Post on 18-Jan-2015
description
Rev. 04/03/09 SJSU Bus. 142 - David Bentley 1
Week 11 - Six-Sigma Management and Tools
6Σ Organization, DMAIC, Taguchi Method, Robust
Design, Design of Experiments, Design for Six
Sigma, Reasons for 6Σ Failure
11/13/07 SJSU Bus. 142 - David Bentley 2
Topics
What is Six-Sigma? Organizing Six-Sigma DMAIC overview DMAIC phases The Taguchi method Design for Six-Sigma Using Six-Sigma from a contingency perspective
Rev. 11/15/07 SJSU Bus. 142 - David Bentley 3
Six Sigma Evolution Started as a simple quality metric at
Motorola in 1986 (Bill Smith) Migrated to Allied Signal
(acquired Honeywell and took its name) Picked up by General Electric
Commitment by CEO Jack Welch in 1995 Grown to be an integrated strategy for
attaining extremely high levels of quality
Rev. 10/29/08 SJSU Bus. 142 - David Bentley 4
What is Six-Sigma?
Sigma () is a Greek letter used to designate a standard deviation (SD)
in statistics Six refers to the number of SD’s
from the specialized limit to the mean. Six sigma is a fairly recent umbrella
approach to achieve quality
11/13/07 SJSU Bus. 142 - David Bentley 5
Percent Not Meeting Specifications
+1Σ = 32% +2Σ = 4.5% +3Σ = 0.3% +6Σ = 0.00034%
11/13/07 SJSU Bus. 142 - David Bentley 6
Six-Sigma Levels
Sigma Level Long-term ppm*
defects
1 691,462
2 308,538
3 66,807
4 6,210
5 233
6 3.4
Rev. 11/10/08 SJSU Bus. 142 - David Bentley 7
Statistics - DPU
Defect Six Sigma: “any mistake or error passed on
to the customer” ??? General view: any variation from
specifications DPU (defects per unit)
Number of defects per unit of work Ex: 3 lost bags ÷ 8,000 customers = .000375
Rev. 11/10/08 SJSU Bus. 142 - David Bentley 8
Statistics – dpmo (defects per million opportunities)
Process may have more than one opportunity for error (e.g., airline baggage)
dpmo = (DPU × 1,000,000) ÷ opportunities for error Ex: (3 lost bags × 1,000,000) ÷ (8,000
customers × 1.6 average bags)= 234.375
or (.000375)(1,000,000) ÷ 1.6 = 234.375
11/13/07 SJSU Bus. 142 - David Bentley 9
Statistics – dpmo (cont’d)
May extend the concept to include higher level processes E.g., may consider all opportunities for
errors for a flight (from ticketing to baggage claim)
11/13/07 SJSU Bus. 142 - David Bentley 10
Statistics - Off-Centering Represents a shift in the process mean Impossible to always keep the process
mean the same (this WOULD be perfection)
Does NOT represent a change in specifications
Control of shift within ± 1.5 σ of the target mean keeps defects to a maximum of 3.4 per million
11/13/07 SJSU Bus. 142 - David Bentley 11
Statistics - Off-Centering (cont’d)Source: Evans & Lindsay, The Management and Control of Quality, Southwestern, 2005
Rev. 11/03/08 SJSU Bus. 142 - David Bentley 12
k-Sigma Quality Levels
Number of defects per million opportunities For a specified off-centering and a desired quality level
11/13/07 SJSU Bus. 142 - David Bentley 13
k-Sigma Quality Levels Source: Evans & Lindsay, The Management and Control of Quality, Southwestern, 2005
11/13/07 SJSU Bus. 142 - David Bentley 14
Six Sigma and Other Techniques
Six-Sigma is … designed to handle the most difficult quality problems.
