Professional Documents
Culture Documents
TV SD South Asia
9/14/2007
INTRODUCTION
TV SD South Asia
9/14/2007
TV SD South Asia
9/14/2007
TV SD South Asia
9/14/2007
What is Six Sigma? Philosophy: The philosophical perspective views all works as a
processes that can be defined, measured, analyzed, improved & controlled (DMAIC). Processes require inputs & produce outputs. If you control the inputs, you will control the outputs. This is generally expressed as the y= f (x) concept.
TV SD South Asia
9/14/2007
What is Six Sigma? Methodology: This view of Six Sigma recognizes the underlying
and rigorous approach known as DMAIC. DMAIC defines the steps a Six Sigma practitioner is expected to follow, starting with identifying the problem and ending with the implementation of long-lasting solutions. While DMAIC is not only Six Sigma Methodology in use, it is certainly the most widely adopted and recognized.
TV SD South Asia
9/14/2007
9/14/2007
Six Sigma is a Standard of Excellence. It means no more than 3.4 Defects per Million Opportunities.
TV SD South Asia
9/14/2007
TV SD South Asia
9/14/2007
TV SD South Asia
9/14/2007
Benchmarking Chart
1,000,000
100,000
Restaurant Bills
10,000
1,000
100
10 World Class 1 2 3 4 5 6 7
11
9/14/2007
Defects are eliminated Production and development costs are reduced Cycle Times and Inventory Levels are reduced Profit Margin and Customer Satisfaction are improved
TV SD South Asia
12
9/14/2007
TV SD South Asia
13
9/14/2007
TV SD South Asia
14
9/14/2007
Process Process
Quality Quality
Control Control
is is
pre-requisite pre-requisite
for for
error error
free free
COPQ COPQis isa aresult resultof ofpoorly poorlycontrolled controlledprocess process Process ProcessControl Controlcan canbe bemeasured measuredin inPPM/Yield PPM/Yield PPM/Yield PPM/Yield measurements measurementsare arecorrelated correlatedto toCOPQ COPQ
Six Six Sigma Sigma has has shown shown that that the the Highest Highest Quality Quality Producer Producer is is also also the theLowest LowestCost CostProducer Producer
TV SD South Asia
15
9/14/2007
TV SD South Asia
16
9/14/2007
TV SD South Asia
17
9/14/2007
Sigma
2.0 3.0 4.0 5.0 6.0
DPMO
308,537 Non Competitive 66,807 6,210 Industry Average 233 3.4 World Class
TV SD South Asia
18
9/14/2007
Sigma
2.0 3.0 4.0 5.0 6.0
Yield
5% 93% 99.4% 99.976% 99.999655% World Class Industry Average Non Competitive
TV SD South Asia
19
9/14/2007
Long Term
LSL
TV SD South Asia
USL
20
9/14/2007
Six-Sigma Six-SigmaObjectives ObjectivesAre AreDirectly Directlyand andQuantifiably Quantifiably Connected Connectedto tothe theObjective Objectiveof ofthe theBusiness. Business.
TV SD South Asia
21
9/14/2007
Six SixSigma Sigmais isa arelentless, relentless,constant constant Journey Journeyof ofImprovement Improvement
TV SD South Asia
22
9/14/2007
TV SD South Asia
23
9/14/2007
Define
Measure Analysis Improve Control
TV SD South Asia
24
9/14/2007
TV SD South Asia
25
9/14/2007
Define
TV SD South Asia
26
9/14/2007
Customer Anyone Anyone internal internal or or external external to to the the Customer
organization organization who who comes comes in in contract contract with with the the product product or oroutput outputof ofwork work
Quality Quality Performance Performance to to the the standard standard expected expected by by
the theCustomer Customer
TV SD South Asia
27
9/14/2007
Customers willnot notrepurchase repurchasea acompanys companysproduct productif ifthey theyare are Customers will
not notsatisfied satisfiedwith withthe thecurrent currentcompany companyproduct. product. The The stronger stronger the the degree degree of of satisfaction satisfaction with with a a companys companys current current product product or or service, service, the the greater greater the the likelihood likelihood that that a a customer customer will will repurchase repurchasefrom fromthe thesame samecompany. company.
