Professional Documents
Culture Documents
SIX SIGMA
1
T 1 2 3 4 5 6
LSL USL
https://www.youtube.com/watch?v=4EDYfSl-fmc
◼ A statistical
measurement
◼ A business
strategy
◼ A philosophy
Six Sigma — A Statistical Measurement
◼It provides a quantitative performance
assessment of our products, services
and processes.
◼It allows for comparisons against other
similar or dissimilar products, services
and processes.
◼It provides feedback on how far ahead
or behind we are.
What Six Sigma Looks Like …
2 308,770 dpmo
3 66,811 dpmo
4 6,210 dpmo
5 233 dpmo
6 3.4 dpmo
Six Sigma — A Business Strategy
Inspection
Warranty Traditional Quality Costs
Scrap
(tangible)
Rework
Rejects
More Setups
Additional Costs of Poor Quality
Expediting Costs
Lost sales (intangible)
Late Delivery Lost Opportunity
Lost Customer Loyalty
Excess Inventory Hidden Factory
Long Cycle-Times
Engineering Change Orders
(Difficult or impossible to measure)
D EFINE
Define the project GOAL & customer (internal & external) deliverables
Project Objective:
What is the goal for the improvement team's project? Make sure that the problem and
goal are SMART(specific, measurable, attainable, relevant, and time-bound)? Has
anyone else (internal or external to the organization) attempted to solve this problem or
a similar one before? If so, what findings are useful for the success of the project?
GOAL = Baseline - (Baseline-entitlement)*70%,
Baseline is the average current process performance.
Entitlement is the best performance attained so far.
Why is this project important? Make sure that the project is linked to key business
goals and objectives. What business indices will be improved by the project? What are
the projected cost savings/opportunities on this project?
DEFINE MEASURE A N A LY Z E IMPROVE C O N TRO L
Trend chart
Plot a trend of the historical data with the baseline(average), entitlement, and goal.
Who will lead the team? Who are the members and their respective roles?
Project Scope:
What is the coverage of the project? Where does it start? What is the end point?
Is the project scope manageable? What are the existing constraints that might affect
the execution of the project?
Milestones:
Insert the roadmap here. Start date? Target completion date? What are the highlights
for the completion of the project and its respective estimated dates of completion?
Pareto Chart: Most Commonly Used Define
Phase Tool
- Generating Pareto chart using the Minitab
Pareto Charts - Why?
⚫ Focus efforts on the problems that offer the
greatest potential for improvement by showing
their relative frequency or size in a descending
bar graph
⚫ A Pareto chart is a powerful graphical tool for
separating the “important few” from the “trivial
many”
⚫ Construct Pareto charts at many levels to dig
down on a problem
Pareto Principle
Vilfredo Pareto - Italian Economist from the
1800’s
“Very few of the people have most of the
money”
80% of the problems Vital Few
are linked to only 20%
Trivial Many
of the causes
Powerful graphical tool
Useful when establishing priorities
Using A Pareto Chart
... ...
DEFINE MEASURE A NA LY Z E IMPROVE CONTROL
M EASURE
Establish a deeper understanding of the problem
by searching for the the potential root causes.
• Process map
• Rolled Throughput Yield
What is a Process Map?
Yes
No
No
No
Will customerallow
Call customerand Consolidated shipments?
Makecorrections
Yes
Consolidated orders
No No No
Can production
Will product Isproduction schedule Will customer No
Manufacture
Beavailable To manufacture Accept back Cancel order
Product in time
Before order ships? Product? Order?
For shipping?
Placeorder on hold
Expedite manufacturing
Until product
schedule
available
OPERATION
All steps in the process where the object undergoes a
change in form or condition.
TRANSPORTATION
All steps in a process where the object moves from one
location to another, outside of the Operation
STORAGE
All steps in the process where the object remains at rest, in
a semi-permanent or storage condition
DELAY
All incidences where the object stops or waits on a an
operation, transportation, or inspection
INSPECTION
All steps in the process where the objects are checked for
completeness, quality, outside of the Operation.
DECISION
Types of Process Maps
Activity flow
diagram
Deployment flow
diagrams
References: Yield the Right Way , Six Sigma and Beyond by Thomas Pyzdek
Learning Objectives
◼ Comparison between Conventional
Metrics and Six Sigma Metrics
◼ Introduction to
DPU
DPMO
Throughput Yield
Rolled Throughput Yield
Definition of Defective and Defect
Defective
◼ A unit of product that does not meet customer’s
requirement or specification.
◼ Also known as a non-conforming unit.
Example
◼ A blouse or shirt with a tear
◼ A cellphone with a malfunctioning LCD
Definition of Defective and Defect
Defect
◼ A flaw or a single quality characteristic that
does not meet customer’s requirement or
specification.
◼ Also known as a non-conformity.
◼ There can be one or more defects in a
defective.
Example
◼ The tear in a brand new blouse or shirt
◼ A scratch on the cellphone LCD
Capability Analysis — Conventional Metrics
◼ Reject Rate
in percent (e.g. 7.4%)
in ratio (e.g. 0.074)
in parts per million (e.g. 74,000 DPPM)
◼ Yield
in first-time yield Are we measuring
in final yield defectives
or defects?
