质量管理第六版英文教辅 foster-mq6-inppt-11

Chapter

11Statistically

BasedQuality

Improvementsfor

VariablesChapter

Objectives•Copyright

©

2017

Pearson

Education,

Inc.11-2Statistical

ThinkingCopyright

©

2017

Pearson

Education,

Inc.11-3Statistical

thinking

is

a

decision-making

skill

demonstrby

the

ability

to

draw

conclusions

based

on

data.Statistical

thinking

is

based

on

three

concepts:All

work

occurs

in

a

system

of

interconnected

processes.

All

processes

have

variation

(the

amount

of

variation

tends

to

beunderestimated).

Understanding

variation

and

reducing

variation

are

important

keysto

success.Why

Do

Statistics

Sometimes

Fail

in

theWorkplace?Copyright

©

2017

Pearson

Education,

Inc.11-4A

lack

of

knowledge

about

the

tools

leads

to

tools

beingmisapplied.General

disdain

for

all

things

mathematical

creates

a

naturbarrier

to

the

use

of

statistics.

When

was

the

last

time

youheard

someone

proclaim

a

love

for

statistics?Cultural

barriers

in

a

company

make

the

use

of

statistics

focontinual

improvement

difficult.Statistical

specialists

have

trouble

communicating

withmanagerial

generalists.Why

Do

Statistics

Sometimes

Fail

in

theWorkplace?

(cont’d)Copyright

©

2017

Pearson

Education,

Inc.11-5Statistics

generally

are

poorly

taught,

emphasizingmathematical

development

rather

than

application.People

have

a

poor

understanding

of

the

scientific

method.Organizations

lack

patience

in

collecting

data.

All

decisiohave

to

be

made

“yesterday.”Statistics

are

viewed

as

something

to

buttress

an

already-held

opinion

rather

than

a

method

for

informing

andimproving

decision

making.Why

Do

Statistics

Sometimes

Fail

in

theWorkplace?

(cont’d)Copyright

©

2017

Pearson

Education,

Inc.11-6

People

fear

using

statistics

because

they

fear

they

may

violatecritical

statistical

assumptions.

Time-ordered

data

are

messy

andrequire

advanced

statistical

techniques

to

be

used

effectively.

Most

people

don’t

understand

random

variation,

resulting

in

toomuch

process

tampering.Statistical

tools

often

are

reactive

and

focus

on

effects

rather

tcauses.

When

either

type

I

or

type

II

errors

occur,

erroneous

decisions

aremade

relative

to

products

that

can

result

in

high

costs

or

lostfuture

sales.Why

Do

Statistics

Sometimes

Fail

in

theWorkplace?

(cont’d)Copyright

©

2017

Pearson

Education,

Inc.11-7Type1

errorProducer’s

riskProbability

that

a

good

product

will

be

rejectedType2

errorConsumer’s

riskProbability

that

a

nonconforming

product

will

be

available

for

salUnderstanding

Process

VariationCopyright

©

2017

Pearson

Education,

Inc.11-8All

processes

exhibit

variationSome

variation

can

be

managed

and

some

cannot

be

managed.Types

of

process

variation:RandomNonrandomRandom

VariationAlso

called

common

causeCentered

around

themeanandoccurs

with

asomewhat

consistentamount

of

dispersionUncontrolled

variationMay

be

either

large

orsmallFigure

11-1Copyright

©

2017

Pearson

Education,

Inc.11-9Nonrandom

VariationAlso

called

special

causevariationResults

from

some

eventwhich

may

be

a

shift

in

aprocess

mean

or

someunexpected

occurrenceDispersion

and

average

ofthe

process

are

changingProcess

is

not

repeatableFigure

11-2Copyright

©

2017

Pearson

Education,

Inc.11-10Process

StabilityCopyright

©

2017

Pearson

Education,

Inc.11-11The

variation

that

we

observe

in

the

process

is

randomvariation

and

not

nonrandom.Process

charts

Graphs

designed

to

signal

process

workers

when

nonrandomvariation

is

occurring

in

a

processSampling

MethodsCopyright

©

2017

Pearson

Education,

Inc.11-12Reasons

why

sampling

is

used:

Samples

are

cheaper,

take

less

time,

are

less

intrusive,

and

allowthe

user

to

frame

the

sample.

