You are on page 1of 92

3rd Batch of 3 Days Refresher Course designed for SIX SIGMA BB/ GB Holders

(Application of Shainin DOE tools)

Six Sigma- D.O.E.


Agenda Day-1 Need of Six Sigma & D.O.E. Pre- knowledge Test DMAIC Process D.O.E. Tools & Their Applications.
Paired comparison Product / Process Search Concentration Chart.
2

Six Sigma- D.O.E.


Agenda Day-2 D.O.E. Tools & Their Application (Contd..)
Recap / Exercise on the previous days tools. Component Search. Modified Component Search. Multi Vari Analysis

Six Sigma- D.O.E.


Agenda Day-3
Recap / Exercise on the previous days tools. B Vs C Analysis Control Phase. Q & A Session to clear the doubts. Post Knowledge Test Presentation by each group / member on the projects selected ( templates given only need to be used) Date & Time for the next meeting in the month of Jan 09
4

Brief Introduction of ACT & SIX SIGMA CLUB

Slide: 5

ACT- ACMA Centre For Technology

Objective of ACT

To make our members globally competitive in order to meet the increasing demand of QCD from the customers

Present ACT Activities


1.ACT Foundation Cluster 2.ACT- UNIDO Cluster 3.Advance Cluster on Lean. 4.ACMA Awards 5.AOTS program in Japan 6.AOTS program in India 7.WP.29- World Forum for Harmonization of Vehicle Regulations

Present ACT Activities


8.NATRIP-National Automotive Testing and R&D Infrastructure Project

9.TIFAC Technology Information, Forecasting & Assessment Council 10.CMVR- Centre Motor Vehicle Rules. 11. AIS-Automotive Industry Standards 12. BIS- Bureau of Indian Standards 13. National QC Convention. 14. Plant Visits & Technical Conferences.
9

ACTPublication
Launch of ACT publication ACTnow- ACT magazine is released on 4 monthly basis, covers the major activities done by ACT during that period while focusing one activity of ACT in detail. Also the cluster journey of one company, who have undergone ACT Cluster, is published in every issue. Five issues have already been released by now.

August08 Issue April08 Issue December07 Issue August 07 Issue April 07 Issue
10

SIX SIGMA
Back Ground:
First Batch of Black Belt Six Sigma Certification program was launched in Northern Region on 11th Nov2003. Program was customized & developed to suit ACMA member companies with the help of outside faculty Mr. Ramnarayan. Overwhelming response received from the members. Program was extended across all the regions.

SIX SIGMA
Status as on Mar08:

SIX SIGMA
Results:

In spite of the best efforts implementation was poor. At the same time the demand of High Quality & Low Cost is increasing day by day.

SIX SIGMA
Challenges ahead:
Keep this momentum alive. Expand the scope and understanding with more companies. Exchange ideas among Six Sigma Black Belt & Green Belt Holders. Enhance implementation of the knowledge . To promote Six Sigma Culture within the organisation for sustenance. To address customers demand of ZERO Defect Supplies.

SIX SIGMA CLUB

Creation & Development of Six Sigma Club was an option to address all these Challenges

Six Sigma Club website was launch on 6th September 07 in ACMAs AGM. This website would provide following features to Club members: (1) Discussion Forum. (2) Latest circulars regarding various training programmes. (3) Case Studies / Technical Documents.

Website Address: www.acmainfo.com/six-sigma


16

SIX SIGMA CLUB


Execution Plan:
Organization Structure: Formation of Six Sigma Club Executive committee, out of the members who had undergone Six Sigma certification course & obtained Black Belt / Green Belt status. Appointment of co-coordinator from ACMA secretariat.

SIX SIGMA CLUB


Eligibility Criteria:
Eligible for all ACMA member companies. Member has to be Six Sigma Black Belt or Green Belt Holder. Membership would be given at the company level only . However company can nominate any number of its employees. Member holding any position in the Executive Committee necessarily have to be Black Belt or Green Belt holder.

