekrnhl . before lhe

requ ir ed l o l e r a~ cesIre viol i~ lrd. l iis is pei~ernl ly chieved by a Co nt ro l ( :hart .

9 6 2 Control Charts

In order

B

provide r;~pi t lcet lback to an ongoing production process . methods son ~cw h;lr

different froni acceptal lce sam pling arc appropriate. Sa n~ pl es

re

lake11

as

s o o ~ lis t l~c y rc

available, rather than waiting for tllc co~ splc tioli f a lot. This affords the c ~pp ortun ily t;

tlcrect unp1;ilined cliiuiges in Ulc prtKe ss. shorlly after they o cc ur an d ti~ kt.I quic.k acliou.

suc h as ;tc ju.;ri~ig lic m;lchiiie.

The niost comliioii devicc used fo r Illis purpose is She wa rt Control C liiut ili lroducctl i l l

1931. Tlic con trol ch ar t is a visu;i l display

o f

th e r es ul t o f i ~ on sp el io n proc.c,ss

incorporaling carefully dcrivetl l inlits to indic;lte unusu;~lhcbaviour. A control cliall cirli

bc

based on ci lteporical i r~ or ma tio i i r aclual n~ei lsurenicnt .Accordingly. they

;LI L:

ci~llccl

co a t r ir l ch a r t f o r v ; ~ r i ah l e slritl con trol ch ar t for at t r ib ute s . Since conlrol c.h;rl\ lor

viuiables are t ilore coinnlonly used a id more powerful . w e will desc ribe t1ic111 ir \( .

9 6 3

x

and R Cllarls

Consicler X Y Z konil,;uiy that uses iui ilutoniatlc ma chin es lo fill kg box es

ot

s u w r ~ . 'he

to1cr;ulces ;Ire spccil'ictl as 1.000 k g

011

tlie low er side (1cg;ll requ irenie nr)

i u ~ t l

.005 h y 011

the higlier side (n o point w ;~slilig ug;lr). Since the spread is only 0.005 llley sclec tPd :I

~ n a c h i n ehat has

a

~ ~ a t u r i l lrocess variability of =

0 0003 The

3 - s i y ~ n aimits t tlli,

nlilchine tlierel'ore arr

3

(0.0003) = 0.00 09 kg ahove an d helow t l ie niean. The . ; p rc ;~~ls

0. 00 18 wliicli is liarrower (11;ln Ihe s pcc itietl toleriu~cc: f 0. 00 5.

They

a t l i u s t l l l e u ~ i ~~h i i i c

o

fill hoxes with

rill

iivc'riige 1.00 25 kg liiilf wily hetwcc li tlie toler anc e liiiuts.

A n W Ch ar t ( ,A v r~ , ay e'r>ii~olCll i~ rt ) as set-up to (lelect

whcn

the I ~ ~ L C ~ I N Cots OUI of

control.

In

ortles to reduce \lie nal ur;~ l rocess va liahility. saniplcs of = hoxes w ~ ~ c

weighed, and Iic i\VCrllgC weiglir

per

box W ii s recorded for each sample . F ' I ~ u I- e0

s h ow s llle c o ~ ~ t r o l1i:lrt used for this niacliine and tlie fou r points p lotted ou he clial-t

represent the

X

values fr om li)ur samples i a total of twe nty

I

kg boxes). We sliiill

L ' X ~ I I ~ I ~ I

the iletails of the c l~ iu t .

U p p p r c o n t r o l ~ i r n i t o r

r -

c 1 . 0021

.

1

L o w e r

Control ~ i m i t o r

R

Figure

9 9

X hart

Tliere is

;m

apparen r trerlil

in

X.

l ie sa r~l ple s een1 to he getting prog ressively he;rvlcr.

However. appewdices not\vithsl:uidil~g,

he

trend in

l l~e

igure 9.9

may

be X t luc Io l ; ~ ~ l d o m

f l ~ ~ t ~ a t I O 1 1 ~ .1

1s

tlllc

1 P ncnn t t ~ n t ~u ~ r tt ~ r . r ~ r n r l r a t c lho

rnnr ont

o

r t - t i r t i o r . 1

r ; , r n ; 4 ; ~ ~ a - a . . .



in our discussion The standard deviation of the sample average

( O F )

is expressed Uirouph

Managenlent of

Ut~al i ty

the fornlula,

Therefore. t l~ e verages of n= boxes ot sugar should have a standard deviation

of

0.0003

2 = 0.000134

kg

The conlrol limits 111 Figure 9.9 represent 3-sigma litnits and are 3) 0.00 0134) = 0 0004

above ruld below the intended average of 1 0025 f a sample a verage falls outside these

control limts

the

deviatiori from the process average is statistically significant.