% Quality Problems
Techniques
90% Basic tools of Quality
< 10% Six-Sigma
< 1% Outside specialists
11/13/07 SJSU Bus. 142 - David Bentley 15
Organizing Six Sigma
11/13/07 SJSU Bus. 142 - David Bentley 16
Key Players Champion. Work with black belts to identify possible projects Master Black Belts. Work with and train new black belts Black Belts. Committed full time to completing cost-reduction projects Green Belts. Trained in basic quality tools
11/13/07 SJSU Bus. 142 - David Bentley 17
Distribution of Six Sigma Trained Employees
In a company with 100 employees there might be: One black belt Sixty green belts Some companies have yellow belts, employees familiar with improvement processes
11/13/07 SJSU Bus. 142 - David Bentley 18
Six Sigma Tools
DMAIC, Taguchi Method, Design for Six Sigma
11/13/07 SJSU Bus. 142 - David Bentley 19
DMAIC
11/13/07 SJSU Bus. 142 - David Bentley 20
DMAIC
DMAIC Overview
Stands for the six phases: Define Measure Analyze Improve Control
11/13/07 SJSU Bus. 142 - David Bentley 21
DMAIC
Define – (1)
Four Sub-Phases:1. Develop the business case2. Project evaluation3. Pareto analysis4. Project definition
11/13/07 SJSU Bus. 142 - David Bentley 22
DMAIC
Define – (2)
Developing the Business Case:1. Identify a group of possible projects2. Writing the business case3. Stratifying the business case into
problem statement and objective statement
Rev. 11/03/08 SJSU Bus. 142 - David Bentley 23
DMAIC
Define – (3)
RUMBA is a device used to check the efficacy of the business case
1. Realistic2. Understandable3. Measurable4. Believable5. Actionable
Rev. 11/15/07 SJSU Bus. 142 - David Bentley 24
DMAIC Measure – (1)
Two major steps: 1. Selecting process outcomes2. Verifying measurements
Rev. 10/29/08 SJSU Bus. 142 - David Bentley 25
DMAIC Measure – (2)
Selecting process outcomes (step 1)
Tools Used: Process map (flowchart) XY matrix (like QFD) FMEA (Failure Modes and Effects
Analysis) (aka DFMEA) Gauge R&R (Repeatability and
Reproducibility) Capability Assessment (cp or cpk)
11/13/07 SJSU Bus. 142 - David Bentley 26
DMAIC Measure – (3)
Verifying measurements (step 2) Tools Used:
Gauges, calipers and other tools. Management System Analysis (MSA) is
used to determine if measurements are consistent
11/13/07 SJSU Bus. 142 - David Bentley 27
DMAIC Measure – (4)
Gauge R&R Most commonly used MSA Determine the accuracy and
precision of your measurements
11/13/07 SJSU Bus. 142 - David Bentley 28
DMAIC Repeatability & Reproducibility
02/26/06 SJSU Bus. 142 - David Bentley 29
Measurement System DMAIC Evaluation
Variation can be due to: Process variation Measurement system error
Random Systematic (bias)
A combination of the two
02/26/06 SJSU Bus. 142 - David Bentley 30
DMAIC Metrology - 1
Definition: The Science of Measurement
Accuracy How close an observation is to a
standard Precision
How close random individual measurements are to each other
02/26/06 SJSU Bus. 142 - David Bentley 31
DMAIC Metrology - 2
Repeatability Instrument variation Variation in measurements using same
instrument and same individual Reproducibility
Operator variation Variation in measurements using same
instrument and different individual
02/26/06 SJSU Bus. 142 - David Bentley 32
DMAIC R&R Studies
Select m operators and n parts Calibrate the measuring instrument Randomly measure each part by each
operator for r trials Compute key statistics to quantify
repeatability and reproducibility
02/25/06 SJSU Bus. 142 - David Bentley 33
DMAIC R&R Spreadsheet Template
Rev. 11/27/06 SJSU Bus. 142 - David Bentley 34
DMAIC R&R Evaluation
Repeatability and/or reproducibility error as a percent of the tolerance Acceptable: < 10% Unacceptable: > 30% Questionable: 10-30%
Decision based on criticality of the quality characteristic being measured and cost factors