SIX SIGMA CLUB


Execution Plan:
Proposed Activities of Six Sigma Club: Launch of Six Sigma Club Web Site. To hold Seminars & Conventions. To share various Six Sigma Projects Completed. To update the status of activities done by individual companies on SIX SIGMA CLUB Website. Sharing information related to Six Sigma across the industries. Refresher course for BB/GB holders

SIX SIGMA CLUB


Execution Plan:
Communication Management: Through an active SIX SIGMA CLUB Website , administratively controlled by ACT Office. Telephonic Conversations. Conferences , Visits & Seminars.

Six Sigma- D.O.E.

D.O.EShainin Tools of Design of Experiments

21

Six Sigma- D.O.E.


Introduction of Six Sigma Six Sigma was first time developed & introduced by Sir Bill Smith in Motorola in 1987.

22

Six Sigma- D.O.E.


Introduction of Six Sigma Sigma represents variation in the process wrt mean (average line). Every process will have certain variation which is natural because no two things are identical.
23

Six Sigma- D.O.E.


Sources of Variations
Process Variations
Input Material Variations

Poor Process Design

Poor Product Design (Tolerancing)


Environment

Machine

Tooling

Parameter Variation

Method Variation

Processing Material

24

Six Sigma- D.O.E.


Lower Spec limit Nominal Target Upper Spec limit

6 Sigma= Tolerance % Defectives= 0.27 % 3 Sigma Level


Defects

Process Mean 6 Sigma represents Part to part variation


25

Six Sigma- D.O.E.


Lower Spec limit Nominal Target Upper Spec limit
6 Sigma= 50 %Tolerance % Defectives= 3 ppb 6 Sigma Level

Objective : Zero Defects Reduce variations

Process Mean
6 Sigma represents Part to part variation 26

Six Sigma- D.O.E.


Six Sigma ? Variation is nothing but its Part to Part Variation (PPV). Control of variation in the process (PPV) less than 50 % of the Tolerance is called Six Sigma.

27

Six Sigma- D.O.E.


Six Sigma ? 6 Sigma = PPV is less than 50 % of the Tolerance. Rejection Level=3ppb.

28

Six Sigma- D.O.E.


Objective of Six Sigma To Achieve Zero Defects by Reducing Process Variations.

29

Six Sigma- D.O.E.


Why DOE Tools?
Most Scientific way to find out the root cause of the problem DOE tools identify the Root Cause of the Problem based on the data not by assumptions.

30

Six Sigma- D.O.E.


DOE Tools application by, Taguchi Methods Shainin Methods

31

Six Sigma- D.O.E.


Introduction of Shainin DOE Tools
DOE (Design of Experiments) Tools, developed by Shainin based on statistics, help in achieving the objective of Six Sigma Level in each process of the organization .

32

Six Sigma- D.O.E.


Why Shainin DOE ?
Simple & Clear- No complexity in the process of data collection & its analysis. Data collection can be done on-line without stopping or disturbing the regular production. No complex mathematics, statistics used while analyzing the data & find out the root cause. Conclusion are drawn based on the data without having any ambiguity.
33

Six Sigma- D.O.E.


Shainin DOE Tools- 8 Problem Solving Tools
1. 2. 3. 4. 5. 6. 7. 8. 9. Paired Comparison. Product / Process Search. Concentration Chart. Component Search. Modified Component Search. Multi-Vari Analysis. B Vs C Analysis. Variable Search. Variable Search.
34

Six Sigma- D.O.E.


Shainin DOE Approach
DMAIC
Phase-3 Improve Phase-2 Measure & Analyze Phase-1 Define Phase-4 Control

35

Six Sigma- D.O.E.


DMAIC
Phase 1-D- Define- After selection of the project for Six Sigma based on Pareto Analysis, CTC,CTQ,Define the Problem Clearly & calculate COPQ( Cost of Poor Quality). Phase-2-M&A-Measure & Analysis the data using appropriate DOE Tools to pin point the Sources of Variation (SSV) & identifying Root Cause . Phase-3-I-Improve Actions are taken to eliminate the source of variation for the root cause identified. Phase-4-C-Control-Process controls methods are set to ensure that the counter measure taken in phase 3 stays permanently.
36

Six Sigma- D.O.E.


DMAIC & Tools
Phase
1-Define 2-Measure & Analyze 3-Improve 1. 2. 1.