TV SD South Asia
28
9/14/2007
Uncertainty Uncertainty
Unknown Unknown Variation Disbelief Disbelief Risk Risk Defect DefectRate Rate
TV SD South Asia
29
9/14/2007
TV SD South Asia
Measure Measureof ofVariation Variationand andQuality Quality Measure Measureof ofProcess ProcessCapability Capability
30
9/14/2007
Customer
Supplier
Need
Do
Customers Customersand andSuppliers SuppliersExchange ExchangeValue Value Through Throughthe theNeed-Do Need-DoInteraction Interaction
TV SD South Asia
31
9/14/2007
Need
Customer
Do
Supplier Supplierstrives strivesfor forperformance performancein inCycle CycleTime, Time,Cost Costand andDefects Defects To ToMeet Meet Customers Customersexpectations expectationsin inDelivery, Delivery,Price Priceand andQuality Quality
TV SD South Asia
32
9/14/2007
TV SD South Asia
33
9/14/2007
Key Questions
What What does does the the phrase phrase critical critical to to satisfaction satisfaction mean mean
in CTS interms termsof ofa acustomer? customer? CTS What What does does the the phrase phrase critical critical to to quality quality mean mean in in terms CTQ termsof ofa aproduct productor orservice? service? CTQ What What does does the the phrase phrase critical critical to to delivery delivery terms termsof ofa aproduct productservice? service? CTD CTD What What does does the the phrase phrase critical critical to to cost cost mean mean mean mean in in in in terms terms
of CTC ofa aproduct productor orservice? service? CTC What What does does the the phrase phrase critical critical to to process process mean mean in in terms CTP termsof ofa aproduct productor orservice? service? CTP
TV SD South Asia
34
9/14/2007
CT Concept
Need Do Quality Defect-free Delivery Cycle time -Critical to Delivery Price Cost -Critical Cost to Cost
CTQ1 -Critical to Quality CTQ2 CTQ3 Processes CTP1 -Critical to Process 1 CTP2
TV SD South Asia
35
9/14/2007
CTQ1 -Critical to Quality CTQ2 CTQ3 Processes CTP1 - Critical to Process CTP2
Output
Y1
Input
f1 (X)
TV SD South Asia
36
9/14/2007
Critical to Characteristics
The The inherent inherent variation variation of of any any dependent dependent variable variable (Y) (Y) is is determined determined by by the the variations variations inherent inherent in in each each of of the the independent independentvariables variablesf(x). f(x). Product or Service requirement that impacts Quality, Delivery or Cost Product Capability Probability of Defects Critical-to-Quality characteristic CTQ1 Defect Opportunity
TV SD South Asia
Y = f (X)
One of the vital few process variable (x) that significantly affect Y
Probability of Defects
Y = f (X)
Y Y Dependent Dependent Output Output Effect Effect Symptom Symptom Monitor Monitor
X ..XN X 1 1..X N Independent Independent Input Input Variables Variables Cause Cause Problem Problem Control Control
TV SD South Asia
38
9/14/2007
f(X)
TV SD South Asia
39
9/14/2007
TV SD South Asia
40
9/14/2007
Business Metrics
Following FollowingBusiness BusinessMetrics Metricscan canbe beused usedin inSix SixSigma SigmaProjects. Projects.