Capability Analysis —
Six Sigma Metrics
◼ Defect Rate
in defects per unit (DPU)
in defects per opportunity (DPO)
or defects per million opportunities (DPMO)
◼ Yield
in throughput yield
in rolled throughput yield
in normalized yield
Defects Per Unit (DPU)
600,000
500,000
400,000
308,770
300,000
200,000
100,000 66,811
6,210 233 3.4 0.02
-
0 1 2 3 4 5 6 7
Sigma Capability
First-Time Yield (Yft)
First-Time Yield(unit-based) is the
number of units that pass a particular
inspection compared to the total number of
units that pass through that point in the
process.
Input 1,000 Process 900 Accept
For the 100 pcs rejected For the 150 pcs rejected For the 200 pcs rejected
Defects = 200 Defects = 350 Defects = 300
Ytp(A)= e–200/1,000 = 0.8187 Ytp(B)= e–350/975 = 0.6984 Ytp(C)= e–300/945 = 0.7280
• Fishbone Diagram
• Cause-and-Effect Matrix
• FMEA
What is a cause-and-effect diagram?
◼ It is a graphical display of potential causes of a problem. The lay-
out shows a cause and effect relationship.
◼ Also known as Ishikawa chart or fishbone diagram.
C/N/X
Material Methods
s
C C
N N
Problem/
Desired
N N Improvemen
N t
C
C C
Machinery Manpower
PROCESS FUNCTION POTENTIAL FAILURE POTENTIAL EFFECT SEV POTENTIAL CAUSES OF FAILURE OCC CURRENT CONTROLS DET. RPN RECOMMENDED ACTIONS RESP. DATE
MODE OF FAILURE
◼ Severity
Severity measures the impact of the effect on your customer.
It is rated on a pre-discussed scale.
Highest number means high impact.
1 means low impact
Descriptive Statistics — An Overview
Descriptive Statistics
Bar Chart
Trend Chart
Determining your Process Capability
a) percentage (%)
b) parts per million (ppm)
c) parts per billion (ppb)
Process Capability
Uses :
◼ Process Potential
➢ Cp
◼ Process Performance
➢ Cpu
➢ Cpl
➢ Cpk
Process Potential
Engineering Tolerance
Cp =
Natural Tolerance
USL − LSL
=
6
Process Performance
The Cpk index relates the scaled distance
between the process mean and the nearest
specification limit.
USL −
C pu =
3
− LSL
C pl =
3
Cpk = Minimum C ,C
pu pl
Within vs Overall Capability
Within Overall
A NALYZE
1. Organize experimentation schedule to determine which of the potential
KPIV’s are the real root causes.
2. Determine the sample size.
3. Analyze the data through the use of appropriate statistical techniques.
4. Implement the best settings as verified from the statistical analysis.
5. Follow-up progress of improvement plans concluded from the process
map.
DEFINE MEASURE ANALYZE IMPROVE CONTROL
➢ Scatterplot/Correlation
➢ Multi-vari Chart
➢ Confidence Intervals
➢ Sample Size Calculation
➢ One-sample tests
➢ Two-sample tests
➢ ANOVA
➢ Contingency Table
➢ Linear Regression
➢ Non-Parametric Tests
Multi Vari chart
The Multi-Vari Chart shows variability due to several sources.
It is very useful for comparison of within-unit variation to other
families of variation.
Variation
Data without dispersion information
is false data.
- Kaoru Ishikawa
Sources of Variation
Variation
Process Variation Measurement Variation
within unit within instrument
between units between instruments
between lots between calibrations
between operators between operators
between machines across time
between set-ups
across time
Three Families of Variation
Positional variation
- Manifests differing locations
simultaneously undergoing
same process
Examples:
- Temperature variations
inside a thermal chamber
1 2 3 4 5 6 7 8 9
- Cavity-to-cavity
Positional variation appears as the
variations in a plastic height of the vertical lines. Top of lines
indicate maximum readings, while bottoms
injection mold correspond to minimum values.
Three Families of Variation
Cyclical variation
-Occurs among sequential
repetitions of a process over a
fairly short time
Examples:
-Variations between consecutive
batches of a process
-Differences from lot to lot of raw 1 2 3 4 5 6 7 8 9
S XY
R =
S XX S YY Sample Covariance
(“+” or “-”)
where :
n
2
S YY = yi – y = (n – 1) s 2y
i=1
Properties of R
–1 R +1
It is a measure of the degree to which a linear
relationship exists between x and y.
R=1 Implies a perfect linear relationship
exists.
R = +1: perfect “positive” correlation
R = -1: perfect “negative” correlation
R= 0 Implies no linear relationship exists,
but does NOT imply no relationship at
all.