Tools
Pareto Analysis. Process FMEA. 8 DOE TOOLS

1. 1. 2. 3.

B vs C Variation analysis. Control Charts. Pre- Control Charts.

4-Control

37

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING

Phase-1: Define
Step-1: Define the Quality Problem clearly. Step-2: Write down Part Name, Part No., Specifications, Tolerances & Instrument used. Step-3: Mention least count of the instruments used( should be 1/10 th of the tolerance allowed) Step-4: Mention the nature of problem.... Attribute or Variable. Step-5: Calculate & mention COPQ. Step-6: Write down the Sources of Variations (SSVs)for the problem.
38

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING

Y = f(x)
Where Y = Response ( Actual Problem)

X = Sources of Variations ( Probable causes of the Problem)


Note : Only those SSVs are to be considered wherein variation is possible & should be able to measure ( Quantify)
39

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING

Phase-2: Measure & Analysis


Step-1: Selection of DOE Tools for all SSVs one by one based on the selection criterion. Step-2: Data Collection and Applying the respective DOE Tool . Step-3: Analysis the data & finding out the Root Cause of the problem.
40

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
DOE TOOL1: Paired Comparison Application:
Can be used only when the SSVs are measurable on both Good & Bad Products. ( Good & Bad parts are selected based on the response / effect of the problem defined)
41

Six Sigma- D.O.E.


SHAININ DOE SOLVING TOOLS FOR PROBLEM
DOE TOOL1: Paired Comparison

Example :
Y= Poor draw ability
(Erickson cup value- Crack ) Xs = 1) % of C in material 2) Thickness of coil 3) Hardness of coil 4) Annealing Time 5) Annealing Temp

We can use Paired comparison tools for 1 & 3 & not for 2,4 & 5 .
42

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
DOE TOOL1: Paired Comparison ( contd..)
Step-1: Select 8 BOB( Best of Best) & 8 WOW ( Worst of Worst) Parts based on Response (Y).Min. should be 6. Step-2: Record all the values of x in a table for a particular SSVs and write the corresponding response Y of that value. Take one by one. Step-3: If SSV is an attribute ,quantify the observations on 1~ 5 Scale. Step-4:Arrange the data in the ascending order.
43

Six Sigma- D.O.E.


DOE TOOL1: Paired Comparison

Example:
Step a-Arrange values in ascending order

Step b From top , Draw line where transition takes place. - From bottom also do the same

Step c- If 2 values above and below the transition line are common reduce 0.5 from the count, upper and lowercase both

Step d If the values above and below the transition line are common more than 2 than. Treat entire block as one data and draw transition line.

XY XY 1.22BOB

XY XY 1.22BOB

XY XY 1.22BOB

XY XY 1.22BOB

1.22BOB 1.23BOB 1.23BOB 1.25BOB 1.25BOB 1.39WOW 1.39WOW 1.39 WOW 1.39 WOW 1.40 WOW 1.40 WOW 1.40BOB 1.40BOB 1.41 BOB 1.41 BOB 1.41 WOW 1.41 WOW 1.45WOW 1.45WOW 1.45WOW 11 1.45WOW .47 BOB .47 BOB

1.22BOB 1.23BOB 1.23BOB 1.25BOB 1.25BOB 1.39WOW 1.39WOW 1.39 WOW 1.39 WOW 1.40 BOB 1.40 BOB 1.40BOB 1.40BOB 1.41 BOB 1.41 BOB 1.42 WOW 1.42 WOW 1.45WOW 1.45WOW 1.45WOW1 1.45WOW1 .47WOW .47WOW

1.22BOB 1.23BOB 1.23BOB 1.25BOB 1.25BOB 1.25 WOW 1.25 WOW 1.39 WOW 1.39 WOW 1.40 BOB 1.40 BOB 1.40BOB 1.40BOB 1.41 BOB 1.41 BOB 1.41 WOW 1.41 WOW 1.45WOW 1.45WOW 1.45WOW1 1.45WOW1 .47WOW .47WOW

1.22BOB 1.23BOB 1.23BOB 1.39BOB 1.39BOB 1.39WOW 1.39WOW 1.39 WOW 1.39 WOW 1.40 BOB 1.40 BOB 1.41BOB 1.41BOB 1.41 BOB 1.41 BOB 1.41 WOW 1.41 WOW 1.45WOW 1.45WOW 1.45WOW1 1.45WOW1 .47WOW .47WOW

44

Six Sigma- D.O.E.