Tlie fourth sam ple X is

o

he control limit and therefore there is a strong statistical

evidence that Inore sugar is put into the boxes than what was intended and the machine

needs adjustment.

The

R

Chart Rang e Control Chart) is also used to control the prcxesses. The range of a

sample is the largest value m n u s Uie smallest. A n

R

chart is appropriate if process

solnctinles goes out of control in such a way that there 1s ~ncon sisten cyn the values, but no

s h ~ f tII the mean value of the process. For examp le, a worker who is basically go od might

produce an Inconsistent set

of

sizes of sorne ~rlanutactured roduct) when he is fatrgued.

The

R

chart used to plot the data of

XYZ

Comp any appears in Figure 9.10.

U D D ~ ~

ontro l

~ i m i t

ar

R=D.00448

.R=0.0007 .

Lower Contro l

~ i r n i t

o r

R 0.0000

Figure 9.10

C art

The co ntrol lilnitq for

K

chart are determined differently. There is n o need for lowe r control

lirmt whlch

is

generally for n

< 6

it is zero) zcro. Table 9 9 is used to set the upper control

limit for

an R

chart and factor

D

fro111 he table is l~iultiplied y the process standard

dev iat~o n, n the case of

X Y Z

Company data, D = 4.918 for a sample size of 5 m d

hence, upper control limit

=

0.0003) 4.918)

=

0.00148. As can

be

secn in Figure 9.10, no

significant shifts are prcsent.

Activity 4

Verify the control chart for the exam ple given in Figure 9.10. The data regarding the

sample is as in Table 9.10.

Table 9.9

R

Chart Factor

1 C .

Sample Size

9

10

n

3.686

4.358

4.698

4.9 18

5 078

5.203

5 307

5 '394

5.469



Managerial

Control

Strategies

4

Average

Range

Table 9 10 Weights of Twenty kg

Boxes

ample 1

Determining the

sample size

is an important decision. It is common to use

= 4

or

n =

in

order to obtain low cost feedback. Large samples such as

n = 15

or

20

are necessary if the

process standard deviation is large. To a large extent this will depend o n the rate of

production, convenience and other considerations as well. Similarly, it is important to

decide the frequency of sampling. It should be in general proportional to

the

average

frequency of o ut-of-control conditions. It is fairly complex decision t o make if one were to

lo ok for some optimality. Generally, it is decided by convenience.

Sample2

9 6 4 Other Control Charts

Sample 3

There are several other control charts, including the p-chart which is used to control the

process when the measurement is by attributes. In other words, the decision is only to decide

whether or not the sample item is acceptable. No measurement is taken. For example in

using the GO/NO

GO

gauges one gets only such a m easurement.

The pc ha rt is based on the fraction defective, p, in a sample of

n

items. If yo represents the

normal process defective i.e. when the process is in control) then the 3 0 control limits are,

This is based on the fact that the number of defectives has th e binomial probab ility. The

control chart is used just like an hart, excep t that the fraction defective is calculated

rather than X for each sam ple of

n

items and a lower control limit is often omitted.

Sam ple sizes are typically larger for p-chart than for X charts. Since the information content

of a yeslno measurem ent is much smaller than the actual variable m easurenient, it can only

be expected. In fact, the required sample size can be computed approximately from the

following formula

:

In this formula

p

is the normal process fraction defective, p is the specified fraction

defective that is unacceptable.

ya

like AQL a n dp l is like LTPD in acceptance sampling.)