02/26/06 SJSU Bus. 142 - David Bentley 35
DMAIC Calibration
Compare 2 instruments or systems 1 with known relationship to national
standards 1 with unknown relationship to national
standards
11/13/07 SJSU Bus. 142 - David Bentley 36
DMAIC
Analyze – (1)
Three major steps: 1. Define your performance objectives
(X’s)2. Identify independent variables3. Analyze sources of variability
11/13/07 SJSU Bus. 142 - David Bentley 37
DMAIC
Analyze – (2)
Define your performance objectives (X’s) (step 1)
11/13/07 SJSU Bus. 142 - David Bentley 38
DMAIC
Analyze – (3)
Identify the independent variables where data will be gathered (step 2) Process maps (flowcharts), XY matrices,
brainstorming, and FMEA’s are the tools used
11/13/07 SJSU Bus. 142 - David Bentley 39
DMAIC
Analyze – (4)
Analyze sources of variability (step 3) Use visual and statistical tools to better
understand the relationships between dependent and independent variables
Rev. 11/03/08 SJSU Bus. 142 - David Bentley 40
DMAIC
Improve
Off-line experimentation
Analysis of variance (ANOVA) Determines whether independent
variable affect variation in dependent variables
Taguchi method or approach
Rev. 10/29/08 SJSU Bus. 142 - David Bentley 41
DMAIC
Control Phase
Manage the improved processes using control charts… covered in:
Variables Attributes
11/13/07 SJSU Bus. 142 - David Bentley 42
The Taguchi Method
Rev. 04/28/08 SJSU Bus. 142 - David Bentley 43
The Taguchi Method provides:
1. A basis for determining the functional relationship between controllable factors
2. A method for adjusting a mean of a process by optimizing controllable variables.
3. A procedure for examining the relationship between random noise … and product or service variability
Rev. 11/03/08 SJSU Bus. 142 - David Bentley 44
Design of Experiments (DOE)
Robust design – designed so that products are inherently defect free
Concept Design – considers process design and technology choices
Parameter Design – selection of control factors and optimal levels
Tolerance Design – specification limits
11/13/07 SJSU Bus. 142 - David Bentley 45
The Taguchi Process
1. Problem identification2. Brainstorming session3. Experimental design4. Experimentation5. Analysis6. Confirming experiment
04/28/08 SJSU Bus. 142 - David Bentley 46
Taguchi Quality Loss Function
Traditional view: anything within specification limits is OK, with no loss
Taguchi Any variation from the target mean
represents a potential loss The greater the distance from the target
mean the greater the potential loss
11/13/07 SJSU Bus. 142 - David Bentley 47
Design for Six Sigma
DFSS
Rev. 11/03/08 SJSU Bus. 142 - David Bentley 48
Design for Six-Sigma (DFSS)
Used in designing new products with high performance, instead of DMAIC1. DMADV (see next slide)2. IDOV (see 2 slides ahead)
Focuses on final engineering design optimization
Relates to new processes and products
11/13/07 SJSU Bus. 142 - David Bentley 49
DMADV
1. Design2. Measure3. Analyze4. Design5. Verify
11/13/07 SJSU Bus. 142 - David Bentley 50
IDOV
1. Identify2. Design3. Optimize4. Verify
11/13/07 SJSU Bus. 142 - David Bentley 51
Reasons for Six Sigma Failure
11/13/07 SJSU Bus. 142 - David Bentley 52
Reasons for Six-Sigma Failure - (1)
1. Lack of leadership by champions2. Misunderstood roles and
responsibility3. Lack of appropriate culture for
improvement
11/13/07 SJSU Bus. 142 - David Bentley 53
Reasons for Six-Sigma Failure - (2)
4. Resistance to change and the Six-Sigma structure
5. Faulty strategies for deployment6. Lack of data
11/13/07 SJSU Bus. 142 - David Bentley 54
Summary
The process for Six-Sigma is define, measure, analyze, improve and control
Keys to Six-Sigma success are skilled management, leadership and long-term commitment