Defect Defectper perUnit Unit(DPU) (DPU) Defects Defectsper perMillion MillionOpportunities Opportunities(DPMO) (DPMO) Throughput ThroughputYield Yield(Yield) (Yield) Rolled RolledThroughput Throughput Yield Yield (RTY) (RTY) Parts Partsper perMillion Million(PPM) (PPM)
TV SD South Asia
41
9/14/2007
Business Metrics
An An example example will will illustrate illustrate the the use use of of business business metrics metrics used used in in
previous previous slides. slides. Example: Example: A A process process produces produces 40000 40000 pencils. pencils. Three Three types types of of defect defect can can occur occur& & number numberof ofoccurrences occurrencesare: are: Blurred Blurredprinting printing 36 36 Wrong Wrong dimensions dimensions 118 118 Rolled Rolledends ends --11 11
TV SD South Asia
42
9/14/2007
Business Metrics Defect Defect per per Unit Unit (DPU): (DPU): = = Total Totalnumber numberof ofdefects defects//No. No.of ofunits units
= =165 165//40000 40000= =0.004125 0.004125
Throughput Throughput Yield Yield (Yield): (Yield): -DPU -0.004125 -DPU = -0.004125 = = =e e =e e = 0.996 0.996 Parts Parts per per Million Million (PPM): (PPM): = = DPU DPUx x10,00,000 10,00,000
= =0.004125 0.004125x x10,00,000 10,00,000= =4125 4125
TV SD South Asia
43
9/14/2007
Business Metrics Defect Defect per per Million Million Opportunities Opportunities (DPMO): (DPMO):
To To Calculate Calculate the the number number of of opportunities, opportunities, it it is is necessary necessary to to find find the the number number of of ways ways each each defect defect can can occur occur on on each each item. item. In In this this product, product, blurred blurredprinting printingoccurs occursin inonly onlyone oneway way (the (the pencil pencil slips slips in in the the fixture), fixture), so so in in the the batch batch there there are are 40,000 40,000 opportunities opportunities for for this this defect defect to to occur. occur. There Thereare arethree threeindependent independentplaces placeswhere wheredimensions dimensionsare arechecked, checked,so so there there are are 3 3x x 40,000 40,000 = = 1,20,000 1,20,000 opportunities opportunities for for dimensional dimensional defects. defects. Rolled Rolled ends ends can can occur occur at at the the top top and and // or or the the bottom bottom of of the the pencil, pencil, so so there thereare are40,000 40,000x x2 2= =80,000 80,000opportunities opportunitiesfor forthis thisdefect defectto to occur. occur. The The total total number number of of opportunities opportunities for for defects defects is is 40,000 40,000 + + 1,20,000 1,20,000 + + 80,000 80,000 = =2,40,000 2,40,000 DPMO DPMO= =(Total (Totalno. no.of ofdefects defectsx x10,00,000) 10,00,000)//(Total (Totalno. no.of ofopportunities) opportunities) = =(165 (165x x10,00,000) 10,00,000)//(2,40,000) (2,40,000)= =687.5 687.5
TV SD South Asia
44
9/14/2007
Business Metrics Rolled Rolled Throughput Throughput Yield Yield (RTY): (RTY):
RTY RTYapplies appliesto tothe theyield yieldfrom froma aseries series of of processes processes and and is is found found by by multiplying multiplying the the individual individual process process yields. yields. If If a a product product goes goes through through four four processes processes whose whose yields yields are are 0.994, 0.994, 0.987, 0.987, 0.951 0.951 & & 0.990, 0.990,then then RTY RTY= =0.994 0.994x x0.987 0.987x x0.951 0.951x x0.990 0.990= =0.924 0.924
TV SD South Asia
45
9/14/2007
Exercise-1 A A car car manufacturer manufacturerproduces produces15000 15000cars carsper permonth month..Three Three
types typesof ofdefect defectcan canoccur occurat atdifferent differentstage stage& &number numberof of occurrences occurrences are: are: Initial InitialAssembly Assembly 50 50(No. (No.of ofOpportunities Opportunities 4) 4) Intermediate IntermediateAssembly Assembly 95 95(No. (No.of ofOpportunities Opportunities 2) 2) Final FinalAssembly Assembly 35 35(No. (No.of ofOpportunities Opportunities 3) 3) Calculate Calculatethe theDefects Defectsper perUnit Unit(DPU), (DPU),Defect Defectper perMillion Million Opportunities Opportunities (DPMO), (DPMO),Throughput ThroughputYield Yield& &Parts Partsper per Million Million(PPM). (PPM).
TV SD South Asia
46
9/14/2007
Exercise-2 A
Process A Mixing 100# in 90# out 90%
B
Process B Making 100# in 80# out 80%
C
Process C Converting 100# in 90# out 90%
D
Process D Inspection 100# in 95# out 95%
Calculate Calculate the the Rolled Rolled Throughput Throughput Yield Yield (RTY) (RTY) for for the the above above process. process.