SHAININ DOE SOLVING TOOLS FOR PROBLEM

DOE TOOL1: Paired Comparison ( contd..)


Step-5:Count the numbers as per following guidelines: a) If both Topmost & Bottommost values of x belong to same category of response y, GOOD or BAD , the count is ZERO. b) If condition a do not fulfill , see for the transition line from Top & Bottom both independently where Good Changing into Bad or Bad is changing into Good. Draw the line at transition. Count the Number above & below the transition line & add. This is the total count for that SSV.
45

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING DOE TOOL1: Paired Comparison
c)Check at the transition line,if the 2 values are common,deduct 0.5 from the total count. Applicable for both upper & lower block. d) Check at the transition line ,if the values above & below are common > 2 , make a block and consider the full block as a transition line.Count above the line for upper block & below the line for lower block.
46

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
DOE TOOL1: Paired Comparison ( contd..)
Step-6:Conclusion If the Total counts are >= 6 , Proposed SSV is a Cause of the problem .If count is < 6 ,its not a cause & to be ignored.

47

Six Sigma- D.O.E.


SHAININ DOE SOLVING TOOLS FOR PROBLEM

DOE TOOL1: Paired Comparison ( contd..)

Count & Confidence Level:


Count
6 7 10 13

CL %
90% 95% 99% 99.9%
48

Six Sigma- D.O.E.


SHAININ DOE SOLVING Phase-3: Improve
DOE TOOL1: Paired Comparison ( contd.) Step-7: Deciding the Spec for SSV having count>6 from the data analyzed.

TOOLS

FOR

PROBLEM

49

Six Sigma- D.O.E.


For all the causes with Count >=6, fix the specification based on the GOOD band Fine tune the specification by adding or subtracting the existing tolerance depending on where the GOOD band is located For geometrical parameters, Do not do fine tuning, Fix only USL based on the data, since the LSL is always zero Calculate the variation as the maximum minimum of all the 12 or 16 data If the Variation is <=75% of the New tolerance, then the Confirmed cause is a ROOT CAUSE If the Variation is >75% of the New tolerance, then the Confirmed cause is only a CAUSE and this should be taken as Y and drilled down further

Six Sigma- D.O.E.


SHAININ DOE SOLVING Phase-3: Improve
DOE TOOL1: Paired Comparison ( contd.) Step-8: Is that the new specification is controllable ? If answer is Yes , make the improvement plan & implement.

TOOLS

FOR

PROBLEM

51

Six Sigma- D.O.E.


SHAININ DOE SOLVING Phase-3: Improve
DOE TOOL1: Paired Comparison ( contd.) Step-9: If answer is No. Investigate the reasons ( XSSV ) for such variation ( Response Y ) & find out the root cause.
52

TOOLS

FOR

PROBLEM

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
DOE TOOL1: Paired Comparison

Example :

Refer Case Study-1

53

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING DOE TOOL2: Product / Process Search
Application:
SSVs related to the Process Parameters or Input Material which can not be measured on both Good & Bad parts, Product / Process Search DOE Tool Used. Examples: Temperature, Pressure ,Pouring time, Input material wherein dimensions get changed during processing like: Drill dia in the case of Thread Loose or Tight Problem. ( Good & Bad parts are selected based on the response / effect of the problem defined )
54

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
Phase-2: Measure & Analysis
DOE TOOL2: Product / Process Search ( contd..)
Step-1: Collect the values of SSV x from the previous process / product & measure the Response Y till we get min. 8 BOB & 8 WOW. Or If collection of the data is not possible from the running process ( Inspection time is higher than the Production Cycle Time), select the lot size based on the historical rejection wherein part to part variation should get captured min 80% of the Tolerances specified. Step-2: Record all the values of x in a table for a particular SSVs and write the corresponding Response Y of that value.
55

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
Phase-2: Measure & Analysis
DOE TOOL2: Product / Process Search ( contd..) Step-3: Find out the Counts & decide the contribution of each SSV one by one in the problems defined. Step - 4 & 5 :are common to Paired Comparison as mentioned in the Previous slides .