If

quality

testing

is

destructive,

100%

inspection

would

beimpossible.Sampling

MethodsCopyright

©

2017

Pearson

Education,

Inc.11-13Reasons

why

100%

inspection

is

used:

When

a

lot

of

material

has

been

rejected

in

the

past

and

materialsmust

be

sorted

to

keep

good

materials

and

return

defectivematerials

for

a

refundWhen

employees

perform

their

own

in-process

inspectionSampling

MethodsCopyright

©

2017

Pearson

Education,

Inc.11-14Random

samples

To

sample

in

such

a

way

that

every

piece

or

product

has

an

equalchance

of

being

selected

for

inspectionSystematic

samplesTo

sample

according

to

time

or

according

to

sequenceRational

subgroup

samplesTo

sample

by

a

group

of

data

that

is

logically

homogenousPlanning

for

InspectionCopyright

©

2017

Pearson

Education,

Inc.11-15Questions

to

answer

about

sampling:What

type

of

planning

will

be

used?Who

will

perform

the

inspection?Who

will

use

in-process

inspection?What

is

the

sample

size?What

are

the

critical

attributes

to

be

inspected?Where

should

the

inspection

be

performed?Control

PlansCopyright

©

2017

Pearson

Education,

Inc.11-16Provide

a

documented,

proactive

approach

to

defining

howto

respond

when

process

control

charts

show

that

aprocess

is

out

of

controlRequired

part

of

an

ISO

9000

quality

management

system(QMS)Control

Plan

SampleFigure

11-3Copyright

©

2017

Pearson

Education,

Inc.11-17Process

Control

ChartsStatistical

process

control

charts:Tools

for

monitoring

process

variationFigure

11-4Copyright

©

2017

Pearson

Education,

Inc.11-18Variables

and

Attributes

Control

ChartsCopyright

©

2017

Pearson

Education,

Inc.11-19VariableContinuous

measurement

such

as

height,

weight,

or

volumeAttribute

An

either-or

situation,

such

as

a

motor

starting

or

not,

or

a

lensbeing

scratched

or

notVariables

and

Attributes

Control

ChartsThe

most

common

types

of

variableand

attribute

chartsTable11-1Copyright

©

2017

Pearson

Education,

Inc.11-20Variables

and

Attributes

Control

ChartsCopyright

©

2017

Pearson

Education,

Inc.11-21Central

requirements

for

properly

using

process

charts:

You

must

understand

this

generic

process

for

implementingprocess

charts.You

must

know

how

to

interpret

process

charts.You

need

to

know

when

different

process

charts

are

used.

You

need

to

know

how

to

compute

limits

for

the

different

typesof

process

charts.We

treat

each

of

these

topics

separately.Variables

and

Attributes

Control

ChartsCopyright

©

2017

Pearson

Education,

Inc.11-22Steps

in

developing

process

control

charts:

Identify

critical

operations

in

the

process

where

inspection

migbe

needed.

These

are

operations

in

which

the

product

will

benegatively

affected

if

the

operation

is

performed

improperly.

Identify

critical

product

characteristics.

These

are

the

aspectsthe

product

that

will

result

in

either

good

or

poor

functioning

ofthe

product.

Determine

whether

the

critical

product

characteristic

is

avariable

or

an

attribute.Variables

and

Attributes

Control

ChartsCopyright

©

2017

Pearson

Education,

Inc.11-23Steps

in

developing

process

control

charts

(cont’d):4.

Select

the

appropriate

process

control

chart

from

among

themany

types

of

control

charts.

(This

decision

process

and

thetypes

of

charts

available

are

discussed

later.)4.

Establish

the

control

limits

and

use

the

chart

to

continuallymonitor

and

improve.4.

Update

the

limits

when

changes

have

been

made

to

the

process.Understanding

Process

ChartsProcess

charts

are

an

application

of

hypothesis

testingwhere

the

null

hypothesis

is

that

the

process

is

stable.For

example:Null

Hypothesis:

Ho:

m

=

11

inchesAlternative

Hypothesis:

H1:

m

11

inchesCopyright

©

2017

Pearson

Education,

Inc.11-24Understanding

Process

ChartsHypothesis

TestingFigure

11-5Process

ChartFigure

11-6Copyright

©

2017

Pearson

Education,

Inc.11-25•Copyright

©

2017

Pearson

Education,

Inc.11-26Figure

11-7Standard

Process

Chart

FormCopyright

©

2017

Pearson

Education,

Inc.11-27Figure

11-8Completed

Process

Chart

FormCopyright

©

2017

Pearson

Education,

Inc.11-28Figure

11-9x

and

R

Charts

Calculation

WorksheetCopyright

©

2017

Pearson

Education,

Inc.11-29Figure

11-10Calculations

for

Figure

11-8

DataCopyright

©

2017

Pearson

Education,

Inc.11-30Interpreting

Control

ChartsSignals

forconcern

sent

bya

control

chartHansen,

Bertrand

L.

