Escolar Documentos
Profissional Documentos
Cultura Documentos
• A Metric
– Something we can measure for any
process
• Philosophy
– Becomes part of our culture
• Is defined to be the standard
deviation
Ζ Million
opportunities
σ
2 308,537
3 66,807
4 6,210
5 233
6 3.4
SQAM Course, Proficience, IISc 6
What is Six Sigma Performance?
99% Good (3.8σ) 99.99966% Good (6σ)
• 20,000 lost articles of mail per • Seven articles lost per hour
hour
• Unsafe drinking water for • One unsafe minute every seven
almost 15 minutes each day months
• 5,000 incorrect surgical • 1.7 incorrect operations per
operations per week week
• Two short or long landings at • One short or long landing every
most major airports each day five years
• 200,000 wrong drug • 68 wrong prescriptions per year
prescriptions each year
Define
Control
Measure
ACT PLAN
CHECK DO
Improve
Analyze
The DMAIC
Methodology
• Define :
– Establish project charter and scope
– Determine return, resources, timing
– Ensure management commitment
– List of customer (internal & external) requirements
• Measure :
– Generate process “is” map
– Prioritize project Y’s and document process input variables
– Access critical Y initial capability and measurement system
– Baseline data on current process performance
– Revisions/Improvements to project charter
– Access measurement system capability
1 R I R C
2 I C R I C
3 R R R I C
4 A C I C
5 R A
SQAM Course, Proficience, IISc 35
The GRPI Tool
•Storming
• Can I trust you?
• Is there group consensus on goals?
• What tasks must the team do?
SQAM Course, Proficience, IISc 49
Team Performance Model :
Four Stages of Team
Performance
•Norming
• How do we manage the team?
• Are conflicts resolved?
• Do we all own responsibility for success?
• Who are our stakeholders how do we resolve
needs?
•Performing
• Are boundaries and limits broken?
• Are activities coordinated?
• Is communication intuitive?
• Is there openness and frank discussion?
SQAM Course, Proficience, IISc 50
Team Management Exercise
Suppliers Inputs
x x Outputs Customers
x
Process
Pande
SQAM Course, Proficience, IISc 53
Six Sigma Process Map :
IPO (Inputs, Process, Outputs)
• The process map is derived from the SIPOC
diagram.
– S: Suppliers
– I : Inputs
– P: Process
– O: Outputs
– C: Customers
• Identifies the current process in terms of inputs,
process, and outputs.
• Provides initial team consensus on all input
variables (X’s) which may affect the process
outputs (Y’s). SQAM Course, Proficience, IISc 54
Steps to Complete the Process
Map
• Gather a cross-functional team which
is familiar with the process.
• At a reasonably high level, list all the
process steps. As a guide, the total
number of process steps should not
exceed six.
• List the outputs for each step.
• List the inputs that cause those
outputs to occur.
SQAM Course, Proficience, IISc 55
DMAIC Methodology :
The Tools
• Measure
– Prioritization Matrix (reducing the
number of X’s to consider for further
evaluation)
– Capability Analysis (Initial baseline
performance)
– Measurement System Analysis (Gage
R&R and %P/T)
– Basic Data Analysis Tools
(histograms, box plots, pareto
diagrams,SQAM
etc.)