Coefficient of Correlation
Gather
evidence 1. Decide test statistic to use
(sampling) 2. Calculate sample size
Ho Ha 3. Collect sample data
DECIDE:
What does the evidence 1. Extract sampling distribution
suggest? 2. Set up Acceptance & Rejection
Reject Ho? zones
Or 3. Use the computed STATISTIC
to decide accept or reject the
Not Reject Ho? Ho
TESTS FOR
MEANS
92
Hypothesis Tests for Variance (2)
TESTS FOR
STANDARD DEVIATIONS
BARTLETT’S LEVENE’S
TEST TEST
Hypothesis Tests for Proportion ()
ATTRIBUTES
PROPORTION
DATA
Process 3 3 Y
3
Variation
What are the within
subgroups
2 hypotheses?
Overview of Regression
Correlation vs Regression
⚫ Regression Analysis
➢ Develops a prediction model, i.e. an equation that relates
the response to the predictor(s).
➢ Used when it is clear which variable is a predictor and
which is a response.
⚫ Correlation Analysis
➢ Assesses the strength of relationship between 2 variables
through correlation coefficient R. R measures how strong
the linear relationship is between the 2 variables.
➢ Used when it is not clear which variable is a predictor and
which is a response.
⚫ Both methods benefit from use of scatter plots.
Overview of Regression
Correlation vs Regression
⚫ Regression and correlation are closely related and
often confused together:
Computes R to quantify strength of the
relationship between 2 variables
➢ Correlation
Also proceeds to establish equation between
Actually regression the variables so that study is of practical value
analysis.
Establishes equation between 1 output variable
and 1 (or more) input variables
➢ Regression
However, also computes R2 to determine how
Actually some form well the model predicts the response
of correlation
analysis.
Francis Galton (1822-1911)
Fathers sons
fathers
Sons
Galton showed that the height of the sons of tall fathers regressed
towards the mean height of the population through several
successive generations. In other words, sons of unusually tall
fathers tend to be shorter than their fathers and sons of unusually
short fathers tend to be taller than their fathers.
Uses of Regression
(x) (Y)
Key Key
Process Process Process
Input Output
Varaibles Variables
Pasta Preparation
Amount of Water
Saltiness of pasta
Cooking time of pasta
Brand of pasta Cooking
Amount of Oil
Spaghetti
Amount of Salt
Tomato sauce for 1/2 kilo Wow!
Ground meat of pasta Delicious
Amount of Cheese
spaghetti!
Factor and Factor Levels
Pasta Preparation
Amount of water
> 4 L vs. 6 L
Amount of Salt How many factors were
> 2 tbsp vs. 4 tbsps. identified?
Cooking time of pasta How many levels for each
> 2 mins. vs. 5 mins factor?
Brand of pasta
> Brand A vs. Brand B vs. Brand C
Amount of Oil
> 1 tbsp. vs. 2 tbsps. Will the combined effect of
Amount of Salt: 0 vs. 1 tsp. vs. 2 tsps. salt and tomato sauce
influence the taste of
Tomato sauce: Brand A vs. Brand B vs.
Brand C spaghetti?
Ground meat: ground beef vs. corned beef
experimental design
consists of 2 factors
with 2 levels each that
will generate ||
2x2 = 4 runs
Coded
Values
2k Full Factorial Design
2-level Design
Number of Number of
Factors Runs A AB AB C
1 2 _
_ _ _ _ _
2 4 + _ _ _
+ +
3 8 _ _ _
+ +
4 16 _
5 32 + + + +
_ _ +
6 64
7 128 + _ +
_ + +
8 256
9 512 + + +
10 1024
The Three Approaches to DOE
➢ Classical DOE
- Based on the work of Ronald Fisher who applied DOE in
agriculture
- This is the full factorial and the most frequently used - the
fractional factorial
➢ Taguchi DOE
- Genichi Taguchi of Japan modified the classical approach,
simplifying it with his orthogonal arrays
➢ Shainin DOE
- Uses 10 distinct techniques invented or perfected by Dorian
Shainin each suited for a particular problem or application
- Easily became the world’s foremost quality problem solver
- However, Shainin techniques have not received wide
publicity because companies that used these techniques
and attained excellent results were unwilling to share them
with others.
DEFINE ME A S UR E A NA LY ZE IMPROVE CONTROL
DE F INE ME A S URE A N A LY Z E IMPROVE C ONTROL
◼ Poka-yoke(POH-kah YOH-kay) is
Japanese for mistake-proofing. Poka
means inadvertent mistake, while yoke
means prevent.
◼ A poka-yoke device is any mechanism
that either prevents a mistake from
being made, or makes the mistake
obvious at a glance.
Shigeo Shingo
◼ One of the industrial engineers at Toyota who
has been credited with creating and formalizing
Zero Quality Control (ZQC), approach to quality
management that relies heavily on the use of
poka-yoke devices.
◼ Has written 14 major books and hundreds of
important papers on manufacturing.
◼ The Shingo Prize is awarded for excellence in
manufacturing as a tribute to Dr. Shingo and his
lifelong work.
◼ He died in 1990.
Mistake and Mistake-Proofing
Machine/Tool
References
◼ Rath and Strong,Six Sigma Pocket Guide, AON Consulting Worldwide,
2001
◼ Subburaj Ramasamy, Total Quality Management, Macgraw-hill International
Ed., 2010