Quality

Control:

Theory

and

Applications.

Upper

Saddle

River,

NJ:

Pearson

Education(1964).

ISBN:

013745208X.

©1964,

p.65.

Reprinted

and

Electronically

reproduced

by

permission

of

PearsonEducation,

Inc.,

New

York,

NY.Figure

11-11Copyright

©

2017

Pearson

Education,

Inc.11-31Interpreting

Control

ChartsSignals

forconcern

sent

bya

control

chart(cont’d)Hansen,

Bertrand

L.

Quality

Control:

Theory

and

Applications.

Upper

Saddle

River,

NJ:

Pearson

Education(1964).

ISBN:

013745208X.

©1964,

p.65.

Reprinted

and

Electronically

reproduced

by

permission

of

PearsonEducation,

Inc.,

New

York,

NY.Figure

11-11Copyright

©

2017

Pearson

Education,

Inc.11-32Interpreting

Control

ChartsCopyright

©

2017

Pearson

Education,

Inc.11-33Out-of-control

situations:

Two

points

in

succession

farther

than

two

standard

deviationsfrom

the

mean

Process

run

–

Five

points

in

succession

either

above

or

below

thecenter

lineProcess

drift

–

Seven

points,

all

increasing

or

decreasing

Erratic

behavior

–

Large

jumps

of

more

than

three

or

fourstandard

deviationsExample

11-1Problem:

The

Sampson

Company

produces

high-tech

radar

that

is

used

in

top-secret

weapons

by

the

Secret

Service

andthe

Green

Berets.

It

has

had

trouble

with

a

particular

roundcomponent

with

a

target

of

6

centimeters.

Samples

of

size

4were

taken

during

four

successive

days.The

results

are

in

the

following

table.Copyright

©

2017

Pearson

Education,

Inc.11-34Example

11-1Copyright

©

2017

Pearson

Education,

Inc.11-35

Develop

a

process

chart

to

determine

whether

the

process

is

stable.Because

these

are

measurements,

use

x

and

R

charts.

Using

the

calculation

work

sheet,

Figure

11-12

shows

the

values

for

theprocess

control

limits.

The

x

control

chart

for

this

problem

is

shown

with

the

appropriatelimits.

The

R

chart

is

also

in

control.

The

sample

averages

were

placedon

the

control

chart,

and

the

process

was

found

to

be

historically

incontrol.

Because

the

averages

and

ranges

fall

within

the

control

limitsand

no

other

signals

of

nonrandom

activity

are

present,

we

concludethat

the

process

variation

is

random.

Note

that

this

example

is

very

simple.

Generally,

you

use

15

to

20subgroups

to

establish

control

charts.Example

11-1•Copyright

©

2017

Pearson

Education,

Inc.Figure

11-1211-36Example

11-1Calculationsusing

ExcelMicrosoft

Excel,

Microsoft

Corporation.

Used

by

permission.Figure

11-13Copyright

©

2017

Pearson

Education,

Inc.11-37X

and

Moving

Range

(MR)

Charts

forPopulation

DataCopyright

©

2017

Pearson

Education,

Inc.11-38X

and

MR

charts

are

used

if

you

have

a

variablemeasurement

that

you

want

to

monitor

and

do

not

haveenough

observations

to

use

sampling.

Central

limit

theorem

does

not

apply,

which

may

result

in

the

databeing

non-normally

distributed.

Therefore,

there

is

an

increase

in

the

likelihood

that

you

will

drawan

erroneous

conclusion.It

is

best

to

first

make

sure

that

the

data

are

normally

distributedX

and

Moving

Range

(MR)

Charts

forPopulation

DatassX

chart

limitCenter

line:Limits:MR

limits

Same

as

R

chart

(where

n=2),except

that

the

ranges

are

computedas

the

differences

from

one

sample

to

the

nextCopyright

©

2017

Pearson

Education,

Inc.11-39Example

11-2Copyright

©

2017

Pearson

Education,

Inc.11-40Problem:

The

EA

Trucking

Company

of

Columbia,

Missourihauls

corn

from

local

fields

to

the

SL

Processing

Plant

inLincoln,

Nebraska.

