標(biāo)準(zhǔn)差教材二

6

OverviewSix

Sigma:A

DefinitionApplied

to

GEGE

Quality

InitiativeWhy

This

Approach?Origin

of

Six

SigmaThe

“Breakthrough

Strategy”Arriving

at

SigmaSix

Sigma

StructureKey

Concepts

&

ToolsA

Practical

ExampleAn

OverviewNot

a

lot

of

Details!!6

Overview“Six

Sigma”If

we

can’t

express

what

we

know

in

the

form

of

numbers,we

really

don’t

know

much

about

it.If

we

don’t

know

much

about

it,

we

can’t

control

it.If

we

can’t

control

it,

we

are

at

the

mercy

ofchance.Mikel

J.

HarryPresident

&CEOSix

Sigma

Academy,

Inc.“...will

bring

GE

to

a

whole

new

level

of

quality

in

a

fraction

of

thetime

it

would

have

taken

to

climb

the

learning

curve

on

our

own.”John

F.

Welch,

Jr.1995

GE

Annual

ReportA

Rigorous

Method

for

Measuring

&

Controlling

Our

Quality6

OverviewWhat

Does

“Sigma”

Mean?Sigma

is

a

Measure

of

the

Consistency

of

a

ProcessIt

(

is

Also

the

18th

Letter

in

the

Greek

Alphabet!Why

Does

GE

Need

A

Quality

Initiative?GE

Raising

The

BarNew

Goal

to

be

“Best

in

the

World”

vs.

#1

or

#2Customers

are

Expecting

More,

we

Must

Deliver“Ship-and-fix”

Approach

no

Longer

Tolerated

in

theMarketAim

to

Speed

Past

Traditional

Competitors

in

5

YearsGoal

Consistent

with

Reduced

Total

CostsWe

Must

Acknowledge

Our

VulnerabilitiesPoor

Quality

That

Impacts

CustomersProblems

withNPIToo

High

Internal

CostsWe

Need

a

Major

Initiative

to

Move

FromWhere

we

Are

to

Where

we

Want

to

be6

Overview6

OverviewWhy

Does

GE

Need

A

Quality

Initiative?Cost

of

Failure

(%

of

Sales)40%35%30%25%20%15%10%5%Defects

per

Million3.4233621066,807308,537500,000Sigma654321Estimated

Cost

of

Failure

in

US

Industry

is

15%

of

Sales;

TakingGE

From

a

3

to

a

6

Company

Will

Save

~

$10.5

Billion

perYear!Why

“Six

Sigma”?

Proven

Successful

in

“Quality-Demanding”

Industries

e.g.,Motorola,

Texas

Instruments

(many

process

steps

in

series)Proven

Method

to

ReduceCostsHighly

Quantitative

Method

–

Science

and

Logic

Instead

of

GutFeelIncludes

Manufacturing

&

Service

(close

to

customer)

and

Provides

Bridgeto

Design

for

Quality

ConceptsHas

Support

and

Commitment

of

Top

ManagementItWorks!!!6

OverviewSigma3

Spelling1.5

Misspelled

Wordsper

Page

in

a

BookMoney$2.7

Million

Indebtednessper

$1

Billion

in

AssetsTime3

1/2

Monthsper

Century4

1

MisspelledWordper

30

Pagesin

a

Book$63,000

Indebtednessper

$1

Billion

in

Assets2

1/2

Daysper

Century5

1

Misspelled

Word

ina

set

ofEncyclopedias$570

Indebtednessper

$1

Billion

in

Assets30

Minutesper

Century6

1

Misspelled

Word

in

allof

the

Books

in

a

SmallLibrary$2

Indebtednessper

$1

Billion

in

Assets6Secondsper

Century6

is

Several

Orders

of

Magnitude

Better

Than

3

!!!Sigma:

A

Measure

of

Quality6

OverviewWhere

Does

“Six

Sigma”

Come

From?Mikel

J.