TV SD South Asia
47
9/14/2007
TV SD South Asia
48
9/14/2007
Pareto Charts
Pareto Pareto Diagrams Diagrams are are an an essential essential tools tools to to help help prioritize prioritize improvement improvement targets. targets. Paretos Paretos usually usually allow allow us us to to focus focus on on the the 20% 20% of of the the problems problems that that causes causes 80% 80% of of the the poor poor performance. performance.
Pareto Chart of Damage
9 8 7 6
Count
100 80 60 40 20 0
Percent
Interpreting the results: Focus on improvements to scratches and chips because 75% of the damage is due to these defects.
TV SD South Asia
49
9/14/2007
20 15 10 5
20 15 10 5 0
Period = Night
Period = Weekend
Interpreting the results: The night shift is producing more flaws overall. Most of the problems are due to scratches and peels. You may learn a lot about the problem if you examine that part of the process during the night shift.
Count
Peel
Flaws
You should drill down using third level, fourth level, etc., as far as it makes sense in solving your problem.
TV SD South Asia
50
9/14/2007
TV SD South Asia
51
9/14/2007
Project Charter
The The Project Project should should be be defined defined through through a a Problem Problem Description/ Description/ Project ProjectObjective Objectiveand andinclude: include: PPM PPMor orDPMO DPMOBaseline Baselinedata data Cost Costof ofPoor PoorQuality Quality(COPQ) (COPQ) Rolled RolledThroughput Throughput Yield Yield (RTY) (RTY) Inventory Inventory or or Other OtherAppropriate AppropriateMetric Metric
TV SD South Asia
52
9/14/2007
Status Summary
Status: Status Date: Review Date: Start Date : End Date : In Progress 30.08.06 30.08.06 10.01.06
Metrics
Business Metric(s): 1 Critical to Delivery 2 Critical to Cost 3 Z-Score Projected Savings: 3.52 Lacs 99.06 Lacs/Annum Baseline 28471 mtrs/shift Current 30442 mtrs/shift 3.52 Lacs Target 47295 mtrs/shift
Y and Complementary Ys
1 Improve OEE 2 Increase Availability
Constraints/Special Conditions
1. Target Productivity calculated based on 90% Efficiency and 7.45 hrs working time(Mini clean time reduced).
MileStones
Define 1 Team 2 Metrics 3 Process 4 Charter Measure MSA Baseline Capability Process Map X Shortlisting Root Cause Analyze 1st level cause Improve Control Statistical SolutionMonitoring Optimum Solution Standardization Implement SolutionTrain Validate Solution Maintain
TV SD South Asia
53
9/14/2007
TV SD South Asia
54
9/14/2007
Team Selection (ARMI) The The Six Six Sigma Sigma Team Team shall shall include include following following members. members. A AR RM M II Approver Approver Resources Resources
Champion Champion
Executive Executive Maser MaserBlack BlackBelt Belt
TV SD South Asia
55
9/14/2007
TV SD South Asia
56
9/14/2007
Top Top Level Level Process Process Map Map the the basic basic steps steps or or activities activities that that will will produce produce the the output output the the essentials, essentials, without without any any extras. extras. Everyone Everyone does does these these steps steps no no argument. argument. The The Top Top Level Level Flow Flow Map Map is is the the minimum minimum level level of of process process flow flow mapping mapping required required in in order order to to begin begin a a FMEA FMEA
TV SD South Asia
57
9/14/2007
INPUTS Assembly Labor Procedures Materials Equipment, Fixtures Environment Cleanliness Rework
OUTPUTS Part to Print Performance to Spec Visually acceptable Leak Free Identified Clean Packaged for use Consistent standard
TV SD South Asia
58
9/14/2007
Suppliers
INPUTS
OUTPUTS
59
9/14/2007
Basic Statistics
TV SD South Asia
60
9/14/2007
TV SD South Asia
61
9/14/2007
Type of Data
Attribute Data (Qualitative)
Categories like Machine 1, Machine 2, Machine 3 Yes, No Go, No Go or Pass/Fail Good/Defective On-Time/Late Discrete (Count) Data # of Maintenance Equipment Failures, # of freight claims Decimal subdivisions are meaningful Cycle Time, Pressure, Conveyor Speed
TV SD South Asia
62
9/14/2007
TV SD South Asia
63
9/14/2007
Data Distributions
Mean: Arithmetic average of a set of values Reflects the influence of all values Strongly Influenced by extreme values
Median: Reflects the 50% rank the center number after a set of numbers has been sorted from low to high. Does not include all values in calculation Is robust to extreme scores
Mode: The value or item occurring most frequently in a series of observations or statistical data.