56

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
Phase-2: Measure & Analysis
DOE TOOL2: Product / Process Search ( contd..) Step-6:Conclusion If the Total counts are >= 6 , Proposed SSV is a Cause of the problem .If count is < 6 ,its not a cause.

57

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
Phase-3: Improve
DOE TOOL2: Product / Process Search( contd.) Step-7: Deciding the Spec for SSV having count>6 from the data analyzed.

58

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
Phase-3: Improve
DOE TOOL2: Product / Process Search ( Contd..) Step-8: Investigation for such variation & make Improvement plan for the above Spec.

59

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
Phase-2: Measure & Analysis
DOE TOOL2: Product / Process Search ( contd..) Step-9 If the Input material found to be the root cause, we have to continue further analysis applying paired comparison or product Process Search. Step-10 If the root cause established is the Process parameter variation, we need to introduce controls in the machine
60

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
Phase-2: Measure & Analysis
DOE TOOL2: Product / Process Search( contd.) Step-11 If the Root cause established is neither related to Input Material nor to Process parameters Variation, We have to apply Variable Search for redesigning the Process parameters.

61

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
DOE TOOL2: Product / Process search

Example :
Y= Crack during Draw operation
Xs = 1) Annealing Time 2) Annealing Temp 3) Hardness variation

We can use Product / Process search tool for 1,2& Paired comparison for 3 .

Y = Internal Thread in the bush is Loose

Xs = 1) I.D. of the Bush 2) Spindle Vibration

W We can use Product / Process search tool for 1 & 2 & not the Paired C Comparison Tool.

62

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
DOE TOOL2: Product / Process search

Example :

Refer Case study-2

63

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
Phase-2: Measure & Analysis
DOE TOOL3: Concentration Chart Application:
Concentration Chart is mostly used for the attribute type of defects to identify Its intensity at various location / sector. Is used when defect can generate at multiple streams from the process and streams are too high to apply Multi-Vari Analysis. Example: Dents, Scratches, Burrs, Paint peel off, Dust, Porosity & Blow Holes in casting.

64

Six Sigma- D.O.E.


Data Collection
Collect 30 bad parts Scale the severity of defect on a scale of 1-3. 1 Less severity 2- Medium severity 3 High severity Divide the part into different zones depending on the geometry Count on every part, the number of defects generated in each zone Note down the severity in each zone

Analysis
Calculate for each zone, the Weighted sum of the severity Calculate the % with respect to the total in each zone

Conclusion
If in any zone the defect severity % is >=50%, then the defect is concentrated in that zone If in all the zones the defect severity is <50%, then the defect is coming randomly

Refer Case Study-2


65

Six Sigma- D.O.E.


SHAININ DOE TOOLS FOR PROBLEM SOLVING
Phase-2: Measure & Analyses
DOE TOOL4: Component Search ( contd.) DOE TOOL-5: Modified Component Se DOE TOOL- 6: Multi -Vari Analysis
Refer Flow Diagram

66

Six Sigma- D.O.E.

Refer Case study for CS & MVA Tool

67

Six Sigma- D.O.E.

B Vs C

68

M&A and Improvement Phase

DOE 7 B Vs C

Application:
Is used to validate the root cause of the problem identified using other DOE tools. (DOE Tool 1-6) If the cause is validated, then we can calculate how much the improvement has taken place Tool is applied only when B & C condition can be created alternately. Response is monitored in terms of Big Y
69 Selection of Correct Tool

Key Skill of BB :

Data Collection
If the tool used for finding out the cause or root cause is Paired comparison or Product/Process search and the total count is full count (16 or 12), then the sample size is 3B and 3C

Analysis
For 3B and 3C, check whether is there any overlap in the data or not. If there is no overlap, then the cause is validated If there is Overlap, then increase the sample size by another 3 and validate as per 6B,6C

Conclusion
For 3B, 3C, if there is no overlap, then the root cause is validated For 3B, 3C, if there is a overlap, then increase the sample size by 3

In all other situations, the sample size is 6B, 6C If the response is variable, produce 6 or 3