For example, consider ABC Company that makes ready-made shirts. It has been found that

4 percent of the shirts are defective when the process is under control. ABC Company wants

to be able to detect

a

shirt to 12 percent defective o he basis of one sample of n items. The

formula for n su ggests a sample size of n = 197.The upper con trol limits w ill be

S Q

3

What is the fundamental difference between the use of acceptance sam pling plrulx

and process control chatrs



9 7

A NOTE O N VARIABILITY IN MANUFAC TURE

Whe n the variability presen t in a production process is limited to chance var iation, the

process is said to be in a state of Statistical Control . The state of statistical control is

at tm ed by avoiding of another variation, called as assignable variation. An assignab le

variation may be due to the extraneo us cause s like defective manoev ring of thc mach ine,

poor quality, raw m aterials, man-m ade error and so o n. Since assignable variation may

contribute substantially to the deviations from the state of statistical control, a systcm atic

method of

dete ting

and eliminating deviations before or during the cou rse of a

nlanu facturin g prcxess is desirab le. Thus the act of getting a process in statistical control

involves the identific ation and elimina tion of assignable variatio n.

A

production process is said to be in a state of statistical control if it is gove rned by chance

variations alone and the assig nable variations are totally absent in the process.

The main objective in any pro ductio n proce ss is to control the quality of the finished

product so that it conforms to specifications.

9 8

QUALITY FUNCTION A ND ITS ASSURANCE

The existence of a comp'my, whether a m anufacturing industry or a process industry or a

job order shop , depends on the incom e that it rcccives from selling of its products and

service s. Again this selling is depen dent on the fitncss for use dictated by custom er.

Therefore, the company functions concerned with quality or achieving fitness for

customer's use are known as quality functions. It includcs the complete journey of the

product right from its conceptio n till it reaches in the hands of custom er. Therefor e, it is said

that quality is not the baby of qua lity control departm ent rather it permeates through each

and every activity of organ isation either directly or indirectly .

The journey of the product bcgins with the mark et survey and market research data to probe

what does customer want out of a product.

Research and development specialists try to create a product which can meet custom er's

stated and implied need s. Depe nding upon the advice of R&D persons, process and product

design engineers prepare product and m aterial specifications considering the quality

rcquirements. Sim ilarly, process engineers specify about the process, machines and the

tooling/instrumentation capable of produ cing the products with requis ite quality.

Purchasing specialists buy materials and c ompon ents processing appropriate qualities after

careful selection of vend ors. Operato rs are also trained to use thc process and tools and

equipme nts so as to ensure that the products are as per desig n and would satisfy custome r's

requirements.

The final judge is ofcourse custom ers. He is king of busines s, therefore feedback either in

the form of complaints or suggestion is carefully tapped and is used as a basis for redesign,

modification or alteration resulting in im provemen ts. Thus the cycle is repeated.

Quality assurance deals with ensu ring the desire d quality , durability and other aspects of

product (perh~rmance,ield complaints or other characteristics) through scientific

engineering and man agement techmqu es. The job of critical evaluation of com pany's

activities with respect to quality reporting the results of eva luation to all conc erned for

information and necessary acuon and subsequen t action taken is known

as

Quality

assurance function .

Obviously, there are three distinct stages in consideration of total quality of product; related

to quality assurance.

1)

esign stage assurance (when the product is on papers, conceptual level)

:

The

idea is to assure that qua lity of designs and en gineering detailing is according to

needs of the customers.

2) Manufacturing stage (transformation stage when the product is ch anging from

conceptual stage to physical reality) : To assure that whether the product

manufactured conforms to design specifications.

(3)

Field stage

(when the product is in the hands of user or user's reactions while

using that products) : t is associated with observing the performance in actual

field, analysing and critically reviewing the performance, maintenance aspects,

life, with a view of im prove men t in the products in next gener ation if any.

Managemr~itf Quality



Managerial

Contrcrl

Strategies

Quality of assurance Assurlu~ce t the design stage (conceptual stage)

Assurance at manufactunng stage (transfor~uatlot~tage)

Assuring quality of performance (field stage or use stage

In short, assurance activities are related with presenting to the management the su~runxlse

facts on what is happening with the quality function. So that corrective actions c;ul be taken

if there is any deviation from the desired level. Quality assurance, thus, helps to inamt;un th

quality of Ule product5 and hence, the reputation of the firm also gets inlproved with gvod

customer relations.

9 OVERVIEW OF IS 0 9000

IS0 90004s perhaps the most talked nlanagement concept today. Ofcourse media must be

given due credit for it but the message has already splasl~ed you will h iiv~o delnon~trat

your ability to provide Consistently high quality products and services. Registration

conforming lo Irltenlational Organisation for Standardisation (ISO) offers such a mechimlsn

Almost every quality conscious colnpany is engaged in the quest for I S0 9000. Eitller it ha

adopted certification and wishes to retain it or is in the process of getting it.