TV SD South Asia
64
9/14/2007
n=
1 2 3 4 5 6 7 8 9
X= X=
n=
TV SD South Asia
65
9/14/2007
n=
1 2 3 4 5 6 7 8 9
X= X=
R= R=
n=
TV SD South Asia
66
9/14/2007
n=
1 2 3 4 5 6 7 8 9
X= X=
R= R=
How do we measure average variability from the Center?
n=
TV SD South Asia
67
9/14/2007
n=
1 2 3 4 5 6 7 8 9
X= X=
R= R=
V= V=
S= S=
n=
TV SD South Asia
68
9/14/2007
Measures of Variability
The Range is the distance between the extreme values of data set. (Highest Lowest) The Variance(S2) is the Average Squared Deviation of each data point from the Mean. The Standard Deviation (s) is the Square Root of the Variance. The range is more sensitive to outliers than the variance. The most common and useful measure of variation is the Standard Deviation.
TV SD South Asia
69
9/14/2007
X = Sample Mean
= Population Mean
= Population Standard
Deviation
Statistics
Estimate
Parameters
TV SD South Asia
70
9/14/2007
Variance
Xi
X=
i =1
S2 =
2 ( ) Xi X i =1
__
n
__
n 1
Standard Deviation
Standard Deviation
= R/ d 2
N 2 3 4 6
TV SD South Asia
S =
2 Xi X ( ) i =1
__
n 1
9/14/2007
__
2 =
2 ( ) Xi X i =1
__
Standard Deviation
=
TV SD South Asia
2 ( ) Xi X i =1
__
n
9/14/2007
72
TV SD South Asia
73
9/14/2007
TV SD South Asia
74
9/14/2007
TV SD South Asia
75
9/14/2007
TV SD South Asia
76
9/14/2007
68 % 95 % 99.73 %
-4
TV SD South Asia
-3
-2
-1
77
Output
4
9/14/2007
Lets Summarize
We need data to study, predict and improve the processes. Data may be Variable or Attribute. To understand a data distribution, we need to know its Center, Spread and Shape. Normal Distribution is the most common but not the only shape.
TV SD South Asia
78
9/14/2007
Exercise-3
TV SD South Asia
79
9/14/2007
Capability Analysis
TV SD South Asia
80
9/14/2007
If Ifwe wecannot cannotmeasure measureit, it,we wecannot cannotimprove improveit. it. Metrics MetricsHelp Helpus usunderstand understandwhere wherewe westand stand Metrics Metricshelp helpus usmove movein inthe theright rightdirection. direction. Metrics Metrics provide provide objectivity objectivity to to peoples peoples feelings feelings and and perceptions. perceptions. Metrics Metrics provide provide a a common common language language and and help help us us share share information informationwithout withoutsubjectivity, subjectivity,biases biasesand andconfusions. confusions. Metrics Metricshelp helpus usset seta acommon commongoal. goal.
TV SD South Asia
81
9/14/2007
Baseline
When When we we start start a a long long journey, journey, we we look look at at the the mile mile stones stones to to find findout outhow howfar farwe wehave havecome come But, But,suppose, suppose,we wedid didnot notknow knowwhere wherewe westarted, started,can canwe wetell tell how howfar farwe wehave havecome? come? Base Baseline linesets setsa astarting startingpoint pointof ofour ourjourney. journey. It Itis isthe thefirst firstmeasure measureand andtells tellsus usabout aboutas-is as-isState. State. It Italso alsohelps helpsus usto toset settargets targetsand andscope scopeout outproject. project.