For 6B, 6C, find out the count using the same rules as Paired comparison

For 6B,6C if the total count is >=6, then the cause is validated For 6B,6C, if the total count is <6, then the cause is not validated and we have to go back to M&A phase

pieces in B condition and 6 or 3 Pieces in C


condition If the response is attribute, produce 6 or 3 Quantifying Improvement

Quantifying improvement

batches in B condition and 6 or 3 batches in C


condition If the response is variable, measure the actual value of the response as the data for analysis Quantify improvement only when the cause as per the above rule is validated If (Xb Xc) is >= K * Sigma (b), then the amount of improvement can be declared at 95% CL. If it is less, then get the K value at 90% CL and do the same check If (Xb-Xc) is < K* Sigma(b) even at 90% CL, increase the sample size in multiples of 3 Maximum boundary for the sample size is 12B, 12C. In 12B, 12C, if the amount of improvement cannot be declared, then go back to M&A phase

If the response is attribute, rejection % in each batch will be taken as the data for analysis If the validation is done using batches, decide the batch quantity to get at-least one bad product. Eg: if the historic rejection is 1%, then the batch quantity will be taken as 100 pcs Always collect data by alternating between the B and C condition

When calculating average, round off the decimal place to one decimal more than the data When calculating Sigma (b), round off the decimal to one decimal more than the data

70

B vs C
VALIDATION OF ROOT CAUSE
ROOT CAUSE: FOR PIPE WELDING OPERATION , FIXTURE IS REQUIRED RESPONSE : VARIATION IN DIM 91.75 FOUNDS AT HIGHER ( 92.2 92.9 ) ASSUMPTOIN : THE RESPONSE WITH B CONDITION MAY OVERLAP WITH RESPONSE WITH C CONDITION

B vs C DATA COLLECTION
HERE WE HAVE TAKEN 6 B PARTS & 6 C PARTS B = DIM 91.75 WITH NEWLY PROVIDED FIXTURE C = DIM 91.75 WITOUT FIXTURE / BY MANUAL OPERATION
ASC ORDER

DIM 91.75 B CAT C CAT 91.66 91.92 91.68 92.32 92.72 91.74 91.60 91.69 92.61 91.81 92.64 92.28

91. 60 91. 61 91. 64 91. 66 91. 68 91. 69 91. 72 91. 74 91. 81 91. 92 92. 28 92. 32

G G G G G B G B B B B B

HERE T.C. = 10 SO ROOT CAUSE IS VALIDATED

B vs C
DATA ANALYSIS & CONCLUSION
X bar-b = 91.651 X bar-c = 91.960 Diff X bar = 0.309 K for 95 % CL = 2.96 Sigma-b = 0.041 Sigma-b * k = 0.121 AS Diff X bar > Sigma-b * k SO ROOT CAUSE IS VALIDATED WITH 95 % C.L. Refer Case study

Six Sigma- D.O.E.


SHAININ DOE SOLVING TOOLS FOR PROBLEM Phase-4: Control Types of Process Variations

Part to Part Variation. .( Range Variation- R ) Time to Time Variation(Mean Variation- X Bar) Stream To Stream Variation.( Fixture or Tool to
Tool Variation)

74

Six Sigma- D.O.E.


Reason for Part to Part Variation(Range)
Machine Input Material Measurements Operator

75

Six Sigma- D.O.E.


Reason for Time to Time Variation ( Mean Shift)
Variation between one time block to another time block like: Shift Change, Tool Change, Material Change, Operator Change etc. Tool Wear

76

Six Sigma- D.O.E.


Reason for Stream to Stream Variation Product Stream ( Diameter at various locations etc) Process Stream ( Multiple cavities, spindles , fixtures etc)

77

Phase-IV: Control

Steps before calculating Process Capability or implementing Control Charts.