Iso

is a greek word n~eaning qual, like Iso Uienllal, Iso

bar

etc. This was taken as a prefix

to a serres of standards developed by Intematirnlal Orgallisation for Standardisation, ISO. I

is worldwide federation

of

Illore than 50 national standard bodies or bureaus

including

Indi

thmugh BIS (Bureau of Indim Standards). Its headquarter is in Geneva, Switzerland.

The basic objective behind the fon~latioil f I S 0 was to meet the needs for worldwide

harmonisation and ratit>n:~lisation

f

terms, standards and methodologies used for assessing

capability of suppliers to fulfill customer s requirements. Thus, the idea was

to

bring

trmsparellcy in the field of Quality Management for the benefit of both mternal quality

management as well as for the customers to see.

IS 0 9000 is a series of six iuternauonally accepted standards to guide and to audit

company s quality managerneilt practices, documents and records. These stan(1ards were

issued in 1987 by IS(

in Geneva and were named as IS() 9000, IS( 9001, IS()

9002.

IS0 9003 and IS0 9004. One inore IS0 8402 is for definitions. Figure 9.1

1

clears the

structure of IS(

.

130-8402

i t

emunology or ~loss nrytandard

Gives definition of all the terms)

Non Contractual

Situation

Basic set of elem ents that a supplier would like to

know about internal documen tations. Quality

Malagemeut and system e lement. :zu~delines.)

Road map to 4 other quality systems

Guideliues for selection and use)

IS .) 9001

Model fo r quality

assurance in

design/development.

production, installation

and servicing

I S 0 9 0 02 I S 0 9 3

Model for quality

Model for quality

assurance in production

assurance in fiual

and installation inspection and tzst

Figure

9 11

S

Structure



Following

1)

are important benefits of IS(.) certification.

Transparency in the system hoth from external custoiner point of view as well

as from internal custom er point.

Better documentati?n of records, facts and pertaining data, therefore

stre mlined

records.

International recognition of ability, credibility and expertise in the

demonstration of consistent quality of products, hence, increase in number

of

customers, i.e. repeat business.

Com petitive edge in the donlestic and global

market.

Culture for con tii~u ous nd consistent iinproveinent in quality.

Change s the attitude of the work-force. the result is mproved house keeping,

work tmosphere arid quality awaren ess.

Encourag es the feeling of doing right things at right time thus no rework or

nothing for rectification.

Process of quali ty i~ npro ven ~en ts maintained by reducing wastes ruld repairs

enhancing profits.

Sapper without I S 0 certification may face higher insurance rates or be denied

insurance in some markets.

9 10 SUMMARY

Quality is fitness for use. Using this definition leads to many o pporty lities to im prove both

quality and productivity sim ultaneously. M any quality problems have cau ses that cross

departmental boundaries. S o a good quality management sys tem must ma ke the entire

organisation responsible fo r quality. Statistical methods are import'mt tools for quality

control. They separate rand om variations from real assignable causes of deviations from

normal. Acceptance sam pling helps in deciding the qua lity of large batch (lot) from an

inspection of sinall sun pl e. The operating cha racteristic curve precisely gives the risks

associated with any sampling plan. The design of a sam pling plan can be based on

Operating Ch aracteristic Curv e as well as Average O utgoing Quality. Control charts display

the results of inspecting a continuous process. This provides conven ient and rapid feedback

suggesting when feed back, overhaul or adjustment, may be needed . Th e design of control

chart is based on sound statistical principle regarding the behaviour of sam ple mean.

The statistical methods described in this unit are used wideJy in n~an ufa ctur ing nd servic e

industry. They are also the basis fo r many of the commonly u sed, yet nlore complex ,

scheme s described in the references.

Ma~lagenlent

f

Quality

9 11 ANSWERS T SAQs

Refer the relevant preceding text in the unit or otner useful books on the topic listed in the

section Further Reading to get the answers of the SAQs as well as guidance for

performing given activities.

Date post:	10-Mar-2016
Category:	Documents
Upload:	jayaneela-prawin
View:	4 times
Download:	0 times

Management of Quality

Documents