TV SD South Asia
82
9/14/2007
We Weneed needto tomeasure measurewhatever whateveris isimportant importantto tothe theproject. project. We Weneed needto totrack trackthe thefollowing followingmetrics metrics 1. 1. 2. 2. Y Y Business BusinessMetrics Metrics
3. 3. Z ZScore Score
TV SD South Asia
83
9/14/2007
Y Yis isthe themetric metric(CTQ/CTD/CTC) (CTQ/CTD/CTC)that thatwe weare arefocusing focusingon. on.E.g. E.g. PPM, PPM, No. No. of of mistakes mistakes in in a a form, form, Dia Dia Variation, Variation, Power Power Factor, Factor,etc. etc. Cycle CycleTime, Time,Lead LeadTime, Time,Inventory InventoryLevel Leveletc. etc. Maintenance Maintenancecost, cost,Utility Utilitycost costetc. etc.
TV SD South Asia
84
9/14/2007
TV SD South Asia
85
9/14/2007
TV SD South Asia
86
9/14/2007
TV SD South Asia
87
9/14/2007
The Z - Transform
Z Z Transform Transform is is used used for for a a Normally NormallyDistributed Distributedprocess. process. It far __ Itexpresses expresseshow how farvalue valueX Xis is from fromthe theCenter Center X in interms termsof of E.g. a E.g. for for a normal normal distribution distribution __ with =10, with X =70 =70 and and =10, a a the the value value X=30 X=30 has has z=(30-70)/10=z=(30-70)/10=4 4 In Inother other words words the the point point X=30 X=30 is is4 4 away awayFrom Fromthe the mean mean on on ve veSide Side
Z=
Z
X X
__
TV SD South Asia
88
9/14/2007
The Z - Score
If If X X was was substituted substituted by by the the Tolerance ToleranceLimits Limits(LSL (LSL& &USL), USL), Z s Z will will tell tell how how many many s are are there there between between Tolerance Tolerance Limits Limits and and the the center center of of the the process. process. This This is is the the Sigma Sigma level level or or the the Z ZScore Score E.g. for E.g. __ for normal normal distribution distribution with =10, withX =70 =70 and and =10, LSL=30, LSL=30, Z L=(30-70)/10=-4 Z L=(30-70)/10=-4 For ForUSL=100,Z=3 USL=100,Z=3
30
Z=
4
SL X
__
70
100
TV SD South Asia
89
9/14/2007
But Wait.
Every Everytime timewe wemake makea abatch, batch,do dowe weget getthe theSAME SAMEamount amountof of Variation? Variation? Why Whydoes doesthe theprocess processVARY VARYfrom frombatch batchto tobatch? batch? Can Canwe weprevent preventBatch Batchto toBatch Batchvariation variationTOTALLY? TOTALLY? Data Data from from WHICH WHICH of of these these batches batches should should be be taken taken for for calculating calculatingZ ZScore? Score? How Howmuch muchdata datais isSUFFICENT? SUFFICENT? If IfI Iget getdifferent differentZ ZScores Scoresfor fordata datacollected collectedat atdifferent differenttimes times which whichof ofthem themis isCORRECT? CORRECT?
TV SD South Asia
90
9/14/2007
Well
Every Every time time we we make make a a batch, batch, we we may may NOT NOT get get the the same same amount amountof ofVariation. Variation. The Theprocess processvaries variesfrom frombatch batchto tobatch batchdue dueto toMEAN MEANSHIFT SHIFT over overa aperiod periodof oftime. time. We Wecan canMINIMIZE MINIMIZEthis thisbut butnot notprevent preventit. it. This This means means that that process process will will have have more more variation variation in in LONG LONG TERM TERMwhen whencompared comparedto toSHORT SHORTTERM TERMvariation. variation. A Aprocess processis isSix SixSigma Sigmaprocess processwhen whenthe theshort shortterm termZ ZScore Score is is6 6
TV SD South Asia
91
9/14/2007
LSL
Xlt
Xst
USL
TV SD South Asia
92
9/14/2007
Capability vs Performance