78

Phase-IV: Control
Construction of Histogram
Number of Groups = Sqrt (Number of data points) Range = Maximum Value - Minimum Value Class Interval (C.I) = Range/Number of groups Class Inverval (C.I) after rounding off to the nearest multiple of the Least count Draw the Histogram here 11 0.004 0.0004

-3

-2

-1

+1 +2 +3
Yes/No Yes/No Yes/No Yes/No

Does the Overall Average lie in the group having maximum frequency or in the adjacent groups Is there a gradual decreasing trend in frequency on both sides of the group having the maximum frequency Are there two Modes (Two groups having maximum frequency) and both the groups are distinctly seperated Based on the above analysis, can you conclude that the histogram is Normal
Note: If the His togram is non-normal, do not proceed further for analys is

79

Phase-IV: Control
Are there 7 consequtive points increasing, decreasing No and one side of the Mean range If yes write causes (if possible) NA

If the Process is not stable even after Homogenization, STOP HERE. Do not proceed further in the analysis. Identify the sources of variations for the instability and correct it Sigma Calculation Sigma = R-bar/d2 6 * Sigma 0.001 0.006
Sample 2 3 4 5 6 d2 1.128 1.693 2.059 2.326 2.534
80

Phase-IV: Control
Six Sigma Analysis
6 Sigma < T olerance 6 Sigma = T olerance 6 Sigma > T olerance Actions de cide d based on Six Sigma Analysis 6 Sigma < T olerance 1. Continue to use Control chart 2. Modify the Control Limits 3. T ighten the T olerance 4. Discontinue 100% inspection 6 Sigma = T olerance 1. Continue to use Control chart 2. Question the basis of T olerancing 3. Use DOE tools to reduce 6 Sigma 4. Do 100% inspection 6 Sigma > T olerance 1. Use DOE tools 2. Change the process 3. Do 100% inspection

Quality-Module

81

Six Sigma- D.O.E.

Control Charts
82

Six Sigma- D.O.E. Types of Control Charts


Control Charts
Variables
Pre control Charts X bar- R chart

Attributes p Chart C Chart

83

Six Sigma- D.O.E.

VARIABLE CONTROL CHARTS

84

Phase-IV: Control
Control Chart for Washer Thickness (mm)
80.000 79.000 78.000 77.000 76.000 75.000 74.000 73.000 72.000 71.000 70.000 16.000 14.000 12.000 10.000 08.000 06.000 04.000 02.000 00.000

Range

Mean)

0900

0930

1000

1030

1100

1130

1200

1230

1300

Time 0900 Hrs 0930 Hrs 1000 Hrs 1030 Hrs 1100 Hrs 1130 Hrs 1200 Hrs 1230 Hrs 1300 Hrs X-bar Range 73.577 5.232 74.585 7.379 74.717 4.746 73.233 6.834 75.408 5.964 75.576 8.581 75.414 7.508 74.274 7.068 74.552 8.408

85

Phase-IV: Control
Interpretation of Control Charts

A Process is in Control if
No sample points are outside limits Most points near process average About equal # points above & below centerline Points appear randomly distributed

86

Phase-IV: Control
Interpretation of Control Charts (Contd.)
The following is set of the rules: 2 out of 3 consecutive values of X on the same side of the CL (Centre line) and more than 2 SD (Standard deviations) from the CL. 4 out of 5 consecutive values of X on the same side of the CL and more than 1 SD from CL. 8 consecutive points are on the same side of the CL. 7 or more consecutive values in a consistently rising or falling pattern. a recurring cyclic pattern. abnormal clustering close to the CL. clustering of values close to both Control limits (suggests two distributions instead of one). is following

87

Phase-IV: Control
TRENDS
UCL CL LCL Sample statistics consistently decreasing.
X-bar Chart Causes Deterioration of Machine Tired Operator Tool Wear

Control Chart Patterns


UCL CL LCL Sample statistics consistently increasing.
Improvement / Deterioration of Operator Skill Tired Operator Change in Incoming Material Quality
88

R Chart Causes

Phase-IV: Control
Control Chart Patterns
UCL

PROCESS IN-CONTROL

CL LCL

Process operating ONLY under the influence of Chance (Inherent) Variation.


89

Six Sigma- D.O.E. Pre Control Chart


Application : Used when the actual Part to Part variation ( R bar) is <= 50 % of the Tolerance.

90

Red Zone Yellow Zone Green Zone Yellow Zone Red Zone

25 % of Tolerance 50 % of Tolerance 25 % of Tolerance

91

92

You might also like