Harry

one

of

the

Original

ArchitectsPreviously

Headed

Quality

Function

at

ABB

and

MotorolaNow

President/CEO

of

Six

Sigma

Academy

in

Phoenix,

ArizonaHas

Consulted

for

Texas

Instruments,

Allied

Signal

(and

others)Currently

Retained

by

GE

to

Teach

the

Implementation, Deployment

and

Application

of

Six

Sigma

Concepts

&

ToolsLearning

from

Those

Who

Have

had

SuccessWith

6

Will

Accelerate

its

Implementation

atGE6

OverviewSo...What

is

Six

Sigma?A

Measurement

SystemAProblem-SolvingApproachA

Disciplined

Change

Process“THE

SIX

SIGMA

BREAKTHROUGH

STRATEGY”MeasureAnalyzeImproveControl6

OverviewMeasurement

SystemHowDo

WeArrive

atSigma?Measuring&

Eliminating

Defects

isthe

“Core”of

SixIdentify

theCTQs“Critical

to

Quality”CharacteristicsortheCustomerRequirements

for

aProduct

orServiceLookforDefects

inProducts

or

ServicesCountDefects

or

failures

tomeetCTQrequirements

inall

process

stepsDefineDefect

OpportunitiesAny

step

in

theprocess

where

aDefectcouldoccurin

aCTQArrive

atDPMOConvertDPMOtoDefects

Per

MillionOpportunities

2PPM308,537SigmaUse

theSIGMATABLE366,80746,210523363.4Defects

perMillion

ofOpportunitySigmaLevel6

Overview2345308,53766,8076,210233PPMSIGMALEVELMILLIONOPPORTUNITYMeasurement

SystemSomeSigma

“Benchmarks”DEFECTSper

6

3.4AverageCompanyin3

toIRSTaxAdviceBestCompaniesAirlineSafetyGEAverageCompanyRestaurant

BillsAirline

BaggageDoctor’s

Prescription6

OverviewMeasurementSystemA

Graphic/Quantitative

Perspective

on

VariationManyData

Sets

Havea

NormalorBellShapeNumber

ofPeopleArrivingat

CRDTime7:007:15

7:30

7:45

8:00

8:15

8:30

8:45

9:00AverageValue9:156

OverviewProblemSolvingApproachOff-TargetCenterProcessReduceSpreadXXXXXXXXXXXXXXXXXXXXXXX

XXXXXXXXXXXXXX

XXXXUnpredictableXOn-Target6

Helps

usIdentify

and

ReduceVARIATIONdueto:Insufficient

Process

CapabilityUnstable

Parts&

Materials-

InadequateDesignMarginTargetUSLLSLTargetUSLLSLTarget“Lower

SpecificationLimit”

LSL

USL

“Upper

SpecificationLimit”LessVariation

MeansFewer

Defects&Higher

ProcessCenterProcessReduceSpreadOff-TargetUnpredictableOn-TargetDefects6

OverviewProblemSolvingApproach6

OverviewProblemSolvingApproachKeyComponents

of“BREAKTHROUGH

STRATEGY”A

Mix

ofConcepts

and

ToolsMeasureAnalyzeImproveControlIdentifyCTQ&CTP(Critical

toProcess)VariablesDoProcessMappingDevelopandValidateMeasurementSystemsBenchmark

andBaselineProcessesCalculateYieldandSigmaTargetOpportunitiesandEstablishImprovementGoalsUseofPareto

Chart&

Fishbone

DiagramsUseDesign

ofExperimentsIsolatethe“Vital

Few”fromthe“TrivialMany”Sourcesof

VariationTestforImprovementinCenteringUseofBrainstormingandAction

WorkoutsSetupControlMechanismsMonitorProcessVariationMaintain“InControl”ProcessesUseofControlChartsandProceduresWillAlso

IntegratewithNPIProcess6

OverviewDisciplinedChange

ProcessA

New

SetofQUALITYMEASURESCustomerSatisfactionCostofPoorQualitySupplierQualityInternalPerformanceDesign

forManufacturabilityWillApplyto

Manufacturing

&Non-ManufacturingProcessesand

be

Tracked&

Reportedby

Each

Busin6

OverviewStructureQuality

CouncilMembers:

Labs

&

Functions“Pipeline”

&

BB

Project

PrioritiesTraining

&

CertificationMeasurements

&

RewardsCommunicationsChampionsLeadership:

Overall

InitiativeProject

FundingHR:

Training

&

RewardsBlack

BeltsLead

6

Project

Teams“Measure/Analyze”“Improve/Control”O(jiān)ut

with

BusinessesHere

at

CRDMaster

Black

BeltsTeach6

Mentor

Black

BeltsMonitor

BB

ProjectsWork

“Pipeline”

ProjectsA

ResourcePoolTeam

MembersLearn/Use

6

ToolsWork

on

BBProjectsPart

of

The

JobOut

with

Businesses6

Projects

with

the

GEBusinesses

Tabulationof

GE

SixSigmaResultsBenefit

Target&

UpdateCurrent

benefits

level

@10.865

MMQPIDloading

:Carryoverfrom1999:CompletedProjects

2000

:3.313Active

Projects2000:3.285Total:10.865

MM4.059Key

Concepts

&Tools6

Overview6

OverviewChanging

FocusFromOutput

toProcessYDependentOutputEffectSymptomMonitorX1.

.

.

XNIndependentInput-ProcessCauseProblemControlIdentifying

andFixing

Root

CausesWillHelpus

Obtainthe

Desired

OutputY

=f

(X)Process

Capability6

OverviewTime

1Time

2Time

3Time

4SustainedCapabilityof

theProcess(longterm)USLTInherent

Capabilityof

theProcess(short

term)LSLTargetOverTime,a

“Typical”

ProcessWillShift

andDrift

byApproximately

1.5

6

Overview“ShortTermCentered”versus

“Long

Term

ShifteLSLUSLTProcessCapabilitySHORTTERMSix

SigmaCentered.001ppm.001ppm+6

LSLUSLT3.4ppmLONGTERMSixSigmaShifted1.5

ProcessCapabilityHigherDefectYieldinLongTermProcessCapabilitythanShortTermProcessCapability-6

4.5

1.5

6

OverviewTyingitAllTogethershiftCDAB0.51.01.52.02.5

123456CONTROLPOORGOODPOOR

GOODTECHNOLOGYProblemCouldbeControl,TechnologyorBothABCDGood

Control/Poor

TechnologyPoor

Control/PoorTechnologyPoor

Control/Good

TechnologyWORLD

CLASS!!!shortterm6

OverviewShortTermCapability

(Cp)ShortTermCapabilityRatioCp=USL

-

LSL6

ExampleUSL

=

3.0

LSL

= -3.0

6

3.0

-(-3.0

Cp=Cp=

1LSLUSL2.5

0.5

3.0

T

TargetProcessMeanA3

ProcessThePotentialPerformanceofaProcess,ifitWereonTarg6

OverviewLongTermCapability(Cpk)Cpk

=

Cp

-LongTermCapabilityRatioExampleCp=Target

==

-0.5

1(previouschart)0

0

-

(-0.5

3

Cpk

=

1-Cpk

=

0.83Off-TargetPenalty

Target

-

μ3

ThePotentialPerformanceofaProcess,CorrectedforanOff-TargetMeanLSLUSL2.5

0.5

3.0

T

TargetProcessMeanA3

Process6

OverviewZ

=Z-ScaleofMeasureAUnitofMeasureEquivalenttotheNumberofStandardDeviationsthataValueisAwayfromtheTargetValue-3.0-0.5

03.0Z

-ValuesUSLLSL2.5

0.5

3.0

=Process

MeanZT

TargetA

3

ProcessTheDefinitionsofYieldFinal

TestProcess(Process

4)PassProcess

3Process

1Process

2100(Units

Tested)65708291Yield

1Yield

2Yield

3Loss

19Rejects5Loss

312Loss

29FirstTime

Yield(Yft)Yield

ExclusiveofRework=Units

PassedUnits

Tested=6570=

0.93= (Yield1)=(Yield2)(Yield

3).

..