Zst Zst represents represents the the CAPABILITY CAPABILITY of of the the process process whereas whereas Zlt Zlt represents representsthe thePERFORMANCE PERFORMANCEover overa aperiod periodof oftime time What Whatis isCapability? Capability? Inherent Inherentability ability Due Dueto toCommon CommonCauses Causesof ofVariation Variation What Whatis isPerformance? Performance? Final FinalOutput Output Due Dueto toCommon Commonas aswell wellas asSpecial SpecialCauses Causesof ofvariation variation
TV SD South Asia
93
9/14/2007
TV SD South Asia
94
9/14/2007
TV SD South Asia
95
9/14/2007
Stability
When Whenthe thevariation variationis isonly onlydue dueto tocommon commoncauses, causes,the theprocess processis issaid saidto tobe be Stable. Stable. A AStable StableProcess Processhas haspredictable predictablevariation. variation. Special causes disturb the stability Special causes disturb the stability of of the the process process due due to to which which the the Variation becomes unpredictable. Variation becomes unpredictable. A AStable Stableprocess processis isalso alsocalled calledaaprocess processin inControl Control
7.5
UCL=5.98
5.0
Sample Mean
2 1 0 -1 -2 -3 -4
1 3 5 7 9 11 13 15 Sample 17 19 21 23 25
_ _ X=0.442
2.5
0.0
_ _ X=0.63
-2.5
LCL=-3.817
-5.0
1
LCL=-4.73
1 3 5 7 9 11 13 15 Sam ple 17 19 21 23 25
TV SD South Asia
96
9/14/2007
Capability vs Stability
Capability Capabilityhas hasa ameaning meaningonly onlywhen whena aprocess processis isstable. stable. If If a a process process is is out out of of control, control, first first we we need need to to stabilize stabilize the the process. process. Improvement Improvement in in the the inherent inherent variation variation can can be be made made only only when whenthe theprocess processis isstable. stable. Control ControlCharts Chartsare areused usedto tostudy studystability stability The The first first job job of of Six Six Sigma Sigma practitioner practitioner is is to to Identify Identify and and remove removeSpecial SpecialCauses Causesof ofVariation. Variation. Once Once the the process process is is made made predictable, predictable, the the next next job job is is to to identify identifythe thecauses causesof ofinherent inherentvariation variationand andremove removethem. them.
TV SD South Asia
97
9/14/2007
Calculating Capability
Capability Can be defined as Tolerable Variation Process Variation
LSL X + 1
USL
X + 2
USL LSL Cp = 6
TV SD South Asia
X + 3
T Cp = 6
9/14/2007
98
Calculating Capability
Marginal Capability
LSL X + 1
USL
X + 2 X + 3
Cp =
TV SD South Asia
T 6
6 Cp = 6
99
Cp = 1
9/14/2007
Calculating Capability
Six Sigma Capability
LSL
USL
X + 3 X + 6
Cp =
TV SD South Asia
T 6
Cp =
12 6
100
Cp = 2
9/14/2007
Calculating Capability
Calculate Cp from Upper & Lower Side
___
Cp L =
___
X LSL 3
CpU =
USL X 3
___
___
X LSL
USL X
10
12
14
16
18
20
Cp K
TV SD South Asia
101
9/14/2007
Calculating Performance
Calculate Cp from Upper & Lower Side
___
PpL =
___
X LSL 3
PpU =
USL X 3
___
___
X LSL
USL X
10
12
14
16
18
20
USL LSL Pp = 6
TV SD South Asia
102
9/14/2007
Capability vs Performance
USL LSL Cp = 6
__ __ X LSL USL X Cp K = Min , 3 3
Pp =
USL LSL 6
If Ifthe theformula formulaare aresame, same,what whatis isthe thedifference? difference? The Thedifference differenceis isin inSigma Sigmacalculation! calculation! Sigma Sigmain inCapability Capabilitycovers coversShort ShortTerm TermVariation. Variation. Sigma Sigmain inPerformance Performancecovers coversLong LongTerm TermVariation. Variation. How Howis isthe theData Datacollection collectionDifferent? Different?