Although

the

trucks

generally

take

6.5hours

to

make

the

daily

trip,

recently

there

seems

to

bemore

variability

in

the

arrival

times.

Mr.

Everett,

the

ownesuspects

that

one

of

his

drivers,

Paul,

may

be

visiting

hisgirlfriend

Janice

en

route

in

Kansas

City.

The

driver

claimsthat

this

is

not

the

case

and

that

the

increase

is

simplyrandom

variation

because

of

variability

in

traffic

flows.

Tdrivers

keep

written

logs

of

departure

and

arrival

times.Example

11-2oMr.

Everett

has

listedthese

times

in

thefollowing

table.

Youare

chosen

as

theanalyst

to

investigatethis

situation.

What

dyou

think?Copyright

©

2017

Pearson

Education,

Inc.11-41Example

11-2Solution:

Youdevelop

an

X

andMR

process

chart

totest

the

hypothesis.The

results

fromExcel

are

in

Figure11-14.Microsoft

Excel,

Microsoft

Corporation.

Used

by

permission.Figure

11-14Copyright

©

2017

Pearson

Education,

Inc.11-42Median

ChartsCopyright

©

2017

Pearson

Education,

Inc.11-43Median

charts

may

be

used

if

it

is

too

time

consuming

orinconvenient

to

compute

subgroup

averages

or

you

haveconcerns

about

the

accuracy

of

computed

means.Need

to

use

an

odd

sample

size,

usually

3,

5,

or

7Median

Chart

Limits::Center

lineControl

limits::Copyright

©

2017

Pearson

Education,

Inc.11-44Example

11-3Copyright

©

2017

Pearson

Education,

Inc.11-45Problem:

The

Luftig

food

company

has

gathered

thefollowing

data

with

weights

of

its

new

health

food

product.Because

the

published

weight

on

the

package

is

6

ounces,Mr.

Luftig

wants

to

know

if

the

company

is

complying

withweight

requirements.Example

11-3Twenty

samplesof

size

5

weredrawn.Copyright

©

2017

Pearson

Education,

Inc.11-46Example

11-3

Solution:

Resultsshow

that

the

processis

not

in

control,

withan

average

median

of6.23.

The

medianprocess

chart

doesshow

that

someproduct

is

being

madethat

is

below

6

ounces.It

also

shows

thatpoints

4,

7,

and

10

areout

of

control.Figure

11-15Microsoft

Excel,

Microsoft

Corporation.

Used

by

permission.Copyright

©

2017

Pearson

Education,

Inc.11-47•Copyright

©

2017

Pearson

Education,

Inc.11-48•Where:Copyright

©

2017

Pearson

Education,

Inc.11-49•Copyright

©

2017

Pearson

Education,

Inc.11-50•Table

11-3Copyright

©

2017

Pearson

Education,

Inc.11-51Example

11-4Copyright

©

2017

Pearson

Education,

Inc.11-52Problem:

Twenty

samples

were

taken

for

a

milled

rod.

Thediameters

are

needed

to

determine

whether

the

process

isin

control.

Because

these

milled

rods

must

be

measuredwithin

1/10,000

of

an

inch,

it

is

determined

that

the

processdispersion

is

important.Therefore,

you

need

to

use

an

s

and

x

chart

to

monitor

theprocess.

The

data

are

found

in

Figure

11-16.

We

have

20samples

with

n

=

3.Example

11-4Solution:

The

controlcharts

in

Figure

11-16

show

that

theprocess

is

in

control.There

is

no

need

forcorrective

action.The

solution

methodis

demonstrated

inthe

next

section.Figure

11-16Microsoft

Excel,

Microsoft

Corporation.