.82917082(

)

(91100)

()

(

)6570=

0.65Rolled

Thruput

Yield(Yrt)Probabilityof1/n(Yrt)==

(0.65)1/4Normalized

Yield(Ynm)AverageYield=

0.89(n:Total

NumberofProcesses)6

OverviewZeroDefectsof

AllProcesses6

OverviewThe

Impact

of

ComplexityProcess

Mean

Centered

on

Each

Operation

ProcessMeanShifted1.5

RolledYield1.000.900.800.700.600.500.400.300.200.100.001

10

100

1,000 10,000

100.000

1,000,000Number

of

Operations1

10100

1,000 10,000

100.000

1,000,000Number

of

OperationsRolledYield1.000.900.800.700.600.500.400.300.200.100.00As

the

Number

ofOperationsIncreases,

aHighRolled

YieldRequires

aHigh

forEachOperation5

4

3

6

6

5

4

3

atEachOperation6

Overviewp

(x)Baselining

&Benchmarking

anExisting

ProcessBenchmarkDefectsBaselineEntitlementBenchmarkA

World-Class

PerformanceEntitlementAchievable

Performance

Giventhe

Investments

Already

MadeBaselineThe

Current

Level

of

PerformanceBaselining=CurrentProcess/Benchmarking=UltimateGSomeBasic

6

-RelatedTools6

OverviewScatterDiagramOverSlept

CarWouldNotStartWeatherFamilyProblemsOtherPareto

DiagramFrequencyofOccurenceReasonsforBeing

Late

for

WorkArrivalTimeat

WorkTimeAlarmWentOffMaterialsPeopleThe

HistogramControl

Charts-6

OverviewSomeBasic6

-Related

ToolsThe

Fishbone

DiagramMeasurements Methods

TechnologyStatementCause

&

EffectBeingLateforWorkPlotof

Daily

Arrival

Time7:00Average

Value7:15

7:30

7:45

8:00

8:15

8:30

8:45

9:00

9:15TimeNumberofPeopleArrivingat

CRD6

OverviewLCLUCL

=

Upper

Control

LimitUCLSomeBasic

6

-RelatedToolsX

Bar

Chart

Range

ChartROut

of

Control

ConditionLCLUCLXLCL

=

Lower

Control

LimitX

=

MeanR

=

Average

RangeMonitorsChangesinAverageor

VariationOver

TimeDesign

of

Experiments6

OverviewSCREENINGOPTIMIZATIONCHARACTERIZATIONForExperimentsInvolving

aLargeNumber

of

FactorsUseful

inIsolatingthe“Vital

Few

“fromthe“TrivialMany”ForExperimentsInvolving

aRelativelySmall

NumberofFactorsUseful

WhenStudyingRelativelyUncomplicatedEffects&

InteractionsForExperimentsInvolvingOnly2or3FactorsUseful

WhenStudyingHighly

ComplicatedEffects&

RelationshipsDOEis

MoreEffective

Than

Testing

One

Factor

ata

T6

OverviewUsing

the

“One

Factor

ata

Time”ApproachThe

GoalReduce

Commute

toWorkto

15

Minutes(withoutworking

an

abnormal

work

schedule)The

VariablesTime

of

Departure

fromHome

&

Route

Takento

WorkThe

ApproachTry

3

Potential

Routes

at

CurrentDeparture

Time

(7:45),

Selectthe

Best

&

Vary

theDepartureTime

‘til

we

get

to

15Minutes7:30

7:45

8:00

8:15TimeofDeparture3217:15RouteCombinationSelectedTheResultUseRoute2andLeaveat7:15

toReach

Goal6

OverviewUsing“D“

esign

ofExperiments”(DOE)DOE(i)

BetterAccountsforInteractiveVariables

Missedby““One

FactorataTime””,and(ii)EfficientlySearchesfo“r“Sweet

Spo”t”in

ParameterSpaceTheVariablesTime

of

DeparturefromHome

&RouteTakentoWorkThe

ApproachVary

time

of

Departure

andRoute

Simultaneously,

in

aSystematic

FashionTheResultABetterCombinationAllowing15MoreMinutesofSleep!!!ActualCommuting

TimeAverages(minutes)3217:15

7:30

7:45

8:00

8:15Time

of

DepartureRoute172023211915182019161215212018OriginalConclusionBestCombination“Sweet

Spot”Reduce

Commute

toWorkto

15

Minutes(withoutworking

an

abnormal

work

schedule)The

GoalA

Practical

Example(The“C“

ookbook”)6

Overview6

and

BakingBreadFLOURYEASTUsinga

12StepProcess6

OverviewThe

“BETTER

BREAD”