Course, Proficience, IISc 56
DMAIC Tools :
Prioritization Matrix
• This is a simplified QFD (Quality Function
Deployment) matrix to rank the importance
of the input variables against the project
and customer requirements
• Key Outputs are scored for their
importance to the project and customer
• Key Inputs are scored based on their
relationship to these outputs
SQAM Course, Proficience, IISc 57
Prioritization Matrix Steps
1. Identify key outputs from the “30,000 ft”
process map
2. Rank order and assign priority factor to
each output (using a 1 to 10 scale)
3. List all process steps and inputs from the
detailed process map
4. Evaluate correlation of each input to each
output using a 0,1,3,9 ranking
– 0 = no correlation ; 9 = strong correlation
5. Cross multiply correlation values with
priority factors and sum for each input (X)
SQAM Course, Proficience, IISc 58
Prioritization Matrix Layout
Rating
Output
Output
Output
Output
Output
Output
Output
Output
Output
Output
Output
Output
Output
Output
Output
Total
1 0
2 0
3 0
4 0
5 0
6 0
7 0
8 0
Average or
20
15 Mean=14.17
10 Center Line
5 LCL=4.195
0
Subgroup 0 10 20 30
Lower
UCL=12.25
Control Limit
Mov ing Range
10
5
R=3.749
0 LCL=0
90
80
[ , ] 95.0% CI
80
70 Percent
70
60
60 50
Percent
Percent
50
40
40 30
30 20
20 10
10 0
1 2 1 2
Appraiser Appraiser
• 2 Operators
• 10 Samples
• 4 Different Sodas (Coke, Pepsi, RC,
HEB)
• 2 trials
A ction R esults
3 _
X=2.795
LCL=1.195
1
3 6 9 12 15 18 21 24 27 30
Observation
1.5
1.0
__
MR=0.602
0.5
0.0 LCL=0
3 6 9 12 15 18 21 24 27 30
Observation
3.5
3.0
FlightTime
2.5
2.0
1.5
WL1 WL2 WL3
WingLength
3.0
FlightTime
2.5
2.0
1.5
1.0 1.5 2.0 2.5 3.0
WL
Analysis of Variance
Source DF SS MS F P
Regression 1 1.0326 1.0326 6.70 0.015 WHAT IS THE
Residual Error 28 4.3160 0.1541 P-VALUE?
Total 29 5.3487
SQAM Course, Proficience, IISc 97
Regression Analysis
Example:
Output Detail
• The Linear relationship
– Algebra days are Y = MX + B
– Regression output is Y = B + MX
• FlightTime = 2.13 + 0.227 WL
• The R-Squared Statistic
– R-Squared is the per cent of the variation in
the response (Y) that is explained by the
model.
– Statistical packages use the method of “least
squares” which finds the best model which
minimizes the difference between the
observed response and the predicted response.
• R-Sq = 19.3%
SQAM Course, Proficience, =
R-Sq(adj) IISc16.4% 98
Regression Analysis Example:
Output Detail
• The ANOVA Table :
– Analysis of Variance
– Source DF SS MS F P
– Regression 1 1.0326 1.0326 6.70 0.015
– Residual Error 28 4.3160 0.1541
– Total 29 5.3487
Standardized Residual
90
Percent
0
50
10 -2
1
-3.0 -1.5 0.0 1.5 3.0 2.4 2.5 2.6 2.7 2.8
Standardized Residual Fitted Value
6
Frequency
0
4
2 -2
0
-3 -2 -1 0 1 2 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Standardized Residual Observation Order
2.25
Wing
Width
1.80
Analysis
Analysis of Variance for Flight Time (coded units)
DOE is a special case of ANOVA. DOE studies several X’s at the same time
in a systematic way.
P-Value = 0. Hence, for at least one of the X’s, changing from the low level
of that variable to the high level of that variable statistically affected flight
time. But which one(s)?
SQAM Course, Proficience, IISc 111
DOE 2 Example : The
3
Analysis
Estimated Effects and Coefficients for Flight Time (coded units)
Length.
Stem Therefore,
Length*Wing Width conclude that each of0.00000
0.00000 these variables
0.02500 statistically
0.00
1.000
affects flight time.
Wing Length*Stem Length*Wing Width 0.10000 0.05000 0.02500 2.00
SQAM Course, Proficience, IISc 112
0.081
DOE 2 Example : The
3
Analysis
• Since DOE is ANOVA but for several X’s at one time, it is
also VERY SIMILAR to regression and the output from a DOE
analysis can be used to predict the response (flight time)
for any combination of the X variables studied (wing length,
stem length, and wing width).
• Note that from the previous slide, the effects and the
coefficients column. The effect is the difference in the
response from the high level of X to the low level of X, or 1
– (-1), a 2 unit change. Regression measures the change in
the response per ONE unit change in X. Hence, the
regression coefficients are half the effect value.
William J. Hill: Six Sigma at Allied Signal, Inc. (Presentation at 1999 Q&P
Research Conference, May 1999)