Used

by

permission.Copyright

©

2017

Pearson

Education,

Inc.11-53Other

Control

ChartsSummary

of

Variables

Chart

FormulasTable11-4Copyright

©

2017

Pearson

Education,

Inc.11-54Moving

Average

ChartCopyright

©

2017

Pearson

Education,

Inc.11-55A

chart

for

monitoring

variables

and

measurement

on

acontinuous

scale

by

using

past

information

to

predict

whatthe

next

process

outcome

will

beCusum

ChartA

chart

used

toidentify

slight

butsustained

shifts

inauniverse

in

whichthere

is

noindependence

between

observationsFigure

11-17Copyright

©

2017

Pearson

Education,

Inc.11-56Choosing

the

Correct

Variables

ControlChartFigure

11-18Copyright

©

2017

Pearson

Education,

Inc.11-57Corrective

ActionCopyright

©

2017

Pearson

Education,

Inc.11-58Corrective

action

steps

when

a

process

is

out

of

control:Carefully

identify

the

quality

problem.Form

the

appropriate

team

to

evaluate

and

solve

the

problem.

Usestructured

brainstorming

along

with

fishbone

diagrams

oraffinity

diagrams

to

identify

causes

of

problems.Brainstorm

to

identify

potential

solutions

to

problems.Eliminate

the

cause.5.

Restart

the

process.5.

Documentthe

problem,

root

causes,

and

solutions.5.

Communicate

the

results

of

the

process

to

all

personnel

so

thisprocess

becomes

reinforced

and

ingrained

in

the

organization.Using

Control

Charts

to

ContinuouslyImproveCopyright

©

2017

Pearson

Education,

Inc.11-59Two

key

concepts:The

focus

of

control

charts

should

be

on

continuous

improvement.

Control

chart

limits

should

be

updated

only

when

there

is

achange

to

the

process.

Otherwise,

any

changes

are

unexpected.Effects

of

Tampering

with

the

ProcessFigure

11-19Copyright

©

2017

Pearson

Education,

Inc.11-60Process

Capability

for

VariablesCopyright

©

2017

Pearson

Education,

Inc.11-61The

capability

of

a

process

to

produce

a

product

thatmeetspecificationWorld-class

levels

of

process

capability

are

measured

byparts

per

million

(ppm)

defect

levels.Process

Capability

for

VariablesSix

Sigma

programs

result

in

highly

capable

processes

and

aaverage

of

only

3.4

defects

per

million

units

produced.Figure

11-20Copyright

©

2017

Pearson

Education,

Inc.11-62Population

versus

Sampling

DistributionsCopyright

©

2017

Pearson

Education,

Inc.11-63Population

distributions

Distributions

with

all

individual

responses

from

an

entirepopulationPopulation

A

collection

of

all

the

items

or

observations

of

interest

to

adecision

makerSampleA

subset

of

the

populationSampling

distributionsDistributions

that

reflect

the

distribution

of

sample

meansPopulation

versus

Sampling

DistributionsPopulation

and

Sampling

Distributions

for

Class

HeightsFigure

11-21Copyright

©

2017

Pearson

Education,

Inc.11-64Population

versus

Sampling

DistributionsCopyright

©

2017

Pearson

Education,

Inc.11-65In

the

context

of

quality,

specifications

and

capabilityassociated

with

population

distributions.Sample-based

process

charts

and

stability

are

computedstatistically

and

reflect

sampling

distributions.Quality

practitioners

should

not

compare

process

chartlimits

with

product

specifications.Capability

StudiesCopyright

©

2017

Pearson

Education,

Inc.11-66Reasons

to

perform

a

process

capability

study:

To

determine

whether

a

process

consistently

results

in

productsthat

meet

specifications

To

determine

whether

a

process

is

in

need

of

monitoringthrough

the

use

of

permanent

process

chartsCapability

StudiesCopyright

©

2017

Pearson

Education,

Inc.11-67Five

steps

to

perform

process

capability

studies:

Select

a

critical

operation.

These

may

be

bottlenecks,

costlysteps

of

the

process,

or

places

in

the

process

in

which

problemshave

occurred

in

the

past.Take

k

samples

of

size

n,

where

x

is

an

individual

observation.Where

19

<

k

<

26If

x

is

an

attribute,

n

>

50

(as

in

the

case

of

a

binomial)Or

if

x

is

a

measurement,

1

<

n

<

11Use

a

trial

control

chart

to

see

whether

the

process

is

stable.Capability

StudiesFive

steps

to

perform

process

capability

studies

(cont’d)4.

Compare

process

natural

tolerancelimits

with

specificationlimits.

Note

that

natural

tolerance

limits

are

three

standarddeviation

limits

for

the

population

distribution.

This

can

becompared

wit