CompanyStep

1Selecting

“Critical

to

Quality”

(CTQs

or

Y)Whatis

Important

totheCustomer?RiseTextureSmellFreshnessTasteY=

Taste!!6

OverviewMeasureStep

2Defining

Performance

Standards

for

CTQs

or

Y6

OverviewMeasureHow

CouldWe

MeasureTaste

(Y)?Panelof

TastersRating

Systemof

1to

10Target:

AverageRating

at8Desired:

NoIndividualRatings(“defects”)

Below

78910TargetDefectsWorstY=

12

34

567BestButIsthistheRightSystem?6

OverviewStep

3Validating

the

Measurement

System

for

YHow

CouldWe

Approach

This?Blindfolded

Panel

RatesSeveral

Loaf

SamplesPut

“Repeat”

PiecesfromSameLoafinDifferentSamplesConsistentRatings*onPieces

fromSameLoaf=

“Repeatability”ConsistentRatings*onSamples

AcrossthePanel=“R“

eproducibility”“Repeatability”

&“R“

eproducibility””Suggest

ValidMeasurementApproachPanelMemberLoaf1Loaf2Loaf3A589B491C492D898E482F591G892*Within

±OneTasteUnitMeasure6

OverviewAnalyzeStep

4Establish

Product

Capability

for

Y

(Taste)Thisisa3

Process!7Defects(ratingsbelow7)24Ratings(fromourpanel)=

.292292,000Defectsper1,ooo,oooLoavesOR76543217

8

9

101

2

3

4

5

6Rating#

ofRatings64321

143Defects

<7Target

=

8HowDoWeApproachThis?BakeSeveralLoavesUnde“r“Norma”l”

ConditionsHaveTasterPanelAgainDotheRatingAverageRatingis7.43

x

10

+

4

x

9

+

6

x

8

+

4

x

7

+

3

x

6

+

2

x

5

+

1

x

4

+

1

x

31

+

1

+

2

+

3

+

4

+

6

+

4

+

3ButVariationistooGreatfora6

Process6

OverviewAnalyzeStep

5Define

Improvement

Objectives

for

Y

(Taste)HowdoweDefineImprovement?BenchmarktheCompetitionFocusonDefects(i.e.tasterating<7)DetermineWhatisan“A“

cceptableSigmaLevel”SetImprovementObjectivesAccordinglyMaybea

5

Process

WillSuffice!1,000,000-100,000-...

...

...

...

...

...

...

...

...

..10,000-..

...

...

...

...

...

...

...

...

...1,000-.

...

...

...

...

...

...

...

...

...

.100-..

...

...

...

...

...

...

...

...

...1-23456710-.

...

...

...

...

.B..

e...s.t..

...

...

.“BETTERBREAD””BakingProcessCompetitorRangeforImprovementDefectsPerMillionSigmaScaleFreihoferWONDERPepperidge

FarmSunbeam6

OverviewStep

6Identify

Sources

of

Variation

in

Y

(Taste)HowdoweDetermine

thePotentialSourcesofVariation(Xs)?Have

theChefsBrainstormSome

LikelyOnes

Might

be:-AmountofSaltUsed-BrandofFlour-BakingTime-BakingTemperature-BrandofYeastYEASTFLOURMultipleSources:Chefs,Suppliers,ControlsAnalyze6

OverviewStep

7Screen

Potential

Causes

of

Variation

(Xs)HowdoweScreen

forCausesofVariation

(Xs)?DesignanExperimentUseDifferent

SourcesofPotentialVariationHave

PanelRatetheBreadUsedintheExperimentResults

Leadtothe“VitalFew”CausesYEASTFLOURSourceConclusionNegligibleMajorCauseNegligibleMajorCauseNegligibleFocusonThe““Vital

Few”Improve6

OverviewStep

8Discover

Variable

Relationships

Between

“Vital

Few”

(Xs)

and

YHowdoweFindtheRelationship

Betweenthe

“VitalFew””(Xs)and

Taste

(Y)?Conduct

aMoreDetailedExperimentFocus:OvenTemperaturefrom

325to375and3BrandsofFlourRUN#TEMPBRAND1325A2325B3325C4350A5350B6