TV SD South Asia
103
9/14/2007
Capability Calculation
Capability Capabilitycovers coversshort shortterm termvariation. variation. It Itrequires requiresdata datacollected collectedover overshort shortperiod periodof oftime. time. Small Small Subgroups Subgroups of of data data (Generally (Generally 3-7 3-7 sample sample size) size) are are taken. taken. Data Data points points within within a a subgroup subgroup need need to to be be of of consecutive consecutive output. output. Many Manysubgroups subgroupsare arecollected collectedover overa aperiod periodof oftime. time. The is considered to be Theaverage averagevariation variationwithin withina asubgroup subgroup is considered to be __ present presentthe theinherent inherentVariation. Variation. R = Sigma is calculated using d2 Sigma is calculated using Should Shouldbe beused usedonly onlywhen whenprocess processis isstable stable
TV SD South Asia
104
9/14/2007
Performance Calculation
Performance Performancecovers coverslong longterm termvariation. variation. It Itrequired requireddata datacollected collectedover overlong longperiod periodof oftime. time. Data Data should should represent represent more more than than one one day, day, if if possible possible more more than thana amonth monthof ofvariation. variation. If If Data Data points points are are too too many, many, one one may may randomly randomly sample sample the the data datato torepresent representall alldays daysand andbatches. batches. The The Root Root Mean Mean Square Square variation variation is is considered considered to to to to 2 __ represent representthe theOverall OverallVariation. Variation. xi X Sigma Sigmais iscalculated calculatedusing using =
n 1
Should Should be be used used only only when when data data is is available available over over a a long long period periodof oftime. time.
TV SD South Asia
105
9/14/2007
Cp K
Pp K
If If the the process process is is Stable, Stable, Collect Collect Short Short term term data data using Subgroups, using Subgroups, Calculate Cpk and Calculate Cpk and Zst=3xCpk Zst=3xCpk If If the the process process is is Unstable Unstable or or Stability Stability is is not not known, known, Collect Collect Long Long term term data data by by random sampling, random sampling, Calculate Ppk and Calculate Ppk and Zlt=3xPpk. Zlt=3xPpk. Convert Convert it it into into Short Short Term Term bench bench mark mark Z Z by byZst=Zlt+1.5 Zst=Zlt+1.5
&
ZST=3 X CpK
TV SD South Asia
106
9/14/2007
107
9/14/2007
Exercise-4
Calculate Calculate the the Z Z Score Score for for the the following following processes. processes.
1. 1. Data Dataof ofLength Lengthin inmachining machining process process has has been been collected. collected. The The required requiredlength lengthis is555 555max. max. The The Process Process has has been been behaving behaving consistently consistentlyfor forlast last6 6months. months. The Thedata datawas wascollected collectedfor forthe the first first5 5pieces piecesfor foreach eachof oflast last35 35 batches. batches.The Themean meanof ofthe thedata data is is540.26857 540.26857& &Standard Standarddeviation deviationis is10.38606. 10.38606. 2. 2. Data Datafor forCycle CycleTime Timeof ofa amolding moldingprocess processhas hasbeen beencollected. collected. The Thespecification specificationis is60-65 60-65sec. sec.The TheProcess Processhas hasbeen been behaving behavingerratically erraticallyfor forlast last8 8 months. months.The Thedata datacollected collectedfor for the thefirst first5 5pieces piecesfor foreach eachof oflast last30 30shifts. shifts.The Themean meanof ofthe the data datais is64.03827 64.03827& &Standard Standarddeviation deviationis is0.54618. 0.54618.
TV SD South Asia
108
9/14/2007
TV SD South Asia
109
9/14/2007
TV SD South Asia
110
9/14/2007
111
9/14/2007
112
9/14/2007
TV SD South Asia
113
9/14/2007
Optimize Process
Test and Verify Critical Input Variables Use Statistical Tools to Optimize Process Identify, Plan and Test Proposed Solutions Select Solutions and Confirm Results Implement Solutions and Improvement Plans Document Improvement Plans and Actions
TV SD South Asia
114
9/14/2007
Optimize Process
Each Standardization and Mistake Proofing Implement Process Controls Verify Effectiveness of Process Controls and System Improvements Monitor Process by Control Plan HOLD and GAINS
TV SD South Asia
115
9/14/2007
Physical Problem
Statistical Problem
Define the Problem in terms of Y and Probable Xs Measures the Extent of Problem Shortlist the Xs for analyses Collect Data Eliminate important Xs that are not
Finalize Xs for Optimization Confirm the short listed Xs Find Y=f(X) or Best case for the selected Xs Identify Physical controls based on constraints Verify controls and Maintain new
Physical Solution
Statistical Solution
TV SD South Asia
116
9/14/2007
THANK YOU
TV SD South Asia
117
9/14/2007